From: Brian Warner Date: Tue, 8 Jan 2008 00:09:16 +0000 (-0700) Subject: mutable: improve logging: mark events with level=log.WEIRD and log.UNUSUAL X-Git-Tag: allmydata-tