* IServer refactoring: pass IServer instances around, instead of peerids
refs #1363
This collapses 88 small incremental changes (each of which passes all tests)
into one big patch. The development process for the long path started with
adding some temporary scaffolding, changing one method at a time, then
removing the scaffolding. The individual pieces are as follows, in reverse
chronological order (the first patch is at the end of this comment):
commit
9bbe4174fd0d98a6cf47a8ef96e85d9ef34b2f9a
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 16:05:00 2011 -0400
immutable/downloader/status.py: correct comment
src/allmydata/immutable/downloader/status.py | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
commit
72146a7c7c91eac2f7c3ceb801eb7a1721376889
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 15:46:20 2011 -0400
remove temporary ServerMap._storage_broker
src/allmydata/mutable/checker.py | 2 +-
src/allmydata/mutable/filenode.py | 2 +-
src/allmydata/mutable/publish.py | 2 +-
src/allmydata/mutable/servermap.py | 5 ++---
src/allmydata/test/test_mutable.py | 8 ++++----
5 files changed, 9 insertions(+), 10 deletions(-)
commit
d703096b41632c47d76414b12672e076a422ff5c
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 15:37:05 2011 -0400
remove temporary storage_broker.get_server_for_id()
src/allmydata/storage_client.py | 3 ---
src/allmydata/test/no_network.py | 13 -------------
2 files changed, 0 insertions(+), 16 deletions(-)
commit
620cc5d80882ef6f7decfd26af8a6c7c1ddf80d1
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 12:50:06 2011 -0400
API of Retrieve._try_to_validate_privkey(), trying to remove reader.server
src/allmydata/mutable/retrieve.py | 10 +++++-----
1 files changed, 5 insertions(+), 5 deletions(-)
commit
92f43f856f4a8b36c207d1b190ed8699b5a4ecb4
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 12:48:08 2011 -0400
API of Retrieve._validate_block(), trying to remove reader.server
src/allmydata/mutable/retrieve.py | 14 +++++++-------
1 files changed, 7 insertions(+), 7 deletions(-)
commit
572d5070761861a2190349d1ed8d85dbc25698a5
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 12:36:58 2011 -0400
API of Retrieve._mark_bad_share(), trying to remove reader.server
src/allmydata/mutable/retrieve.py | 21 +++++++++------------
1 files changed, 9 insertions(+), 12 deletions(-)
commit
a793ff00c0de1e2eec7b46288fdf388c7a2bec89
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 12:06:13 2011 -0400
remove now-unused get_rref_for_serverid()
src/allmydata/mutable/servermap.py | 3 ---
1 files changed, 0 insertions(+), 3 deletions(-)
commit
1b9827cc9366bf90b93297fdd6832f2ad0480ce7
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 12:03:09 2011 -0400
Retrieve: stop adding .serverid attributes to readers
src/allmydata/mutable/retrieve.py | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
commit
5d4e9d491b19e49d2e443a1dfff2c672842c36ef
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 12:03:34 2011 -0400
return value of Retrieve(verify=True)
src/allmydata/mutable/checker.py | 11 ++++++-----
src/allmydata/mutable/retrieve.py | 3 +--
2 files changed, 7 insertions(+), 7 deletions(-)
commit
e9ab7978c384e1f677cb7779dc449b1044face82
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 11:54:23 2011 -0400
Retrieve._bad_shares (but not return value, used by Verifier)
src/allmydata/mutable/retrieve.py | 7 ++++---
1 files changed, 4 insertions(+), 3 deletions(-)
commit
2d91926de233ec5c881f30e36b4a30ad92ab42a9
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 11:51:23 2011 -0400
Publish: stop adding .serverid attributes to writers
src/allmydata/mutable/publish.py | 9 ++-------
1 files changed, 2 insertions(+), 7 deletions(-)
commit
47c7a0105dec7cbf4f7e0a3ce800bbb85b15df4a
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 11:56:33 2011 -0400
API of get_write_enabler()
src/allmydata/mutable/filenode.py | 7 ++++---
src/allmydata/mutable/publish.py | 4 ++--
src/allmydata/test/no_network.py | 3 +++
3 files changed, 9 insertions(+), 5 deletions(-)
commit
9196a5c6590fdbfd660325ea8358b345887d3db0
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 11:46:24 2011 -0400
API of get_(renewal|cancel)_secret()
src/allmydata/mutable/filenode.py | 14 ++++++++------
src/allmydata/mutable/publish.py | 8 ++++----
src/allmydata/mutable/servermap.py | 5 ++---
3 files changed, 14 insertions(+), 13 deletions(-)
commit
de7c1552f8c163eff5b6d820b5fb3b21c1b47cb5
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 11:41:52 2011 -0400
API of CorruptShareError. Also comment out some related+unused test_web.py code
src/allmydata/mutable/common.py | 13 +++++--------
src/allmydata/mutable/retrieve.py | 10 +++++-----
src/allmydata/mutable/servermap.py | 8 +++-----
src/allmydata/test/common.py | 13 ++++++++-----
4 files changed, 21 insertions(+), 23 deletions(-)
commit
2c1c314046b620c16f1e66d030c150d768b7d01e
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 12:01:46 2011 -0400
API of ServerMap.mark_bad_share()
src/allmydata/mutable/publish.py | 2 +-
src/allmydata/mutable/retrieve.py | 6 +++---
src/allmydata/mutable/servermap.py | 6 ++----
src/allmydata/test/test_mutable.py | 3 +--
4 files changed, 7 insertions(+), 10 deletions(-)
commit
1bed349030779fd0c378ae4e821384f953c6f6ff
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 11:11:17 2011 -0400
API+name of ServerMap.shares_on_server() : only for tests, so debug_ prefix
src/allmydata/mutable/servermap.py | 7 ++-----
src/allmydata/test/test_mutable.py | 6 +++---
2 files changed, 5 insertions(+), 8 deletions(-)
commit
2d32e448677d6b818692e801045d4115b29abf21
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 11:07:10 2011 -0400
API of ServerMap.all_servers_for_version()
src/allmydata/mutable/servermap.py | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
commit
48f3204d1889c3e7179578125c4bdef515af3d6a
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 11:04:50 2011 -0400
internals of ServerMap methods that use make_versionmap(), remove temp copy
src/allmydata/mutable/servermap.py | 28 +++++++++----------------
1 files changed, 10 insertions(+), 18 deletions(-)
commit
5c3da77b6c777a145bd5ddfaa4db849dc9495548
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 11:01:28 2011 -0400
API of ServerMap.make_versionmap()
src/allmydata/mutable/checker.py | 4 ++--
src/allmydata/mutable/retrieve.py | 5 ++---
src/allmydata/mutable/servermap.py | 4 ++--
src/allmydata/test/test_mutable.py | 7 ++++---
4 files changed, 10 insertions(+), 10 deletions(-)
commit
b6882ece49afb4c507d118af2db346fa329209dc
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 10:53:38 2011 -0400
make a copy of ServerMap.make_versionmap() (_make_versionmap2) for internal use
src/allmydata/mutable/servermap.py | 18 +++++++++++++-----
1 files changed, 13 insertions(+), 5 deletions(-)
commit
963f8e63faf32b950eb1b8103cd2ff16fe8f0151
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 00:45:58 2011 -0400
API of RetrieveStatus.add_problem()
src/allmydata/mutable/retrieve.py | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
commit
4976d29ffae565a048851601c29013bbae2976d8
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 00:45:05 2011 -0400
API of RetrieveStatus.add_fetch_timing()
src/allmydata/mutable/retrieve.py | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
commit
d057d3bbba72663ee148a8b916bc2d52be2e3982
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 00:44:04 2011 -0400
API of Retrieve.notify_server_corruption()
src/allmydata/mutable/retrieve.py | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
commit
8a2a81e46671c860610e0e96d6add1a57551f22d
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 00:42:32 2011 -0400
remove unused _outstanding_queries
src/allmydata/mutable/retrieve.py | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
commit
56d12cc9968d03ccd53764455c671122c4f391d1
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 00:40:57 2011 -0400
change Retrieve.remaining_sharemap
src/allmydata/mutable/retrieve.py | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
commit
4f0b7af4821f43290bfc70f2b1fc30149ad81281
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 10:40:18 2011 -0400
accessor for PublishStatus._problems
src/allmydata/mutable/publish.py | 4 +++-
src/allmydata/web/status.py | 2 +-
2 files changed, 4 insertions(+), 2 deletions(-)
commit
627087cf66d0b8cc519f4d551a967a7bd9b6a741
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 10:36:39 2011 -0400
accessor for RetrieveStatus._problems
src/allmydata/mutable/retrieve.py | 8 ++++++--
src/allmydata/web/status.py | 2 +-
2 files changed, 7 insertions(+), 3 deletions(-)
commit
ca7dea81f03801b1c7353fc00ecba689268109cf
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 00:35:32 2011 -0400
add .server to "reader", so we can get at it later
src/allmydata/mutable/retrieve.py | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
commit
6ef516e24908ec195af084a7550d1921a5e983b0
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 00:32:32 2011 -0400
temporarily give Retrieve a _storage_broker, so it can map serverids to servers
src/allmydata/mutable/checker.py | 3 ++-
src/allmydata/mutable/filenode.py | 6 ++++--
src/allmydata/mutable/retrieve.py | 5 +++--
src/allmydata/test/test_mutable.py | 4 ++--
4 files changed, 11 insertions(+), 7 deletions(-)
commit
afe08e4dd3f4ff9ff7e8a2a8d28b181e3625bcc9
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 00:21:51 2011 -0400
mutable/retrieve.py: s/peer/server/
src/allmydata/mutable/retrieve.py | 82 +++++++++++++-------------
src/allmydata/test/test_mutable.py | 6 +-
2 files changed, 44 insertions(+), 44 deletions(-)
commit
910afcb5d7f274880f68dd6cdb5b05f2bbc29adc
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 00:16:01 2011 -0400
web.status.PublishStatusPage: add comment, I think .problems isn't exercised
src/allmydata/web/status.py | 2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
commit
311466dd8c931bbba40d590ade867704282e7f1a
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:48:16 2011 -0400
API of PublishStatus.add_per_server_time()
src/allmydata/mutable/publish.py | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
commit
2df5faa1b6cbfbaded520d2320305a62fe961118
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:46:37 2011 -0400
more simplifications
src/allmydata/mutable/publish.py | 4 +---
1 files changed, 1 insertions(+), 3 deletions(-)
commit
6ac4544a3da385f2aad9392f906b90192f4f919a
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:44:08 2011 -0400
API of ServerMap.version_on_server()
src/allmydata/mutable/publish.py | 2 +-
src/allmydata/mutable/servermap.py | 4 ++--
src/allmydata/test/test_mutable.py | 5 ++---
3 files changed, 5 insertions(+), 6 deletions(-)
commit
3e187e322511072e4683329df6b2c6c733a66dba
Author: Brian Warner <warner@lothar.com>
Date: Tue Oct 4 00:16:32 2011 -0400
API of ServerMap.make_sharemap()
src/allmydata/mutable/servermap.py | 4 ++--
src/allmydata/test/test_mutable.py | 7 ++++---
src/allmydata/web/status.py | 4 ++--
3 files changed, 8 insertions(+), 7 deletions(-)
commit
318feed8437bdd8d4943c6569d38f7b54b6313cc
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:36:19 2011 -0400
small cleanups
src/allmydata/mutable/publish.py | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
commit
bd459ed5714e1db5a7163935c54b7b0b56db8349
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:33:39 2011 -0400
API of ServerMap.add_new_share()
src/allmydata/mutable/publish.py | 4 ++--
src/allmydata/mutable/servermap.py | 6 ++----
2 files changed, 4 insertions(+), 6 deletions(-)
commit
f2804fb6ed11d80088e0da8ed48e6c2922f2ffef
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:30:26 2011 -0400
API of ServerMap.get_bad_shares()
src/allmydata/mutable/publish.py | 3 +--
src/allmydata/mutable/servermap.py | 9 ++++-----
2 files changed, 5 insertions(+), 7 deletions(-)
commit
965074a47b3ce1431cb46d9a233840afcf9105f5
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:26:58 2011 -0400
more small cleanups
src/allmydata/mutable/publish.py | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
commit
38020da34f034f8889947dd3dc05e087ffff7106
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:18:47 2011 -0400
change Publish.bad_share_checkstrings
src/allmydata/mutable/publish.py | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
commit
5efebcbd2ee0c2f299ea86f7591d856c0f265304
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:16:31 2011 -0400
change internals of Publish.update_goal()
src/allmydata/mutable/publish.py | 8 +++-----
1 files changed, 3 insertions(+), 5 deletions(-)
commit
e91b55ff4c2a69165b71f2c7b217ac319ff4c527
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:11:42 2011 -0400
get rid of Publish.connections
src/allmydata/mutable/publish.py | 27 +++++----------------------
1 files changed, 5 insertions(+), 22 deletions(-)
commit
64e9a53b3229ebe2f9ebf7ed502d539311d0e037
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:05:32 2011 -0400
change Publish.bad_servers
src/allmydata/mutable/publish.py | 10 +++++-----
1 files changed, 5 insertions(+), 5 deletions(-)
commit
b85a934bef315a06bcfe00c9c12a3627fed2b918
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:03:07 2011 -0400
Publish.bad_servers: fix bug, this should be a set of serverids, not writers
src/allmydata/mutable/publish.py | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
commit
605ea15ec15ed671513819003ccd211cdb9761e0
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 23:00:21 2011 -0400
change .placed
src/allmydata/mutable/publish.py | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
commit
f7aba37b1b345d5b6d5cb16e3b3f6f3c1afb658e
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 22:59:22 2011 -0400
temporarily stash IServer as .server on the "writer" object
src/allmydata/mutable/publish.py | 2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
commit
f9b551d788e7db1f187fce5ab98ab5d5fe4e1c36
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 22:48:18 2011 -0400
change Publish.goal and API of log_goal() to use IServer, not serverid
src/allmydata/mutable/publish.py | 48 ++++++++++++++--------------
1 files changed, 24 insertions(+), 24 deletions(-)
commit
75f20616558e4900b8b1f685dd99aa838de6d452
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 15:27:02 2011 -0400
API of ServerMap.get_known_shares()
src/allmydata/mutable/publish.py | 16 ++++++++++------
src/allmydata/mutable/servermap.py | 7 ++-----
2 files changed, 12 insertions(+), 11 deletions(-)
commit
1c38c9d37bb08221b4418762234b1a62397b3b4b
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 15:20:29 2011 -0400
Publish.full_serverlist
src/allmydata/mutable/publish.py | 10 +++++-----
1 files changed, 5 insertions(+), 5 deletions(-)
commit
b6cbd215a04b9cde31a7d92a97a7f048622b16f1
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 15:12:31 2011 -0400
API of ServerMap.all_servers()
src/allmydata/mutable/servermap.py | 19 ++++++-------------
1 files changed, 6 insertions(+), 13 deletions(-)
commit
e63cd0315fae65357b1727ec6d5ff3c6e0d27c98
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 15:10:18 2011 -0400
remove ServerMap.connections, set_rref_for_serverid()
src/allmydata/mutable/servermap.py | 11 +----------
1 files changed, 1 insertions(+), 10 deletions(-)
commit
4df52db2f80eb12eefa5d57103c24893cde89553
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 15:04:06 2011 -0400
API of ServerMap.mark_server_reachable()
src/allmydata/mutable/servermap.py | 7 ++-----
1 files changed, 2 insertions(+), 5 deletions(-)
commit
69c715bde77944dc25181b3dbbeb042c816f9a1b
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 15:03:21 2011 -0400
API of ServerMap.mark_server_unreachable()
src/allmydata/mutable/servermap.py | 9 +++------
1 files changed, 3 insertions(+), 6 deletions(-)
commit
3d784d60eec1c508858e3a617e4411ffbcc3c1fa
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 15:02:03 2011 -0400
API of status.set_privkey_from()
src/allmydata/mutable/servermap.py | 7 +++----
1 files changed, 3 insertions(+), 4 deletions(-)
commit
544ed3ea29bed7e66da7fd29ca3f6f076f27a9e6
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 15:01:15 2011 -0400
API of status.add_per_server_time()
src/allmydata/mutable/servermap.py | 7 ++++---
1 files changed, 4 insertions(+), 3 deletions(-)
commit
fffe5008b6320bd1e04c3c68389a2bf2ee383fa8
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:59:02 2011 -0400
remove unused .versionmap
src/allmydata/mutable/servermap.py | 7 -------
1 files changed, 0 insertions(+), 7 deletions(-)
commit
2816562e090d2294179db3588dafcca18de1bc2b
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:57:51 2011 -0400
remove serverid from all log messages. Also one unused lambda.
src/allmydata/mutable/servermap.py | 30 +++++++++++++-------------
1 files changed, 15 insertions(+), 15 deletions(-)
commit
28fa6b1a2738fa98c1f1dbd3d0e01ae98912d11f
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:54:30 2011 -0400
removed unused _readers
src/allmydata/mutable/servermap.py | 3 ---
1 files changed, 0 insertions(+), 3 deletions(-)
commit
a8e4ed3d645ab592d1add6a1e69b6d1ebfb77817
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:54:16 2011 -0400
remove unused _sharemap
src/allmydata/mutable/servermap.py | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
commit
3f072e55cf1d0700f9fffe23f8f3a475725df588
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:49:03 2011 -0400
_must_query
src/allmydata/mutable/servermap.py | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
commit
c599a059b8df3f5785e4bf89fb6ecc6d8dcd708b
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:48:05 2011 -0400
_queries_outstanding
src/allmydata/mutable/servermap.py | 16 +++++++---------
1 files changed, 7 insertions(+), 9 deletions(-)
commit
7743759f98ac2c07926b2fdbd80bf52dfab33085
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:46:17 2011 -0400
_empty_servers
src/allmydata/mutable/servermap.py | 5 ++---
1 files changed, 2 insertions(+), 3 deletions(-)
commit
6bb1825916828a713a32cdf7f7411fa3ea2e1e5d
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:45:39 2011 -0400
_good_servers
src/allmydata/mutable/servermap.py | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
commit
1768fab1b51d8dd93ecabbaaabfadfa20cf6c3d4
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:44:59 2011 -0400
_bad_servers
src/allmydata/mutable/servermap.py | 14 +++++++-------
1 files changed, 7 insertions(+), 7 deletions(-)
commit
dccbaef30f0ba714c746bf6d4a1a803c36e17b65
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:41:54 2011 -0400
API of _try_to_set_pubkey()
src/allmydata/mutable/servermap.py | 7 ++++---
1 files changed, 4 insertions(+), 3 deletions(-)
commit
0481ea70042ba3575f15eac7fd0780f8ece580cc
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:35:02 2011 -0400
API of notify_server_corruption()
src/allmydata/mutable/servermap.py | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
commit
bea9cba18fb3b9c11bb22f18356a263ecec7351e
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:34:09 2011 -0400
API of _got_signature_one_share()
src/allmydata/mutable/servermap.py | 9 +++++----
1 files changed, 5 insertions(+), 4 deletions(-)
commit
1520123583cf78650706e114b15bb5b0ac1f4a14
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:32:33 2011 -0400
API of _try_to_validate_privkey()
src/allmydata/mutable/servermap.py | 9 +++++----
1 files changed, 5 insertions(+), 4 deletions(-)
commit
938852c9c8519c7a078f58a9b1f4dd8ec8b6715e
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:31:48 2011 -0400
API and internals of _add_lease_failed()
src/allmydata/mutable/servermap.py | 8 ++++----
1 files changed, 4 insertions(+), 4 deletions(-)
commit
3843dba367e3c19e176a622ab853cb51d2472ddf
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:30:37 2011 -0400
API of _privkey_query_failed()
src/allmydata/mutable/servermap.py | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
commit
2219a710e1633cd57d0ca0786490de87b3e19ba7
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:29:43 2011 -0400
fix bug in call to _privkey_query_failed, unrelated to refactoring
src/allmydata/mutable/servermap.py | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
commit
ae615bec7d0d1b269710b6902797b12f9592ad62
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:27:17 2011 -0400
API of _got_corrupt_share()
src/allmydata/mutable/servermap.py | 17 +++++++++--------
1 files changed, 9 insertions(+), 8 deletions(-)
commit
cb51c95a6f4e077278157a77dab060c8c1ad7a81
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:23:16 2011 -0400
API of _got_results()
src/allmydata/mutable/servermap.py | 9 +++++----
1 files changed, 5 insertions(+), 4 deletions(-)
commit
bac9154fe0af18f226999a58ffc2362d8cf4b802
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:19:19 2011 -0400
API of _query_failed()
src/allmydata/mutable/servermap.py | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
commit
fdc29a8ca95d4b5c503e5382b9e5d4d02141ba12
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:17:20 2011 -0400
API of _do_read()
src/allmydata/mutable/servermap.py | 6 ++++--
1 files changed, 4 insertions(+), 2 deletions(-)
commit
e7e9e338f28d004aa4d423d11c65f1e271ac7322
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:20:21 2011 -0400
API of _do_query()
src/allmydata/mutable/servermap.py | 15 +++++++--------
1 files changed, 7 insertions(+), 8 deletions(-)
commit
330625b9dac4cdbe72a11464a893065b9aeed453
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 14:43:05 2011 -0400
next step: first batch of updates to ServermapUpdater
updates:
most method-local variables in update()
API of _build_initial_querylist()
API of _send_initial_requests()
.full_serverlist
.extra_servers
src/allmydata/mutable/servermap.py | 39 ++++++++++++++------------
1 files changed, 21 insertions(+), 18 deletions(-)
commit
4aadc584fa7dcb2daa86b048c81dee0049ba26d9
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 15:07:00 2011 -0400
internal change: index _bad_shares with IServer
src/allmydata/mutable/servermap.py | 20 ++++++++++----------
1 files changed, 10 insertions(+), 10 deletions(-)
commit
16d4e6fa82a9907dbdc92094213387c6a4164e41
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 18:20:47 2011 +0100
internal change: index _known_shares with IServer instead of serverid
callers are unchanged
src/allmydata/mutable/servermap.py | 42 +++++++++++++++----------
1 files changed, 25 insertions(+), 17 deletions(-)
commit
ceeb5f4938cc814a0c75d1b8f4018aed965c2176
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 18:11:43 2011 +0100
accessors and name cleanup for servermap.Servermap.last_update_mode/time
src/allmydata/mutable/filenode.py | 6 +++---
src/allmydata/mutable/publish.py | 4 ++--
src/allmydata/mutable/servermap.py | 17 +++++++++++------
3 files changed, 16 insertions(+), 11 deletions(-)
commit
8d3cbda82661c0a7e5c3d3b65cf7a5d5ab7e32c0
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 18:11:14 2011 +0100
accessors and name cleanup for servermap.Servermap.problems
src/allmydata/mutable/servermap.py | 21 +++++++++++++--------
src/allmydata/test/test_mutable.py | 6 +++---
2 files changed, 16 insertions(+), 11 deletions(-)
commit
348f57988f79389db0aab7672e6eaa9a6d8e3219
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 18:10:41 2011 +0100
accessors and name cleanup for servermap.Servermap.bad_shares
src/allmydata/mutable/publish.py | 2 +-
src/allmydata/mutable/servermap.py | 30 ++++++++++++++-----------
2 files changed, 18 insertions(+), 14 deletions(-)
commit
520c9368134673cdf76c653c5e1bb91c2ab5d51e
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 18:10:05 2011 +0100
accessors and name cleanup for servermap.Servermap.servermap .
src/allmydata/mutable/publish.py | 14 +++++----
src/allmydata/mutable/servermap.py | 38 ++++++++++++++-----------
2 files changed, 29 insertions(+), 23 deletions(-)
commit
b8b8dc38287a91dbdf494426ac801d9381ce5841
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 18:08:02 2011 +0100
fix reachable_servers
src/allmydata/mutable/checker.py | 3 ++-
src/allmydata/mutable/publish.py | 4 +++-
src/allmydata/mutable/servermap.py | 12 ++++++++++--
3 files changed, 15 insertions(+), 4 deletions(-)
commit
cb0cfd1adfefad357c187aaaf690c3df68b622bc
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 18:06:03 2011 +0100
fix Servermap.unreachable_servers
src/allmydata/mutable/servermap.py | 11 ++++++++---
1 files changed, 8 insertions(+), 3 deletions(-)
commit
2d9ea79b94bd4db674d40386fda90825785ac495
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 18:03:48 2011 +0100
give ServerMap a StorageFarmBroker, temporary
this makes it possible for the ServerMap to accept bare serverids and still
build data structures with IServers
src/allmydata/mutable/checker.py | 2 +-
src/allmydata/mutable/filenode.py | 2 +-
src/allmydata/mutable/publish.py | 2 +-
src/allmydata/mutable/servermap.py | 5 +++--
src/allmydata/test/test_mutable.py | 8 ++++----
5 files changed, 10 insertions(+), 9 deletions(-)
commit
718d1aeff6fded893f65397806d22ece928b0dd4
Author: Brian Warner <warner@lothar.com>
Date: Mon Oct 3 13:43:30 2011 -0400
add StorageFarmBroker.get_server_for_id(), temporary helper
This will go away once we're passing IServers everywhere.
src/allmydata/storage_client.py | 2 ++
src/allmydata/test/no_network.py | 13 +++++++++++++
2 files changed, 15 insertions(+), 0 deletions(-)
commit
ece20231d7fda0d503704842a4aa068dfbc2e54e
Author: Brian Warner <warner@lothar.com>
Date: Sun Oct 2 01:11:50 2011 +0100
add proper accessors for Servermap.connections, to make refactoring easier
src/allmydata/mutable/publish.py | 6 +++---
src/allmydata/mutable/retrieve.py | 10 +++++-----
src/allmydata/mutable/servermap.py | 17 +++++++++++------
3 files changed, 19 insertions(+), 14 deletions(-)
commit
3b943d6bf302ff702668081a612fc4fe2604cf9c
Author: Brian Warner <warner@lothar.com>
Date: Fri Sep 23 10:34:30 2011 -0700
mutable/servermap.py and neighbors: s/peer/server/
src/allmydata/mutable/checker.py | 22 +-
src/allmydata/mutable/publish.py | 204 +++++++-------
src/allmydata/mutable/servermap.py | 402 +++++++++++++-------------
src/allmydata/test/test_mutable.py | 18 +-
4 files changed, 323 insertions(+), 323 deletions(-)
# response_length (None until success)
self.block_requests = []
- self.known_shares = [] # (serverid, shnum)
+ self.known_shares = [] # (server, shnum)
self.problems = []
from allmydata.uri import from_string
-from allmydata.util import base32, idlib, log
+from allmydata.util import base32, log
from allmydata.check_results import CheckAndRepairResults, CheckResults
from allmydata.mutable.common import MODE_CHECK, CorruptShareError
self._storage_broker = storage_broker
self._history = history
self._monitor = monitor
- self.bad_shares = [] # list of (nodeid,shnum,failure)
+ self.bad_shares = [] # list of (server,shnum,failure)
self._storage_index = self._node.get_storage_index()
self.results = CheckResults(from_string(node.get_uri()), self._storage_index)
self.need_repair = False
if not self.best_version:
return
- r = Retrieve(self._node, servermap, self.best_version, verify=True)
+ r = Retrieve(self._node, self._storage_broker, servermap,
+ self.best_version, verify=True)
d = r.download()
d.addCallback(self._process_bad_shares)
return d
counters["count-shares-good"] = num_distinct_shares
counters["count-shares-needed"] = k
counters["count-shares-expected"] = N
- good_hosts = smap.all_peers_for_version(version)
+ good_hosts = smap.all_servers_for_version(version)
counters["count-good-share-hosts"] = len(good_hosts)
vmap = smap.make_versionmap()
counters["count-wrong-shares"] = sum([len(shares)
report.append("Unhealthy: best version has only %d shares "
"(encoding is %d-of-%d)" % (s, k, N))
summary.append("%d shares (enc %d-of-%d)" % (s, k, N))
- hosts = smap.all_peers_for_version(best_version)
+ hosts = smap.all_servers_for_version(best_version)
needs_rebalancing = bool( len(hosts) < N )
elif unrecoverable:
healthy = False
data["list-corrupt-shares"] = locators = []
report.append("Corrupt Shares:")
summary.append("Corrupt Shares:")
- for (peerid, shnum, f) in sorted(self.bad_shares):
- locators.append( (peerid, self._storage_index, shnum) )
- s = "%s-sh%d" % (idlib.shortnodeid_b2a(peerid), shnum)
+ for (server, shnum, f) in sorted(self.bad_shares):
+ serverid = server.get_serverid()
+ locators.append( (serverid, self._storage_index, shnum) )
+ s = "%s-sh%d" % (server.get_name(), shnum)
if f.check(CorruptShareError):
ft = f.value.reason
else:
ft = str(f)
report.append(" %s: %s" % (s, ft))
summary.append(s)
- p = (peerid, self._storage_index, shnum, f)
+ p = (serverid, self._storage_index, shnum, f)
r.problems.append(p)
msg = ("CorruptShareError during mutable verify, "
- "peerid=%(peerid)s, si=%(si)s, shnum=%(shnum)d, "
+ "serverid=%(serverid)s, si=%(si)s, shnum=%(shnum)d, "
"where=%(where)s")
- log.msg(format=msg, peerid=idlib.nodeid_b2a(peerid),
+ log.msg(format=msg, serverid=server.get_name(),
si=base32.b2a(self._storage_index),
shnum=shnum,
where=ft,
sharemap = {}
for verinfo in vmap:
- for (shnum, peerid, timestamp) in vmap[verinfo]:
+ for (shnum, server, timestamp) in vmap[verinfo]:
shareid = "%s-sh%d" % (smap.summarize_version(verinfo), shnum)
if shareid not in sharemap:
sharemap[shareid] = []
- sharemap[shareid].append(peerid)
+ sharemap[shareid].append(server.get_serverid())
data["sharemap"] = sharemap
- data["servers-responding"] = list(smap.reachable_peers)
+ data["servers-responding"] = [s.get_serverid() for s in
+ list(smap.get_reachable_servers())]
r.set_healthy(healthy)
r.set_recoverable(bool(recoverable))
-from allmydata.util import idlib
from allmydata.util.spans import DataSpans
MODE_CHECK = "MODE_CHECK" # query all peers
self.first_error = first_error
class CorruptShareError(Exception):
- def __init__(self, peerid, shnum, reason):
- self.args = (peerid, shnum, reason)
- self.peerid = peerid
+ def __init__(self, server, shnum, reason):
+ self.args = (server, shnum, reason)
+ self.server = server
self.shnum = shnum
self.reason = reason
def __str__(self):
- short_peerid = idlib.nodeid_b2a(self.peerid)[:8]
- return "<CorruptShareError peerid=%s shnum[%d]: %s" % (short_peerid,
- self.shnum,
- self.reason)
+ return "<CorruptShareError server=%s shnum[%d]: %s" % \
+ (self.server.get_name(), self.shnum, self.reason)
class UnknownVersionError(Exception):
"""The share we received was of a version we don't recognize."""
def _read_from_cache(self, verinfo, shnum, offset, length):
return self._cache.read(verinfo, shnum, offset, length)
- def get_write_enabler(self, peerid):
- assert len(peerid) == 20
- return hashutil.ssk_write_enabler_hash(self._writekey, peerid)
- def get_renewal_secret(self, peerid):
- assert len(peerid) == 20
+ def get_write_enabler(self, server):
+ seed = server.get_foolscap_write_enabler_seed()
+ assert len(seed) == 20
+ return hashutil.ssk_write_enabler_hash(self._writekey, seed)
+ def get_renewal_secret(self, server):
crs = self._secret_holder.get_renewal_secret()
frs = hashutil.file_renewal_secret_hash(crs, self._storage_index)
- return hashutil.bucket_renewal_secret_hash(frs, peerid)
- def get_cancel_secret(self, peerid):
- assert len(peerid) == 20
+ lease_seed = server.get_lease_seed()
+ assert len(lease_seed) == 20
+ return hashutil.bucket_renewal_secret_hash(frs, lease_seed)
+ def get_cancel_secret(self, server):
ccs = self._secret_holder.get_cancel_secret()
fcs = hashutil.file_cancel_secret_hash(ccs, self._storage_index)
- return hashutil.bucket_cancel_secret_hash(fcs, peerid)
+ lease_seed = server.get_lease_seed()
+ assert len(lease_seed) == 20
+ return hashutil.bucket_cancel_secret_hash(fcs, lease_seed)
def get_writekey(self):
return self._writekey
recoverable version that I can find in there.
"""
# XXX: wording ^^^^
- if servermap and servermap.last_update_mode == mode:
+ if servermap and servermap.get_last_update()[0] == mode:
d = defer.succeed(servermap)
else:
d = self._get_servermap(mode)
def _overwrite(self, new_contents):
assert IMutableUploadable.providedBy(new_contents)
- assert self._servermap.last_update_mode == MODE_WRITE
+ assert self._servermap.get_last_update()[0] == MODE_WRITE
return self._upload(new_contents)
I attempt to apply a modifier to the contents of the mutable
file.
"""
- assert self._servermap.last_update_mode != MODE_READ
+ assert self._servermap.get_last_update()[0] != MODE_READ
# download_to_data is serialized, so we have to call this to
# avoid deadlock.
"""
I am the serialized companion of read.
"""
- r = Retrieve(self._node, self._servermap, self._version, fetch_privkey)
+ r = Retrieve(self._node, self._storage_broker, self._servermap,
+ self._version, fetch_privkey)
if self._history:
self._history.notify_retrieve(r.get_status())
d = r.download(consumer, offset, size)
used by the new uploadable. I return a Deferred that fires with
the segments.
"""
- r = Retrieve(self._node, self._servermap, self._version)
+ r = Retrieve(self._node, self._storage_broker, self._servermap,
+ self._version)
# decode: takes in our blocks and salts from the servermap,
# returns a Deferred that fires with the corresponding plaintext
# segments. Does not download -- simply takes advantage of
from twisted.python import failure
from allmydata.interfaces import IPublishStatus, SDMF_VERSION, MDMF_VERSION, \
IMutableUploadable
-from allmydata.util import base32, hashutil, mathutil, idlib, log
+from allmydata.util import base32, hashutil, mathutil, log
from allmydata.util.dictutil import DictOfSets
from allmydata import hashtree, codec
from allmydata.storage.server import si_b2a
self.timings["encrypt"] = 0.0
self.timings["encode"] = 0.0
self.servermap = None
- self.problems = {}
+ self._problems = {}
self.active = True
self.storage_index = None
self.helper = False
self.counter = self.statusid_counter.next()
self.started = time.time()
- def add_per_server_time(self, peerid, elapsed):
- if peerid not in self.timings["send_per_server"]:
- self.timings["send_per_server"][peerid] = []
- self.timings["send_per_server"][peerid].append(elapsed)
+ def add_per_server_time(self, server, elapsed):
+ serverid = server.get_serverid()
+ if serverid not in self.timings["send_per_server"]:
+ self.timings["send_per_server"][serverid] = []
+ self.timings["send_per_server"][serverid].append(elapsed)
def accumulate_encode_time(self, elapsed):
self.timings["encode"] += elapsed
def accumulate_encrypt_time(self, elapsed):
return self.active
def get_counter(self):
return self.counter
+ def get_problems(self):
+ return self._problems
def set_storage_index(self, si):
self.storage_index = si
it on their own.
"""
# How this works:
- # 1: Make peer assignments. We'll assign each share that we know
- # about on the grid to that peer that currently holds that
+ # 1: Make server assignments. We'll assign each share that we know
+ # about on the grid to that server that currently holds that
# share, and will not place any new shares.
# 2: Setup encoding parameters. Most of these will stay the same
# -- datalength will change, as will some of the offsets.
# first, which servers will we publish to? We require that the
# servermap was updated in MODE_WRITE, so we can depend upon the
- # peerlist computed by that process instead of computing our own.
+ # serverlist computed by that process instead of computing our own.
assert self._servermap
- assert self._servermap.last_update_mode in (MODE_WRITE, MODE_CHECK)
+ assert self._servermap.get_last_update()[0] in (MODE_WRITE, MODE_CHECK)
# we will push a version that is one larger than anything present
# in the grid, according to the servermap.
self._new_seqnum = self._servermap.highest_seqnum() + 1
self._encprivkey = self._node.get_encprivkey()
sb = self._storage_broker
- full_peerlist = [(s.get_serverid(), s.get_rref())
- for s in sb.get_servers_for_psi(self._storage_index)]
- self.full_peerlist = full_peerlist # for use later, immutable
- self.bad_peers = set() # peerids who have errbacked/refused requests
+ full_serverlist = list(sb.get_servers_for_psi(self._storage_index))
+ self.full_serverlist = full_serverlist # for use later, immutable
+ self.bad_servers = set() # servers who have errbacked/refused requests
# This will set self.segment_size, self.num_segments, and
# self.fec. TODO: Does it know how to do the offset? Probably
# we keep track of three tables. The first is our goal: which share
# we want to see on which servers. This is initially populated by the
# existing servermap.
- self.goal = set() # pairs of (peerid, shnum) tuples
+ self.goal = set() # pairs of (server, shnum) tuples
# the number of outstanding queries: those that are in flight and
# may or may not be delivered, accepted, or acknowledged. This is
# the third is a table of successes: share which have actually been
# placed. These are populated when responses come back with success.
# When self.placed == self.goal, we're done.
- self.placed = set() # (peerid, shnum) tuples
-
- # we also keep a mapping from peerid to RemoteReference. Each time we
- # pull a connection out of the full peerlist, we add it to this for
- # use later.
- self.connections = {}
+ self.placed = set() # (server, shnum) tuples
self.bad_share_checkstrings = {}
# try to update each existing share in place. Since we're
# updating, we ignore damaged and missing shares -- callers must
# do a repair to repair and recreate these.
- for (peerid, shnum) in self._servermap.servermap:
- self.goal.add( (peerid, shnum) )
- self.connections[peerid] = self._servermap.connections[peerid]
+ self.goal = set(self._servermap.get_known_shares())
self.writers = {}
# SDMF files are updated differently.
self._version = MDMF_VERSION
writer_class = MDMFSlotWriteProxy
- # For each (peerid, shnum) in self.goal, we make a
- # write proxy for that peer. We'll use this to write
- # shares to the peer.
- for key in self.goal:
- peerid, shnum = key
- write_enabler = self._node.get_write_enabler(peerid)
- renew_secret = self._node.get_renewal_secret(peerid)
- cancel_secret = self._node.get_cancel_secret(peerid)
+ # For each (server, shnum) in self.goal, we make a
+ # write proxy for that server. We'll use this to write
+ # shares to the server.
+ for (server,shnum) in self.goal:
+ write_enabler = self._node.get_write_enabler(server)
+ renew_secret = self._node.get_renewal_secret(server)
+ cancel_secret = self._node.get_cancel_secret(server)
secrets = (write_enabler, renew_secret, cancel_secret)
self.writers[shnum] = writer_class(shnum,
- self.connections[peerid],
+ server.get_rref(),
self._storage_index,
secrets,
self._new_seqnum,
self.total_shares,
self.segment_size,
self.datalength)
- self.writers[shnum].peerid = peerid
- assert (peerid, shnum) in self._servermap.servermap
- old_versionid, old_timestamp = self._servermap.servermap[key]
+ self.writers[shnum].server = server
+ known_shares = self._servermap.get_known_shares()
+ assert (server, shnum) in known_shares
+ old_versionid, old_timestamp = known_shares[(server,shnum)]
(old_seqnum, old_root_hash, old_salt, old_segsize,
old_datalength, old_k, old_N, old_prefix,
old_offsets_tuple) = old_versionid
# first, which servers will we publish to? We require that the
# servermap was updated in MODE_WRITE, so we can depend upon the
- # peerlist computed by that process instead of computing our own.
+ # serverlist computed by that process instead of computing our own.
if self._servermap:
- assert self._servermap.last_update_mode in (MODE_WRITE, MODE_CHECK)
+ assert self._servermap.get_last_update()[0] in (MODE_WRITE, MODE_CHECK)
# we will push a version that is one larger than anything present
# in the grid, according to the servermap.
self._new_seqnum = self._servermap.highest_seqnum() + 1
self._encprivkey = self._node.get_encprivkey()
sb = self._storage_broker
- full_peerlist = [(s.get_serverid(), s.get_rref())
- for s in sb.get_servers_for_psi(self._storage_index)]
- self.full_peerlist = full_peerlist # for use later, immutable
- self.bad_peers = set() # peerids who have errbacked/refused requests
+ full_serverlist = list(sb.get_servers_for_psi(self._storage_index))
+ self.full_serverlist = full_serverlist # for use later, immutable
+ self.bad_servers = set() # servers who have errbacked/refused requests
# This will set self.segment_size, self.num_segments, and
# self.fec.
# we keep track of three tables. The first is our goal: which share
# we want to see on which servers. This is initially populated by the
# existing servermap.
- self.goal = set() # pairs of (peerid, shnum) tuples
+ self.goal = set() # pairs of (server, shnum) tuples
# the number of outstanding queries: those that are in flight and
# may or may not be delivered, accepted, or acknowledged. This is
# the third is a table of successes: share which have actually been
# placed. These are populated when responses come back with success.
# When self.placed == self.goal, we're done.
- self.placed = set() # (peerid, shnum) tuples
-
- # we also keep a mapping from peerid to RemoteReference. Each time we
- # pull a connection out of the full peerlist, we add it to this for
- # use later.
- self.connections = {}
+ self.placed = set() # (server, shnum) tuples
self.bad_share_checkstrings = {}
# we use the servermap to populate the initial goal: this way we will
# try to update each existing share in place.
- for (peerid, shnum) in self._servermap.servermap:
- self.goal.add( (peerid, shnum) )
- self.connections[peerid] = self._servermap.connections[peerid]
+ self.goal = set(self._servermap.get_known_shares())
+
# then we add in all the shares that were bad (corrupted, bad
# signatures, etc). We want to replace these.
- for key, old_checkstring in self._servermap.bad_shares.items():
- (peerid, shnum) = key
- self.goal.add(key)
- self.bad_share_checkstrings[key] = old_checkstring
- self.connections[peerid] = self._servermap.connections[peerid]
+ for key, old_checkstring in self._servermap.get_bad_shares().items():
+ (server, shnum) = key
+ self.goal.add( (server,shnum) )
+ self.bad_share_checkstrings[(server,shnum)] = old_checkstring
- # TODO: Make this part do peer selection.
+ # TODO: Make this part do server selection.
self.update_goal()
self.writers = {}
if self._version == MDMF_VERSION:
else:
writer_class = SDMFSlotWriteProxy
- # For each (peerid, shnum) in self.goal, we make a
- # write proxy for that peer. We'll use this to write
- # shares to the peer.
- for key in self.goal:
- peerid, shnum = key
- write_enabler = self._node.get_write_enabler(peerid)
- renew_secret = self._node.get_renewal_secret(peerid)
- cancel_secret = self._node.get_cancel_secret(peerid)
+ # For each (server, shnum) in self.goal, we make a
+ # write proxy for that server. We'll use this to write
+ # shares to the server.
+ for (server,shnum) in self.goal:
+ write_enabler = self._node.get_write_enabler(server)
+ renew_secret = self._node.get_renewal_secret(server)
+ cancel_secret = self._node.get_cancel_secret(server)
secrets = (write_enabler, renew_secret, cancel_secret)
self.writers[shnum] = writer_class(shnum,
- self.connections[peerid],
+ server.get_rref(),
self._storage_index,
secrets,
self._new_seqnum,
self.total_shares,
self.segment_size,
self.datalength)
- self.writers[shnum].peerid = peerid
- if (peerid, shnum) in self._servermap.servermap:
- old_versionid, old_timestamp = self._servermap.servermap[key]
+ self.writers[shnum].server = server
+ known_shares = self._servermap.get_known_shares()
+ if (server, shnum) in known_shares:
+ old_versionid, old_timestamp = known_shares[(server,shnum)]
(old_seqnum, old_root_hash, old_salt, old_segsize,
old_datalength, old_k, old_N, old_prefix,
old_offsets_tuple) = old_versionid
self.writers[shnum].set_checkstring(old_seqnum,
old_root_hash,
old_salt)
- elif (peerid, shnum) in self.bad_share_checkstrings:
- old_checkstring = self.bad_share_checkstrings[(peerid, shnum)]
+ elif (server, shnum) in self.bad_share_checkstrings:
+ old_checkstring = self.bad_share_checkstrings[(server, shnum)]
self.writers[shnum].set_checkstring(old_checkstring)
# Our remote shares will not have a complete checkstring until
def log_goal(self, goal, message=""):
logmsg = [message]
- for (shnum, peerid) in sorted([(s,p) for (p,s) in goal]):
- logmsg.append("sh%d to [%s]" % (shnum,
- idlib.shortnodeid_b2a(peerid)))
+ for (shnum, server) in sorted([(s,p) for (p,s) in goal]):
+ logmsg.append("sh%d to [%s]" % (shnum, server.get_name()))
self.log("current goal: %s" % (", ".join(logmsg)), level=log.NOISY)
self.log("we are planning to push new seqnum=#%d" % self._new_seqnum,
level=log.NOISY)
if True:
self.log_goal(self.goal, "before update: ")
- # first, remove any bad peers from our goal
- self.goal = set([ (peerid, shnum)
- for (peerid, shnum) in self.goal
- if peerid not in self.bad_peers ])
+ # first, remove any bad servers from our goal
+ self.goal = set([ (server, shnum)
+ for (server, shnum) in self.goal
+ if server not in self.bad_servers ])
# find the homeless shares:
- homefull_shares = set([shnum for (peerid, shnum) in self.goal])
+ homefull_shares = set([shnum for (server, shnum) in self.goal])
homeless_shares = set(range(self.total_shares)) - homefull_shares
homeless_shares = sorted(list(homeless_shares))
# place them somewhere. We prefer unused servers at the beginning of
- # the available peer list.
+ # the available server list.
if not homeless_shares:
return
# if an old share X is on a node, put the new share X there too.
- # TODO: 1: redistribute shares to achieve one-per-peer, by copying
- # shares from existing peers to new (less-crowded) ones. The
+ # TODO: 1: redistribute shares to achieve one-per-server, by copying
+ # shares from existing servers to new (less-crowded) ones. The
# old shares must still be updated.
# TODO: 2: move those shares instead of copying them, to reduce future
# update work
# this is a bit CPU intensive but easy to analyze. We create a sort
- # order for each peerid. If the peerid is marked as bad, we don't
+ # order for each server. If the server is marked as bad, we don't
# even put them in the list. Then we care about the number of shares
# which have already been assigned to them. After that we care about
# their permutation order.
old_assignments = DictOfSets()
- for (peerid, shnum) in self.goal:
- old_assignments.add(peerid, shnum)
+ for (server, shnum) in self.goal:
+ old_assignments.add(server, shnum)
- peerlist = []
- for i, (peerid, ss) in enumerate(self.full_peerlist):
- if peerid in self.bad_peers:
+ serverlist = []
+ for i, server in enumerate(self.full_serverlist):
+ serverid = server.get_serverid()
+ if server in self.bad_servers:
continue
- entry = (len(old_assignments.get(peerid, [])), i, peerid, ss)
- peerlist.append(entry)
- peerlist.sort()
+ entry = (len(old_assignments.get(server, [])), i, serverid, server)
+ serverlist.append(entry)
+ serverlist.sort()
- if not peerlist:
+ if not serverlist:
raise NotEnoughServersError("Ran out of non-bad servers, "
"first_error=%s" %
str(self._first_write_error),
self._first_write_error)
- # we then index this peerlist with an integer, because we may have to
- # wrap. We update the goal as we go.
+ # we then index this serverlist with an integer, because we may have
+ # to wrap. We update the goal as we go.
i = 0
for shnum in homeless_shares:
- (ignored1, ignored2, peerid, ss) = peerlist[i]
+ (ignored1, ignored2, ignored3, server) = serverlist[i]
# if we are forced to send a share to a server that already has
# one, we may have two write requests in flight, and the
# servermap (which was computed before either request was sent)
# this, otherwise it would cause the publish to fail with an
# UncoordinatedWriteError. See #546 for details of the trouble
# this used to cause.
- self.goal.add( (peerid, shnum) )
- self.connections[peerid] = ss
+ self.goal.add( (server, shnum) )
i += 1
- if i >= len(peerlist):
+ if i >= len(serverlist):
i = 0
if True:
self.log_goal(self.goal, "after update: ")
# bother checking it.
return
- peerid = writer.peerid
+ server = writer.server
lp = self.log("_got_write_answer from %s, share %d" %
- (idlib.shortnodeid_b2a(peerid), writer.shnum))
+ (server.get_name(), writer.shnum))
now = time.time()
elapsed = now - started
- self._status.add_per_server_time(peerid, elapsed)
+ self._status.add_per_server_time(server, elapsed)
wrote, read_data = answer
surprise_shares = set(read_data.keys()) - set([writer.shnum])
# We need to remove from surprise_shares any shares that we are
- # knowingly also writing to that peer from other writers.
+ # knowingly also writing to that server from other writers.
# TODO: Precompute this.
known_shnums = [x.shnum for x in self.writers.values()
- if x.peerid == peerid]
+ if x.server == server]
surprise_shares -= set(known_shnums)
self.log("found the following surprise shares: %s" %
str(surprise_shares))
if checkstring == self._checkstring:
# they have the right share, somehow
- if (peerid,shnum) in self.goal:
+ if (server,shnum) in self.goal:
# and we want them to have it, so we probably sent them a
# copy in an earlier write. This is ok, and avoids the
# #546 problem.
else:
# the new shares are of a different version
- if peerid in self._servermap.reachable_peers:
+ if server in self._servermap.get_reachable_servers():
# we asked them about their shares, so we had knowledge
# of what they used to have. Any surprising shares must
# have come from someone else, so UCW.
# mapupdate should have wokred harder and asked more
# servers before concluding that it knew about them all.
- # signal UCW, but make sure to ask this peer next time,
+ # signal UCW, but make sure to ask this server next time,
# so we'll remember to update it if/when we retry.
surprised = True
- # TODO: ask this peer next time. I don't yet have a good
+ # TODO: ask this server next time. I don't yet have a good
# way to do this. Two insufficient possibilities are:
#
- # self._servermap.add_new_share(peerid, shnum, verinfo, now)
+ # self._servermap.add_new_share(server, shnum, verinfo, now)
# but that requires fetching/validating/parsing the whole
# version string, and all we have is the checkstring
- # self._servermap.mark_bad_share(peerid, shnum, checkstring)
+ # self._servermap.mark_bad_share(server, shnum, checkstring)
# that will make publish overwrite the share next time,
# but it won't re-query the server, and it won't make
# mapupdate search further
# a way to tell these two apart (in fact, the storage server code
# doesn't have the option of refusing our share).
#
- # If the server is full, mark the peer as bad (so we don't ask
+ # If the server is full, mark the server as bad (so we don't ask
# them again), but don't set self.surprised. The loop() will find
# a new server.
#
# If the testv failed, log it, set self.surprised, but don't
- # bother adding to self.bad_peers .
+ # bother adding to self.bad_servers .
self.log("our testv failed, so the write did not happen",
parent=lp, level=log.WEIRD, umid="8sc26g")
self.surprised = True
- self.bad_peers.add(writer) # don't ask them again
+ self.bad_servers.add(server) # don't ask them again
# use the checkstring to add information to the log message
unknown_format = False
for (shnum,readv) in read_data.items():
other_IV) = unpack_sdmf_checkstring(checkstring)
else:
unknown_format = True
- expected_version = self._servermap.version_on_peer(peerid,
- shnum)
+ expected_version = self._servermap.version_on_server(server,
+ shnum)
if expected_version:
(seqnum, root_hash, IV, segsize, datalength, k, N, prefix,
offsets_tuple) = expected_version
(other_seqnum, other_roothash)
self.log(msg, parent=lp, level=log.NOISY)
# if expected_version==None, then we didn't expect to see a
- # share on that peer, and the 'surprise_shares' clause above
- # will have logged it.
+ # share on that server, and the 'surprise_shares' clause
+ # above will have logged it.
return
# and update the servermap
# shares, and can safely execute these statements.
if self.versioninfo:
self.log("wrote successfully: adding new share to servermap")
- self._servermap.add_new_share(peerid, writer.shnum,
+ self._servermap.add_new_share(server, writer.shnum,
self.versioninfo, started)
- self.placed.add( (peerid, writer.shnum) )
+ self.placed.add( (server, writer.shnum) )
self._update_status()
# the next method in the deferred chain will check to see if
# we're done and successful.
self.timings["decode"] = 0.0
self.timings["decrypt"] = 0.0
self.timings["cumulative_verify"] = 0.0
- self.problems = {}
+ self._problems = {}
self.active = True
self.storage_index = None
self.helper = False
return self.active
def get_counter(self):
return self.counter
-
- def add_fetch_timing(self, peerid, elapsed):
- if peerid not in self.timings["fetch_per_server"]:
- self.timings["fetch_per_server"][peerid] = []
- self.timings["fetch_per_server"][peerid].append(elapsed)
+ def get_problems(self):
+ return self._problems
+
+ def add_fetch_timing(self, server, elapsed):
+ serverid = server.get_serverid()
+ if serverid not in self.timings["fetch_per_server"]:
+ self.timings["fetch_per_server"][serverid] = []
+ self.timings["fetch_per_server"][serverid].append(elapsed)
def accumulate_decode_time(self, elapsed):
self.timings["decode"] += elapsed
def accumulate_decrypt_time(self, elapsed):
self.progress = value
def set_active(self, value):
self.active = value
+ def add_problem(self, server, f):
+ serverid = server.get_serverid()
+ self._problems[serverid] = f
class Marker:
pass
# will use a single ServerMap instance.
implements(IPushProducer)
- def __init__(self, filenode, servermap, verinfo, fetch_privkey=False,
- verify=False):
+ def __init__(self, filenode, storage_broker, servermap, verinfo,
+ fetch_privkey=False, verify=False):
self._node = filenode
assert self._node.get_pubkey()
+ self._storage_broker = storage_broker
self._storage_index = filenode.get_storage_index()
assert self._node.get_readkey()
self._last_failure = None
prefix = si_b2a(self._storage_index)[:5]
self._log_number = log.msg("Retrieve(%s): starting" % prefix)
- self._outstanding_queries = {} # maps (peerid,shnum) to start_time
self._running = True
self._decoding = False
self._bad_shares = set()
self.log("starting download")
self._started_fetching = time.time()
# The download process beyond this is a state machine.
- # _add_active_peers will select the peers that we want to use
+ # _add_active_servers will select the servers that we want to use
# for the download, and then attempt to start downloading. After
# each segment, it will check for doneness, reacting to broken
- # peers and corrupt shares as necessary. If it runs out of good
- # peers before downloading all of the segments, _done_deferred
+ # servers and corrupt shares as necessary. If it runs out of good
+ # servers before downloading all of the segments, _done_deferred
# will errback. Otherwise, it will eventually callback with the
# contents of the mutable file.
self.loop()
def loop(self):
d = fireEventually(None) # avoid #237 recursion limit problem
- d.addCallback(lambda ign: self._activate_enough_peers())
+ d.addCallback(lambda ign: self._activate_enough_servers())
d.addCallback(lambda ign: self._download_current_segment())
# when we're done, _download_current_segment will call _done. If we
# aren't, it will call loop() again.
shares = versionmap[self.verinfo]
# this sharemap is consumed as we decide to send requests
self.remaining_sharemap = DictOfSets()
- for (shnum, peerid, timestamp) in shares:
- self.remaining_sharemap.add(shnum, peerid)
+ for (shnum, server, timestamp) in shares:
+ self.remaining_sharemap.add(shnum, server)
# If the servermap update fetched anything, it fetched at least 1
# KiB, so we ask for that much.
# TODO: Change the cache methods to allow us to fetch all of the
# data that they have, then change this method to do that.
any_cache = self._node._read_from_cache(self.verinfo, shnum,
0, 1000)
- ss = self.servermap.connections[peerid]
- reader = MDMFSlotReadProxy(ss,
+ reader = MDMFSlotReadProxy(server.get_rref(),
self._storage_index,
shnum,
any_cache)
- reader.peerid = peerid
+ reader.server = server
self.readers[shnum] = reader
assert len(self.remaining_sharemap) >= k
self._current_segment = self._start_segment
- def _activate_enough_peers(self):
+ def _activate_enough_servers(self):
"""
I populate self._active_readers with enough active readers to
retrieve the contents of this mutable file. I am called before
"""
# TODO: It would be cool to investigate other heuristics for
# reader selection. For instance, the cost (in time the user
- # spends waiting for their file) of selecting a really slow peer
+ # spends waiting for their file) of selecting a really slow server
# that happens to have a primary share is probably more than
- # selecting a really fast peer that doesn't have a primary
+ # selecting a really fast server that doesn't have a primary
# share. Maybe the servermap could be extended to provide this
# information; it could keep track of latency information while
# it gathers more important data, and then this routine could
else:
new_shnums = []
- self.log("adding %d new peers to the active list" % len(new_shnums))
+ self.log("adding %d new servers to the active list" % len(new_shnums))
for shnum in new_shnums:
reader = self.readers[shnum]
self._active_readers.append(reader)
# segment decoding, then we'll take more drastic measures.
if self._need_privkey and not self._node.is_readonly():
d = reader.get_encprivkey()
- d.addCallback(self._try_to_validate_privkey, reader)
+ d.addCallback(self._try_to_validate_privkey, reader, reader.server)
# XXX: don't just drop the Deferred. We need error-reporting
# but not flow-control here.
assert len(self._active_readers) >= self._required_shares
def _remove_reader(self, reader):
"""
- At various points, we will wish to remove a peer from
+ At various points, we will wish to remove a server from
consideration and/or use. These include, but are not necessarily
limited to:
self._active_readers.remove(reader)
# TODO: self.readers.remove(reader)?
for shnum in list(self.remaining_sharemap.keys()):
- self.remaining_sharemap.discard(shnum, reader.peerid)
+ self.remaining_sharemap.discard(shnum, reader.server)
- def _mark_bad_share(self, reader, f):
+ def _mark_bad_share(self, server, shnum, reader, f):
"""
- I mark the (peerid, shnum) encapsulated by my reader argument as
- a bad share, which means that it will not be used anywhere else.
+ I mark the given (server, shnum) as a bad share, which means that it
+ will not be used anywhere else.
There are several reasons to want to mark something as a bad
share. These include:
- - A connection error to the peer.
+ - A connection error to the server.
- A mismatched prefix (that is, a prefix that does not match
our local conception of the version information string).
- A failing block hash, salt hash, share hash, or other
This method will ensure that readers that we wish to mark bad
(for these reasons or other reasons) are not used for the rest
of the download. Additionally, it will attempt to tell the
- remote peer (with no guarantee of success) that its share is
+ remote server (with no guarantee of success) that its share is
corrupt.
"""
self.log("marking share %d on server %s as bad" % \
- (reader.shnum, reader))
+ (shnum, server.get_name()))
prefix = self.verinfo[-2]
- self.servermap.mark_bad_share(reader.peerid,
- reader.shnum,
- prefix)
+ self.servermap.mark_bad_share(server, shnum, prefix)
self._remove_reader(reader)
- self._bad_shares.add((reader.peerid, reader.shnum, f))
- self._status.problems[reader.peerid] = f
+ self._bad_shares.add((server, shnum, f))
+ self._status.add_problem(server, f)
self._last_failure = f
- self.notify_server_corruption(reader.peerid, reader.shnum,
- str(f.value))
+ self.notify_server_corruption(server, shnum, str(f.value))
def _download_current_segment(self):
d = reader.get_block_and_salt(segnum)
d2 = self._get_needed_hashes(reader, segnum)
dl = defer.DeferredList([d, d2], consumeErrors=True)
- dl.addCallback(self._validate_block, segnum, reader, started)
+ dl.addCallback(self._validate_block, segnum, reader, reader.server, started)
dl.addErrback(self._validation_or_decoding_failed, [reader])
ds.append(dl)
dl = defer.DeferredList(ds)
I am called when a block or a salt fails to correctly validate, or when
the decryption or decoding operation fails for some reason. I react to
this failure by notifying the remote server of corruption, and then
- removing the remote peer from further activity.
+ removing the remote server from further activity.
"""
assert isinstance(readers, list)
bad_shnums = [reader.shnum for reader in readers]
- self.log("validation or decoding failed on share(s) %s, peer(s) %s "
+ self.log("validation or decoding failed on share(s) %s, server(s) %s "
", segment %d: %s" % \
(bad_shnums, readers, self._current_segment, str(f)))
for reader in readers:
- self._mark_bad_share(reader, f)
+ self._mark_bad_share(reader.server, reader.shnum, reader, f)
return
- def _validate_block(self, results, segnum, reader, started):
+ def _validate_block(self, results, segnum, reader, server, started):
"""
I validate a block from one share on a remote server.
"""
self.log("validating share %d for segment %d" % (reader.shnum,
segnum))
elapsed = time.time() - started
- self._status.add_fetch_timing(reader.peerid, elapsed)
+ self._status.add_fetch_timing(server, elapsed)
self._set_current_status("validating blocks")
# Did we fail to fetch either of the things that we were
# supposed to? Fail if so.
assert isinstance(results[0][1], failure.Failure)
f = results[0][1]
- raise CorruptShareError(reader.peerid,
+ raise CorruptShareError(server,
reader.shnum,
"Connection error: %s" % str(f))
bht.set_hashes(blockhashes)
except (hashtree.BadHashError, hashtree.NotEnoughHashesError, \
IndexError), e:
- raise CorruptShareError(reader.peerid,
+ raise CorruptShareError(server,
reader.shnum,
"block hash tree failure: %s" % e)
bht.set_hashes(leaves={segnum: blockhash})
except (hashtree.BadHashError, hashtree.NotEnoughHashesError, \
IndexError), e:
- raise CorruptShareError(reader.peerid,
+ raise CorruptShareError(server,
reader.shnum,
"block hash tree failure: %s" % e)
leaves={reader.shnum: bht[0]})
except (hashtree.BadHashError, hashtree.NotEnoughHashesError, \
IndexError), e:
- raise CorruptShareError(reader.peerid,
+ raise CorruptShareError(server,
reader.shnum,
"corrupt hashes: %s" % e)
return plaintext
- def notify_server_corruption(self, peerid, shnum, reason):
- ss = self.servermap.connections[peerid]
- ss.callRemoteOnly("advise_corrupt_share",
- "mutable", self._storage_index, shnum, reason)
+ def notify_server_corruption(self, server, shnum, reason):
+ rref = server.get_rref()
+ rref.callRemoteOnly("advise_corrupt_share",
+ "mutable", self._storage_index, shnum, reason)
- def _try_to_validate_privkey(self, enc_privkey, reader):
+ def _try_to_validate_privkey(self, enc_privkey, reader, server):
alleged_privkey_s = self._node._decrypt_privkey(enc_privkey)
alleged_writekey = hashutil.ssk_writekey_hash(alleged_privkey_s)
if alleged_writekey != self._node.get_writekey():
(reader, reader.shnum),
level=log.WEIRD, umid="YIw4tA")
if self._verify:
- self.servermap.mark_bad_share(reader.peerid, reader.shnum,
+ self.servermap.mark_bad_share(server, reader.shnum,
self.verinfo[-2])
- e = CorruptShareError(reader.peerid,
+ e = CorruptShareError(server,
reader.shnum,
"invalid privkey")
f = failure.Failure(e)
- self._bad_shares.add((reader.peerid, reader.shnum, f))
+ self._bad_shares.add((server, reader.shnum, f))
return
# it's good
self._node._populate_total_shares(N)
if self._verify:
- ret = list(self._bad_shares)
+ ret = self._bad_shares
self.log("done verifying, found %d bad shares" % len(ret))
else:
# TODO: upload status here?
def _raise_notenoughshareserror(self):
"""
- I am called by _activate_enough_peers when there are not enough
- active peers left to complete the download. After making some
+ I am called by _activate_enough_servers when there are not enough
+ active servers left to complete the download. After making some
useful logging statements, I throw an exception to that effect
to the caller of this Retrieve object through
self._done_deferred.
"""
- format = ("ran out of peers: "
+ format = ("ran out of servers: "
"have %(have)d of %(total)d segments "
"found %(bad)d bad shares "
"encoding %(k)d-of-%(n)d")
from twisted.python import failure
from foolscap.api import DeadReferenceError, RemoteException, eventually, \
fireEventually
-from allmydata.util import base32, hashutil, idlib, log, deferredutil
+from allmydata.util import base32, hashutil, log, deferredutil
from allmydata.util.dictutil import DictOfSets
from allmydata.storage.server import si_b2a
from allmydata.interfaces import IServermapUpdaterStatus
self.started = time.time()
self.finished = None
- def add_per_server_time(self, peerid, op, sent, elapsed):
+ def add_per_server_time(self, server, op, sent, elapsed):
+ serverid = server.get_serverid()
assert op in ("query", "late", "privkey")
- if peerid not in self.timings["per_server"]:
- self.timings["per_server"][peerid] = []
- self.timings["per_server"][peerid].append((op,sent,elapsed))
+ if serverid not in self.timings["per_server"]:
+ self.timings["per_server"][serverid] = []
+ self.timings["per_server"][serverid].append((op,sent,elapsed))
def get_started(self):
return self.started
self.storage_index = si
def set_mode(self, mode):
self.mode = mode
- def set_privkey_from(self, peerid):
- self.privkey_from = peerid
+ def set_privkey_from(self, server):
+ self.privkey_from = server.get_serverid()
def set_status(self, status):
self.status = status
def set_progress(self, value):
has changed since I last retrieved this data'. This reduces the chances
of clobbering a simultaneous (uncoordinated) write.
- @ivar servermap: a dictionary, mapping a (peerid, shnum) tuple to a
- (versionid, timestamp) tuple. Each 'versionid' is a
- tuple of (seqnum, root_hash, IV, segsize, datalength,
- k, N, signed_prefix, offsets)
-
- @ivar connections: maps peerid to a RemoteReference
-
- @ivar bad_shares: dict with keys of (peerid, shnum) tuples, describing
- shares that I should ignore (because a previous user of
- the servermap determined that they were invalid). The
- updater only locates a certain number of shares: if
- some of these turn out to have integrity problems and
- are unusable, the caller will need to mark those shares
- as bad, then re-update the servermap, then try again.
- The dict maps (peerid, shnum) tuple to old checkstring.
+ @var _known_shares: a dictionary, mapping a (server, shnum) tuple to a
+ (versionid, timestamp) tuple. Each 'versionid' is a
+ tuple of (seqnum, root_hash, IV, segsize, datalength,
+ k, N, signed_prefix, offsets)
+
+ @ivar _bad_shares: dict with keys of (server, shnum) tuples, describing
+ shares that I should ignore (because a previous user
+ of the servermap determined that they were invalid).
+ The updater only locates a certain number of shares:
+ if some of these turn out to have integrity problems
+ and are unusable, the caller will need to mark those
+ shares as bad, then re-update the servermap, then try
+ again. The dict maps (server, shnum) tuple to old
+ checkstring.
"""
def __init__(self):
- self.servermap = {}
- self.connections = {}
- self.unreachable_peers = set() # peerids that didn't respond to queries
- self.reachable_peers = set() # peerids that did respond to queries
- self.problems = [] # mostly for debugging
- self.bad_shares = {} # maps (peerid,shnum) to old checkstring
- self.last_update_mode = None
- self.last_update_time = 0
+ self._known_shares = {}
+ self.unreachable_servers = set() # servers that didn't respond to queries
+ self.reachable_servers = set() # servers that did respond to queries
+ self._problems = [] # mostly for debugging
+ self._bad_shares = {} # maps (server,shnum) to old checkstring
+ self._last_update_mode = None
+ self._last_update_time = 0
self.update_data = {} # (verinfo,shnum) => data
def copy(self):
s = ServerMap()
- s.servermap = self.servermap.copy() # tuple->tuple
- s.connections = self.connections.copy() # str->RemoteReference
- s.unreachable_peers = set(self.unreachable_peers)
- s.reachable_peers = set(self.reachable_peers)
- s.problems = self.problems[:]
- s.bad_shares = self.bad_shares.copy() # tuple->str
- s.last_update_mode = self.last_update_mode
- s.last_update_time = self.last_update_time
+ s._known_shares = self._known_shares.copy() # tuple->tuple
+ s.unreachable_servers = set(self.unreachable_servers)
+ s.reachable_servers = set(self.reachable_servers)
+ s._problems = self._problems[:]
+ s._bad_shares = self._bad_shares.copy() # tuple->str
+ s._last_update_mode = self._last_update_mode
+ s._last_update_time = self._last_update_time
return s
- def mark_bad_share(self, peerid, shnum, checkstring):
+ def get_reachable_servers(self):
+ return self.reachable_servers
+
+ def mark_server_reachable(self, server):
+ self.reachable_servers.add(server)
+
+ def mark_server_unreachable(self, server):
+ self.unreachable_servers.add(server)
+
+ def mark_bad_share(self, server, shnum, checkstring):
"""This share was found to be bad, either in the checkstring or
signature (detected during mapupdate), or deeper in the share
(detected at retrieve time). Remove it from our list of useful
corrupted or badly signed) so that a repair operation can do the
test-and-set using it as a reference.
"""
- key = (peerid, shnum) # record checkstring
- self.bad_shares[key] = checkstring
- self.servermap.pop(key, None)
+ key = (server, shnum) # record checkstring
+ self._bad_shares[key] = checkstring
+ self._known_shares.pop(key, None)
- def add_new_share(self, peerid, shnum, verinfo, timestamp):
+ def get_bad_shares(self):
+ # key=(server,shnum) -> checkstring
+ return self._bad_shares
+
+ def add_new_share(self, server, shnum, verinfo, timestamp):
"""We've written a new share out, replacing any that was there
before."""
- key = (peerid, shnum)
- self.bad_shares.pop(key, None)
- self.servermap[key] = (verinfo, timestamp)
+ key = (server, shnum)
+ self._bad_shares.pop(key, None)
+ self._known_shares[key] = (verinfo, timestamp)
+
+ def add_problem(self, f):
+ self._problems.append(f)
+ def get_problems(self):
+ return self._problems
+
+ def set_last_update(self, mode, when):
+ self._last_update_mode = mode
+ self._last_update_time = when
+ def get_last_update(self):
+ return (self._last_update_mode, self._last_update_time)
def dump(self, out=sys.stdout):
print >>out, "servermap:"
- for ( (peerid, shnum), (verinfo, timestamp) ) in self.servermap.items():
+ for ( (server, shnum), (verinfo, timestamp) ) in self._known_shares.items():
(seqnum, root_hash, IV, segsize, datalength, k, N, prefix,
offsets_tuple) = verinfo
print >>out, ("[%s]: sh#%d seq%d-%s %d-of-%d len%d" %
- (idlib.shortnodeid_b2a(peerid), shnum,
+ (server.get_name(), shnum,
seqnum, base32.b2a(root_hash)[:4], k, N,
datalength))
- if self.problems:
- print >>out, "%d PROBLEMS" % len(self.problems)
- for f in self.problems:
+ if self._problems:
+ print >>out, "%d PROBLEMS" % len(self._problems)
+ for f in self._problems:
print >>out, str(f)
return out
- def all_peers(self):
- return set([peerid
- for (peerid, shnum)
- in self.servermap])
+ def all_servers(self):
+ return set([server for (server, shnum) in self._known_shares])
- def all_peers_for_version(self, verinfo):
- """Return a set of peerids that hold shares for the given version."""
- return set([peerid
- for ( (peerid, shnum), (verinfo2, timestamp) )
- in self.servermap.items()
+ def all_servers_for_version(self, verinfo):
+ """Return a set of servers that hold shares for the given version."""
+ return set([server
+ for ( (server, shnum), (verinfo2, timestamp) )
+ in self._known_shares.items()
if verinfo == verinfo2])
+ def get_known_shares(self):
+ # maps (server,shnum) to (versionid,timestamp)
+ return self._known_shares
+
def make_sharemap(self):
- """Return a dict that maps shnum to a set of peerds that hold it."""
+ """Return a dict that maps shnum to a set of servers that hold it."""
sharemap = DictOfSets()
- for (peerid, shnum) in self.servermap:
- sharemap.add(shnum, peerid)
+ for (server, shnum) in self._known_shares:
+ sharemap.add(shnum, server)
return sharemap
def make_versionmap(self):
- """Return a dict that maps versionid to sets of (shnum, peerid,
+ """Return a dict that maps versionid to sets of (shnum, server,
timestamp) tuples."""
versionmap = DictOfSets()
- for ( (peerid, shnum), (verinfo, timestamp) ) in self.servermap.items():
- versionmap.add(verinfo, (shnum, peerid, timestamp))
+ for ( (server, shnum), (verinfo, timestamp) ) in self._known_shares.items():
+ versionmap.add(verinfo, (shnum, server, timestamp))
return versionmap
- def shares_on_peer(self, peerid):
- return set([shnum
- for (s_peerid, shnum)
- in self.servermap
- if s_peerid == peerid])
+ def debug_shares_on_server(self, server): # used by tests
+ return set([shnum for (s, shnum) in self._known_shares if s == server])
- def version_on_peer(self, peerid, shnum):
- key = (peerid, shnum)
- if key in self.servermap:
- (verinfo, timestamp) = self.servermap[key]
+ def version_on_server(self, server, shnum):
+ key = (server, shnum)
+ if key in self._known_shares:
+ (verinfo, timestamp) = self._known_shares[key]
return verinfo
return None
all_shares = {}
for verinfo, shares in versionmap.items():
s = set()
- for (shnum, peerid, timestamp) in shares:
+ for (shnum, server, timestamp) in shares:
s.add(shnum)
(seqnum, root_hash, IV, segsize, datalength, k, N, prefix,
offsets_tuple) = verinfo
bits = []
for (verinfo, shares) in versionmap.items():
vstr = self.summarize_version(verinfo)
- shnums = set([shnum for (shnum, peerid, timestamp) in shares])
+ shnums = set([shnum for (shnum, server, timestamp) in shares])
bits.append("%d*%s" % (len(shnums), vstr))
return "/".join(bits)
for (verinfo, shares) in versionmap.items():
(seqnum, root_hash, IV, segsize, datalength, k, N, prefix,
offsets_tuple) = verinfo
- shnums = set([shnum for (shnum, peerid, timestamp) in shares])
+ shnums = set([shnum for (shnum, server, timestamp) in shares])
if len(shnums) >= k:
# this one is recoverable
recoverable_versions.add(verinfo)
for (verinfo, shares) in versionmap.items():
(seqnum, root_hash, IV, segsize, datalength, k, N, prefix,
offsets_tuple) = verinfo
- shnums = set([shnum for (shnum, peerid, timestamp) in shares])
+ shnums = set([shnum for (shnum, server, timestamp) in shares])
if len(shnums) < k:
unrecoverable_versions.add(verinfo)
for (verinfo, shares) in versionmap.items():
(seqnum, root_hash, IV, segsize, datalength, k, N, prefix,
offsets_tuple) = verinfo
- shnums = set([shnum for (shnum, peerid, timestamp) in shares])
+ shnums = set([shnum for (shnum, server, timestamp) in shares])
healths[verinfo] = (len(shnums),k)
if len(shnums) < k:
unrecoverable.add(verinfo)
# avoid re-checking the signatures for each share.
self._valid_versions = set()
- # self.versionmap maps verinfo tuples to sets of (shnum, peerid,
- # timestamp) tuples. This is used to figure out which versions might
- # be retrievable, and to make the eventual data download faster.
- self.versionmap = DictOfSets()
-
self._done_deferred = defer.Deferred()
- # first, which peers should be talk to? Any that were in our old
+ # first, which servers should be talk to? Any that were in our old
# servermap, plus "enough" others.
self._queries_completed = 0
sb = self._storage_broker
- # All of the peers, permuted by the storage index, as usual.
- full_peerlist = [(s.get_serverid(), s.get_rref())
- for s in sb.get_servers_for_psi(self._storage_index)]
- self.full_peerlist = full_peerlist # for use later, immutable
- self.extra_peers = full_peerlist[:] # peers are removed as we use them
- self._good_peers = set() # peers who had some shares
- self._empty_peers = set() # peers who don't have any shares
- self._bad_peers = set() # peers to whom our queries failed
- self._readers = {} # peerid -> dict(sharewriters), filled in
- # after responses come in.
+ # All of the servers, permuted by the storage index, as usual.
+ full_serverlist = list(sb.get_servers_for_psi(self._storage_index))
+ self.full_serverlist = full_serverlist # for use later, immutable
+ self.extra_servers = full_serverlist[:] # servers are removed as we use them
+ self._good_servers = set() # servers who had some shares
+ self._empty_servers = set() # servers who don't have any shares
+ self._bad_servers = set() # servers to whom our queries failed
k = self._node.get_required_shares()
# For what cases can these conditions work?
if N is None:
N = 10
self.EPSILON = k
- # we want to send queries to at least this many peers (although we
+ # we want to send queries to at least this many servers (although we
# might not wait for all of their answers to come back)
- self.num_peers_to_query = k + self.EPSILON
+ self.num_servers_to_query = k + self.EPSILON
if self.mode == MODE_CHECK:
- # We want to query all of the peers.
- initial_peers_to_query = dict(full_peerlist)
- must_query = set(initial_peers_to_query.keys())
- self.extra_peers = []
+ # We want to query all of the servers.
+ initial_servers_to_query = list(full_serverlist)
+ must_query = set(initial_servers_to_query)
+ self.extra_servers = []
elif self.mode == MODE_WRITE:
# we're planning to replace all the shares, so we want a good
# chance of finding them all. We will keep searching until we've
# seen epsilon that don't have a share.
- # We don't query all of the peers because that could take a while.
- self.num_peers_to_query = N + self.EPSILON
- initial_peers_to_query, must_query = self._build_initial_querylist()
- self.required_num_empty_peers = self.EPSILON
+ # We don't query all of the servers because that could take a while.
+ self.num_servers_to_query = N + self.EPSILON
+ initial_servers_to_query, must_query = self._build_initial_querylist()
+ self.required_num_empty_servers = self.EPSILON
# TODO: arrange to read lots of data from k-ish servers, to avoid
# the extra round trip required to read large directories. This
# private key.
else: # MODE_READ, MODE_ANYTHING
- # 2k peers is good enough.
- initial_peers_to_query, must_query = self._build_initial_querylist()
-
- # this is a set of peers that we are required to get responses from:
- # they are peers who used to have a share, so we need to know where
- # they currently stand, even if that means we have to wait for a
- # silently-lost TCP connection to time out. We remove peers from this
- # set as we get responses.
- self._must_query = must_query
-
- # now initial_peers_to_query contains the peers that we should ask,
- # self.must_query contains the peers that we must have heard from
- # before we can consider ourselves finished, and self.extra_peers
- # contains the overflow (peers that we should tap if we don't get
- # enough responses)
+ # 2*k servers is good enough.
+ initial_servers_to_query, must_query = self._build_initial_querylist()
+
+ # this is a set of servers that we are required to get responses
+ # from: they are servers who used to have a share, so we need to know
+ # where they currently stand, even if that means we have to wait for
+ # a silently-lost TCP connection to time out. We remove servers from
+ # this set as we get responses.
+ self._must_query = set(must_query)
+
+ # now initial_servers_to_query contains the servers that we should
+ # ask, self.must_query contains the servers that we must have heard
+ # from before we can consider ourselves finished, and
+ # self.extra_servers contains the overflow (servers that we should
+ # tap if we don't get enough responses)
# I guess that self._must_query is a subset of
- # initial_peers_to_query?
- assert set(must_query).issubset(set(initial_peers_to_query))
+ # initial_servers_to_query?
+ assert must_query.issubset(initial_servers_to_query)
- self._send_initial_requests(initial_peers_to_query)
+ self._send_initial_requests(initial_servers_to_query)
self._status.timings["initial_queries"] = time.time() - self._started
return self._done_deferred
def _build_initial_querylist(self):
- initial_peers_to_query = {}
- must_query = set()
- for peerid in self._servermap.all_peers():
- ss = self._servermap.connections[peerid]
- # we send queries to everyone who was already in the sharemap
- initial_peers_to_query[peerid] = ss
- # and we must wait for responses from them
- must_query.add(peerid)
-
- while ((self.num_peers_to_query > len(initial_peers_to_query))
- and self.extra_peers):
- (peerid, ss) = self.extra_peers.pop(0)
- initial_peers_to_query[peerid] = ss
-
- return initial_peers_to_query, must_query
-
- def _send_initial_requests(self, peerlist):
- self._status.set_status("Sending %d initial queries" % len(peerlist))
+ # we send queries to everyone who was already in the sharemap
+ initial_servers_to_query = set(self._servermap.all_servers())
+ # and we must wait for responses from them
+ must_query = set(initial_servers_to_query)
+
+ while ((self.num_servers_to_query > len(initial_servers_to_query))
+ and self.extra_servers):
+ initial_servers_to_query.add(self.extra_servers.pop(0))
+
+ return initial_servers_to_query, must_query
+
+ def _send_initial_requests(self, serverlist):
+ self._status.set_status("Sending %d initial queries" % len(serverlist))
self._queries_outstanding = set()
- self._sharemap = DictOfSets() # shnum -> [(peerid, seqnum, R)..]
- for (peerid, ss) in peerlist.items():
- self._queries_outstanding.add(peerid)
- self._do_query(ss, peerid, self._storage_index, self._read_size)
+ for server in serverlist:
+ self._queries_outstanding.add(server)
+ self._do_query(server, self._storage_index, self._read_size)
- if not peerlist:
+ if not serverlist:
# there is nobody to ask, so we need to short-circuit the state
# machine.
d = defer.maybeDeferred(self._check_for_done, None)
# might produce a result.
return None
- def _do_query(self, ss, peerid, storage_index, readsize):
- self.log(format="sending query to [%(peerid)s], readsize=%(readsize)d",
- peerid=idlib.shortnodeid_b2a(peerid),
+ def _do_query(self, server, storage_index, readsize):
+ self.log(format="sending query to [%(name)s], readsize=%(readsize)d",
+ name=server.get_name(),
readsize=readsize,
level=log.NOISY)
- self._servermap.connections[peerid] = ss
started = time.time()
- self._queries_outstanding.add(peerid)
- d = self._do_read(ss, peerid, storage_index, [], [(0, readsize)])
- d.addCallback(self._got_results, peerid, readsize, (ss, storage_index),
+ self._queries_outstanding.add(server)
+ d = self._do_read(server, storage_index, [], [(0, readsize)])
+ d.addCallback(self._got_results, server, readsize, storage_index,
started)
- d.addErrback(self._query_failed, peerid)
+ d.addErrback(self._query_failed, server)
# errors that aren't handled by _query_failed (and errors caused by
# _query_failed) get logged, but we still want to check for doneness.
d.addErrback(log.err)
d.addCallback(self._check_for_done)
return d
- def _do_read(self, ss, peerid, storage_index, shnums, readv):
+ def _do_read(self, server, storage_index, shnums, readv):
+ ss = server.get_rref()
if self._add_lease:
# send an add-lease message in parallel. The results are handled
# separately. This is sent before the slot_readv() so that we can
# be sure the add_lease is retired by the time slot_readv comes
# back (this relies upon our knowledge that the server code for
# add_lease is synchronous).
- renew_secret = self._node.get_renewal_secret(peerid)
- cancel_secret = self._node.get_cancel_secret(peerid)
+ renew_secret = self._node.get_renewal_secret(server)
+ cancel_secret = self._node.get_cancel_secret(server)
d2 = ss.callRemote("add_lease", storage_index,
renew_secret, cancel_secret)
# we ignore success
- d2.addErrback(self._add_lease_failed, peerid, storage_index)
+ d2.addErrback(self._add_lease_failed, server, storage_index)
d = ss.callRemote("slot_readv", storage_index, shnums, readv)
return d
- def _got_corrupt_share(self, e, shnum, peerid, data, lp):
+ def _got_corrupt_share(self, e, shnum, server, data, lp):
"""
I am called when a remote server returns a corrupt share in
response to one of our queries. By corrupt, I mean a share
self.log(format="bad share: %(f_value)s", f_value=str(f),
failure=f, parent=lp, level=log.WEIRD, umid="h5llHg")
# Notify the server that its share is corrupt.
- self.notify_server_corruption(peerid, shnum, str(e))
- # By flagging this as a bad peer, we won't count any of
- # the other shares on that peer as valid, though if we
+ self.notify_server_corruption(server, shnum, str(e))
+ # By flagging this as a bad server, we won't count any of
+ # the other shares on that server as valid, though if we
# happen to find a valid version string amongst those
# shares, we'll keep track of it so that we don't need
# to validate the signature on those again.
- self._bad_peers.add(peerid)
+ self._bad_servers.add(server)
self._last_failure = f
# XXX: Use the reader for this?
checkstring = data[:SIGNED_PREFIX_LENGTH]
- self._servermap.mark_bad_share(peerid, shnum, checkstring)
- self._servermap.problems.append(f)
+ self._servermap.mark_bad_share(server, shnum, checkstring)
+ self._servermap.add_problem(f)
def _cache_good_sharedata(self, verinfo, shnum, now, data):
self._node._add_to_cache(verinfo, shnum, 0, data)
- def _got_results(self, datavs, peerid, readsize, stuff, started):
- lp = self.log(format="got result from [%(peerid)s], %(numshares)d shares",
- peerid=idlib.shortnodeid_b2a(peerid),
+ def _got_results(self, datavs, server, readsize, storage_index, started):
+ lp = self.log(format="got result from [%(name)s], %(numshares)d shares",
+ name=server.get_name(),
numshares=len(datavs))
+ ss = server.get_rref()
now = time.time()
elapsed = now - started
def _done_processing(ignored=None):
- self._queries_outstanding.discard(peerid)
- self._servermap.reachable_peers.add(peerid)
- self._must_query.discard(peerid)
+ self._queries_outstanding.discard(server)
+ self._servermap.mark_server_reachable(server)
+ self._must_query.discard(server)
self._queries_completed += 1
if not self._running:
self.log("but we're not running, so we'll ignore it", parent=lp)
_done_processing()
- self._status.add_per_server_time(peerid, "late", started, elapsed)
+ self._status.add_per_server_time(server, "late", started, elapsed)
return
- self._status.add_per_server_time(peerid, "query", started, elapsed)
+ self._status.add_per_server_time(server, "query", started, elapsed)
if datavs:
- self._good_peers.add(peerid)
+ self._good_servers.add(server)
else:
- self._empty_peers.add(peerid)
+ self._empty_servers.add(server)
- ss, storage_index = stuff
ds = []
for shnum,datav in datavs.items():
storage_index,
shnum,
data)
- self._readers.setdefault(peerid, dict())[shnum] = reader
# our goal, with each response, is to validate the version
# information and share data as best we can at this point --
# we do this by validating the signature. To do this, we
if not self._node.get_pubkey():
# fetch and set the public key.
d = reader.get_verification_key()
- d.addCallback(lambda results, shnum=shnum, peerid=peerid:
- self._try_to_set_pubkey(results, peerid, shnum, lp))
+ d.addCallback(lambda results, shnum=shnum:
+ self._try_to_set_pubkey(results, server, shnum, lp))
# XXX: Make self._pubkey_query_failed?
- d.addErrback(lambda error, shnum=shnum, peerid=peerid, data=data:
- self._got_corrupt_share(error, shnum, peerid, data, lp))
+ d.addErrback(lambda error, shnum=shnum, data=data:
+ self._got_corrupt_share(error, shnum, server, data, lp))
else:
# we already have the public key.
d = defer.succeed(None)
# bytes of the share on the storage server, so we
# shouldn't need to fetch anything at this step.
d2 = reader.get_verinfo()
- d2.addErrback(lambda error, shnum=shnum, peerid=peerid, data=data:
- self._got_corrupt_share(error, shnum, peerid, data, lp))
+ d2.addErrback(lambda error, shnum=shnum, data=data:
+ self._got_corrupt_share(error, shnum, server, data, lp))
# - Next, we need the signature. For an SDMF share, it is
# likely that we fetched this when doing our initial fetch
# to get the version information. In MDMF, this lives at
# the end of the share, so unless the file is quite small,
# we'll need to do a remote fetch to get it.
d3 = reader.get_signature()
- d3.addErrback(lambda error, shnum=shnum, peerid=peerid, data=data:
- self._got_corrupt_share(error, shnum, peerid, data, lp))
+ d3.addErrback(lambda error, shnum=shnum, data=data:
+ self._got_corrupt_share(error, shnum, server, data, lp))
# Once we have all three of these responses, we can move on
# to validating the signature
# fetch it here.
if self._need_privkey:
d4 = reader.get_encprivkey()
- d4.addCallback(lambda results, shnum=shnum, peerid=peerid:
- self._try_to_validate_privkey(results, peerid, shnum, lp))
- d4.addErrback(lambda error, shnum=shnum, peerid=peerid, data=data:
- self._privkey_query_failed(error, shnum, data, lp))
+ d4.addCallback(lambda results, shnum=shnum:
+ self._try_to_validate_privkey(results, server, shnum, lp))
+ d4.addErrback(lambda error, shnum=shnum:
+ self._privkey_query_failed(error, server, shnum, lp))
else:
d4 = defer.succeed(None)
dl = defer.DeferredList([d, d2, d3, d4, d5])
dl.addBoth(self._turn_barrier)
- dl.addCallback(lambda results, shnum=shnum, peerid=peerid:
- self._got_signature_one_share(results, shnum, peerid, lp))
+ dl.addCallback(lambda results, shnum=shnum:
+ self._got_signature_one_share(results, shnum, server, lp))
dl.addErrback(lambda error, shnum=shnum, data=data:
- self._got_corrupt_share(error, shnum, peerid, data, lp))
- dl.addCallback(lambda verinfo, shnum=shnum, peerid=peerid, data=data:
- self._cache_good_sharedata(verinfo, shnum, now, data))
+ self._got_corrupt_share(error, shnum, server, data, lp))
+ dl.addCallback(lambda verinfo, shnum=shnum, data=data:
+ self._cache_good_sharedata(verinfo, shnum, now, data))
ds.append(dl)
# dl is a deferred list that will fire when all of the shares
- # that we found on this peer are done processing. When dl fires,
+ # that we found on this server are done processing. When dl fires,
# we know that processing is done, so we can decrement the
# semaphore-like thing that we incremented earlier.
dl = defer.DeferredList(ds, fireOnOneErrback=True)
return fireEventually(result)
- def _try_to_set_pubkey(self, pubkey_s, peerid, shnum, lp):
+ def _try_to_set_pubkey(self, pubkey_s, server, shnum, lp):
if self._node.get_pubkey():
return # don't go through this again if we don't have to
fingerprint = hashutil.ssk_pubkey_fingerprint_hash(pubkey_s)
assert len(fingerprint) == 32
if fingerprint != self._node.get_fingerprint():
- raise CorruptShareError(peerid, shnum,
- "pubkey doesn't match fingerprint")
+ raise CorruptShareError(server, shnum,
+ "pubkey doesn't match fingerprint")
self._node._populate_pubkey(self._deserialize_pubkey(pubkey_s))
assert self._node.get_pubkey()
- def notify_server_corruption(self, peerid, shnum, reason):
- ss = self._servermap.connections[peerid]
- ss.callRemoteOnly("advise_corrupt_share",
- "mutable", self._storage_index, shnum, reason)
+ def notify_server_corruption(self, server, shnum, reason):
+ rref = server.get_rref()
+ rref.callRemoteOnly("advise_corrupt_share",
+ "mutable", self._storage_index, shnum, reason)
- def _got_signature_one_share(self, results, shnum, peerid, lp):
+ def _got_signature_one_share(self, results, shnum, server, lp):
# It is our job to give versioninfo to our caller. We need to
# raise CorruptShareError if the share is corrupt for any
# reason, something that our caller will handle.
- self.log(format="_got_results: got shnum #%(shnum)d from peerid %(peerid)s",
+ self.log(format="_got_results: got shnum #%(shnum)d from serverid %(name)s",
shnum=shnum,
- peerid=idlib.shortnodeid_b2a(peerid),
+ name=server.get_name(),
level=log.NOISY,
parent=lp)
if not self._running:
assert self._node.get_pubkey()
valid = self._node.get_pubkey().verify(prefix, signature[1])
if not valid:
- raise CorruptShareError(peerid, shnum,
+ raise CorruptShareError(server, shnum,
"signature is invalid")
# ok, it's a valid verinfo. Add it to the list of validated
# versions.
self.log(" found valid version %d-%s from %s-sh%d: %d-%d/%d/%d"
% (seqnum, base32.b2a(root_hash)[:4],
- idlib.shortnodeid_b2a(peerid), shnum,
+ server.get_name(), shnum,
k, n, segsize, datalen),
parent=lp)
self._valid_versions.add(verinfo)
# version info again, that its signature checks out and that
# we're okay to skip the signature-checking step.
- # (peerid, shnum) are bound in the method invocation.
- if (peerid, shnum) in self._servermap.bad_shares:
+ # (server, shnum) are bound in the method invocation.
+ if (server, shnum) in self._servermap.get_bad_shares():
# we've been told that the rest of the data in this share is
# unusable, so don't add it to the servermap.
self.log("but we've been told this is a bad share",
# Add the info to our servermap.
timestamp = time.time()
- self._servermap.add_new_share(peerid, shnum, verinfo, timestamp)
- # and the versionmap
- self.versionmap.add(verinfo, (shnum, peerid, timestamp))
+ self._servermap.add_new_share(server, shnum, verinfo, timestamp)
return verinfo
return verifier
- def _try_to_validate_privkey(self, enc_privkey, peerid, shnum, lp):
+ def _try_to_validate_privkey(self, enc_privkey, server, shnum, lp):
"""
Given a writekey from a remote server, I validate it against the
writekey stored in my node. If it is valid, then I set the
alleged_writekey = hashutil.ssk_writekey_hash(alleged_privkey_s)
if alleged_writekey != self._node.get_writekey():
self.log("invalid privkey from %s shnum %d" %
- (idlib.nodeid_b2a(peerid)[:8], shnum),
+ (server.get_name(), shnum),
parent=lp, level=log.WEIRD, umid="aJVccw")
return
# it's good
- self.log("got valid privkey from shnum %d on peerid %s" %
- (shnum, idlib.shortnodeid_b2a(peerid)),
+ self.log("got valid privkey from shnum %d on serverid %s" %
+ (shnum, server.get_name()),
parent=lp)
privkey = rsa.create_signing_key_from_string(alleged_privkey_s)
self._node._populate_encprivkey(enc_privkey)
self._node._populate_privkey(privkey)
self._need_privkey = False
- self._status.set_privkey_from(peerid)
+ self._status.set_privkey_from(server)
- def _add_lease_failed(self, f, peerid, storage_index):
+ def _add_lease_failed(self, f, server, storage_index):
# Older versions of Tahoe didn't handle the add-lease message very
# well: <=1.1.0 throws a NameError because it doesn't implement
# remote_add_lease(), 1.2.0/1.3.0 throw IndexError on unknown buckets
# this may ignore a bit too much, but that only hurts us
# during debugging
return
- self.log(format="error in add_lease from [%(peerid)s]: %(f_value)s",
- peerid=idlib.shortnodeid_b2a(peerid),
+ self.log(format="error in add_lease from [%(name)s]: %(f_value)s",
+ name=server.get_name(),
f_value=str(f.value),
failure=f,
level=log.WEIRD, umid="iqg3mw")
return
# local errors are cause for alarm
log.err(f,
- format="local error in add_lease to [%(peerid)s]: %(f_value)s",
- peerid=idlib.shortnodeid_b2a(peerid),
+ format="local error in add_lease to [%(name)s]: %(f_value)s",
+ name=server.get_name(),
f_value=str(f.value),
level=log.WEIRD, umid="ZWh6HA")
- def _query_failed(self, f, peerid):
+ def _query_failed(self, f, server):
if not self._running:
return
level = log.WEIRD
self.log(format="error during query: %(f_value)s",
f_value=str(f.value), failure=f,
level=level, umid="IHXuQg")
- self._must_query.discard(peerid)
- self._queries_outstanding.discard(peerid)
- self._bad_peers.add(peerid)
- self._servermap.problems.append(f)
- # a peerid could be in both ServerMap.reachable_peers and
- # .unreachable_peers if they responded to our query, but then an
+ self._must_query.discard(server)
+ self._queries_outstanding.discard(server)
+ self._bad_servers.add(server)
+ self._servermap.add_problem(f)
+ # a server could be in both ServerMap.reachable_servers and
+ # .unreachable_servers if they responded to our query, but then an
# exception was raised in _got_results.
- self._servermap.unreachable_peers.add(peerid)
+ self._servermap.mark_server_unreachable(server)
self._queries_completed += 1
self._last_failure = f
- def _privkey_query_failed(self, f, peerid, shnum, lp):
- self._queries_outstanding.discard(peerid)
+ def _privkey_query_failed(self, f, server, shnum, lp):
+ self._queries_outstanding.discard(server)
if not self._running:
return
level = log.WEIRD
self.log(format="error during privkey query: %(f_value)s",
f_value=str(f.value), failure=f,
parent=lp, level=level, umid="McoJ5w")
- self._servermap.problems.append(f)
+ self._servermap.add_problem(f)
self._last_failure = f
# return : keep waiting, no new queries
lp = self.log(format=("_check_for_done, mode is '%(mode)s', "
"%(outstanding)d queries outstanding, "
- "%(extra)d extra peers available, "
- "%(must)d 'must query' peers left, "
+ "%(extra)d extra servers available, "
+ "%(must)d 'must query' servers left, "
"need_privkey=%(need_privkey)s"
),
mode=self.mode,
outstanding=len(self._queries_outstanding),
- extra=len(self.extra_peers),
+ extra=len(self.extra_servers),
must=len(self._must_query),
need_privkey=self._need_privkey,
level=log.NOISY,
return
if self._must_query:
- # we are still waiting for responses from peers that used to have
+ # we are still waiting for responses from servers that used to have
# a share, so we must continue to wait. No additional queries are
# required at this time.
- self.log("%d 'must query' peers left" % len(self._must_query),
+ self.log("%d 'must query' servers left" % len(self._must_query),
level=log.NOISY, parent=lp)
return
- if (not self._queries_outstanding and not self.extra_peers):
- # all queries have retired, and we have no peers left to ask. No
+ if (not self._queries_outstanding and not self.extra_servers):
+ # all queries have retired, and we have no servers left to ask. No
# more progress can be made, therefore we are done.
- self.log("all queries are retired, no extra peers: done",
+ self.log("all queries are retired, no extra servers: done",
parent=lp)
return self._done()
# version, and we haven't seen any unrecoverable higher-seqnum'ed
# versions, then we're done.
- if self._queries_completed < self.num_peers_to_query:
+ if self._queries_completed < self.num_servers_to_query:
self.log(format="%(completed)d completed, %(query)d to query: need more",
completed=self._queries_completed,
- query=self.num_peers_to_query,
+ query=self.num_servers_to_query,
level=log.NOISY, parent=lp)
return self._send_more_queries(MAX_IN_FLIGHT)
if not recoverable_versions:
states = []
found_boundary = False
- for i,(peerid,ss) in enumerate(self.full_peerlist):
- if peerid in self._bad_peers:
+ for i,server in enumerate(self.full_serverlist):
+ if server in self._bad_servers:
# query failed
states.append("x")
- #self.log("loop [%s]: x" % idlib.shortnodeid_b2a(peerid))
- elif peerid in self._empty_peers:
+ #self.log("loop [%s]: x" % server.get_name()
+ elif server in self._empty_servers:
# no shares
states.append("0")
- #self.log("loop [%s]: 0" % idlib.shortnodeid_b2a(peerid))
+ #self.log("loop [%s]: 0" % server.get_name()
if last_found != -1:
num_not_found += 1
if num_not_found >= self.EPSILON:
found_boundary = True
break
- elif peerid in self._good_peers:
+ elif server in self._good_servers:
# yes shares
states.append("1")
- #self.log("loop [%s]: 1" % idlib.shortnodeid_b2a(peerid))
+ #self.log("loop [%s]: 1" % server.get_name()
last_found = i
num_not_found = 0
else:
# not responded yet
states.append("?")
- #self.log("loop [%s]: ?" % idlib.shortnodeid_b2a(peerid))
+ #self.log("loop [%s]: ?" % server.get_name()
last_not_responded = i
num_not_responded += 1
return self._send_more_queries(num_not_responded)
# if we hit here, we didn't find our boundary, so we're still
- # waiting for peers
+ # waiting for servers
self.log("no boundary yet, %s" % "".join(states), parent=lp,
level=log.NOISY)
return self._send_more_queries(MAX_IN_FLIGHT)
active_queries = len(self._queries_outstanding) + len(more_queries)
if active_queries >= num_outstanding:
break
- if not self.extra_peers:
+ if not self.extra_servers:
break
- more_queries.append(self.extra_peers.pop(0))
+ more_queries.append(self.extra_servers.pop(0))
self.log(format="sending %(more)d more queries: %(who)s",
more=len(more_queries),
- who=" ".join(["[%s]" % idlib.shortnodeid_b2a(peerid)
- for (peerid,ss) in more_queries]),
+ who=" ".join(["[%s]" % s.get_name() for s in more_queries]),
level=log.NOISY)
- for (peerid, ss) in more_queries:
- self._do_query(ss, peerid, self._storage_index, self._read_size)
+ for server in more_queries:
+ self._do_query(server, self._storage_index, self._read_size)
# we'll retrigger when those queries come back
def _done(self):
self._status.set_status("Finished")
self._status.set_active(False)
- self._servermap.last_update_mode = self.mode
- self._servermap.last_update_time = self._started
+ self._servermap.set_last_update(self.mode, self._started)
# the servermap will not be touched after this
self.log("servermap: %s" % self._servermap.summarize_versions())
return self.servers[serverid].get_nickname()
return None
-
class IServer(Interface):
"""I live in the client, and represent a single server."""
def start_connecting(tub, trigger_cb):
MDMF_VERSION
from allmydata.check_results import CheckResults, CheckAndRepairResults, \
DeepCheckResults, DeepCheckAndRepairResults
-from allmydata.mutable.common import CorruptShareError
from allmydata.mutable.layout import unpack_header
from allmydata.mutable.publish import MutableData
from allmydata.storage.mutable import MutableShareFile
r.set_recoverable(True)
data["count-shares-good"] = 9
data["list-corrupt-shares"] = [(nodeid, self.storage_index, 0)]
- r.problems = failure.Failure(CorruptShareError(is_bad))
+ # XXX: this whole 'is_bad' clause is unused. When a test is added
+ # to take advantage of it, we must find a way to provide 'server'
+ # (and IServer instance) to the CorruptShareError
+ #r.problems = failure.Failure(CorruptShareError(server, 0, is_bad))
else:
r.set_healthy(True)
r.set_recoverable(True)
r.set_healthy(False)
r.set_recoverable(True)
data["count-shares-good"] = 9
- r.problems = failure.Failure(CorruptShareError("peerid",
- 0, # shnum
- is_bad))
+ # XXX: this whole 'is_bad' clause is unused. When a test is added
+ # to take advantage of it, we must find a way to provide 'server'
+ # (and IServer instance) to the CorruptShareError
+ #r.problems = failure.Failure(CorruptShareError(server, 0, is_bad))
else:
r.set_healthy(True)
r.set_recoverable(True)
return self.serverid
def get_lease_seed(self):
return self.serverid
+ def get_foolscap_write_enabler_seed(self):
+ return self.serverid
+
def get_name(self):
return idlib.shortnodeid_b2a(self.serverid)
def get_longname(self):
self.failUnlessEqual(sm.recoverable_versions(), set([best]))
self.failUnlessEqual(len(sm.shares_available()), 1)
self.failUnlessEqual(sm.shares_available()[best], (num_shares, 3, 10))
- shnum, peerids = sm.make_sharemap().items()[0]
- peerid = list(peerids)[0]
- self.failUnlessEqual(sm.version_on_peer(peerid, shnum), best)
- self.failUnlessEqual(sm.version_on_peer(peerid, 666), None)
+ shnum, servers = sm.make_sharemap().items()[0]
+ server = list(servers)[0]
+ self.failUnlessEqual(sm.version_on_server(server, shnum), best)
+ self.failUnlessEqual(sm.version_on_server(server, 666), None)
return sm
def test_basic(self):
# mark the first 5 shares as corrupt, then update the servermap.
# The map should not have the marked shares it in any more, and
# new shares should be found to replace the missing ones.
- for (shnum, peerid, timestamp) in shares:
+ for (shnum, server, timestamp) in shares:
if shnum < 5:
- self._corrupted.add( (peerid, shnum) )
- sm.mark_bad_share(peerid, shnum, "")
+ self._corrupted.add( (server, shnum) )
+ sm.mark_bad_share(server, shnum, "")
return self.update_servermap(sm, MODE_WRITE)
d.addCallback(_made_map)
def _check_map(sm):
v = sm.best_recoverable_version()
vm = sm.make_versionmap()
shares = list(vm[v])
- for (peerid, shnum) in self._corrupted:
- peer_shares = sm.shares_on_peer(peerid)
- self.failIf(shnum in peer_shares,
- "%d was in %s" % (shnum, peer_shares))
+ for (server, shnum) in self._corrupted:
+ server_shares = sm.debug_shares_on_server(server)
+ self.failIf(shnum in server_shares,
+ "%d was in %s" % (shnum, server_shares))
self.failUnlessEqual(len(shares), 5)
d.addCallback(_check_map)
return d
def do_download(self, servermap, version=None):
if version is None:
version = servermap.best_recoverable_version()
- r = Retrieve(self._fn, servermap, version)
+ r = Retrieve(self._fn, self._storage_broker, servermap, version)
c = consumer.MemoryConsumer()
d = r.download(consumer=c)
d.addCallback(lambda mc: "".join(mc.chunks))
shares.clear()
d1 = self.shouldFail(NotEnoughSharesError,
"test_all_shares_vanished",
- "ran out of peers",
+ "ran out of servers",
self.do_download, servermap)
return d1
d.addCallback(_remove_shares)
self.failUnlessEqual(servermap.best_recoverable_version(), None)
self.failIf(servermap.recoverable_versions())
self.failIf(servermap.unrecoverable_versions())
- self.failIf(servermap.all_peers())
+ self.failIf(servermap.all_servers())
d.addCallback(_check_servermap)
return d
# no recoverable versions == not succeeding. The problem
# should be noted in the servermap's list of problems.
if substring:
- allproblems = [str(f) for f in servermap.problems]
+ allproblems = [str(f) for f in servermap.get_problems()]
self.failUnlessIn(substring, "".join(allproblems))
return servermap
if should_succeed:
f = res[0]
self.failUnless(f.check(NotEnoughSharesError))
self.failUnless("uncoordinated write" in str(f))
- return self._test_corrupt_all(1, "ran out of peers",
+ return self._test_corrupt_all(1, "ran out of servers",
corrupt_early=False,
failure_checker=_check)
shnums_to_corrupt=range(0, N-k))
d.addCallback(lambda res: self.make_servermap())
def _do_retrieve(servermap):
- self.failUnless(servermap.problems)
+ self.failUnless(servermap.get_problems())
self.failUnless("pubkey doesn't match fingerprint"
- in str(servermap.problems[0]))
+ in str(servermap.get_problems()[0]))
ver = servermap.best_recoverable_version()
- r = Retrieve(self._fn, servermap, ver)
+ r = Retrieve(self._fn, self._storage_broker, servermap, ver)
c = consumer.MemoryConsumer()
return r.download(c)
d.addCallback(_do_retrieve)
d.addCallback(lambda res:
self.shouldFail(NotEnoughSharesError,
"test_retrieve_surprise",
- "ran out of peers: have 0 of 1",
+ "ran out of servers: have 0 of 1",
n.download_version,
self.old_map,
self.old_map.best_recoverable_version(),
# stash the old state of the file
self.old_map = smap
# now shut down one of the servers
- peer0 = list(smap.make_sharemap()[0])[0]
+ peer0 = list(smap.make_sharemap()[0])[0].get_serverid()
self.g.remove_server(peer0)
# then modify the file, leaving the old map untouched
log.msg("starting winning write")
return ctx.tag["Encoding: %s of %s" % (k, n)]
def render_problems(self, ctx, data):
- problems = data.problems
+ problems = data.get_problems()
if not problems:
return ""
l = T.ul()
sharemap = servermap.make_sharemap()
for shnum in sorted(sharemap.keys()):
l[T.li["%d -> Placed on " % shnum,
- ", ".join(["[%s]" % idlib.shortnodeid_b2a(peerid)
- for peerid in sharemap[shnum]])]]
+ ", ".join(["[%s]" % server.get_name()
+ for server in sharemap[shnum]])]]
return ctx.tag["Sharemap:", l]
def render_problems(self, ctx, data):
- problems = data.problems
+ problems = data.get_problems()
if not problems:
return ""
l = T.ul()
+ # XXX: is this exercised? I don't think PublishStatus.problems is
+ # ever populated
for peerid in sorted(problems.keys()):
peerid_s = idlib.shortnodeid_b2a(peerid)
l[T.li["[%s]: %s" % (peerid_s, problems[peerid])]]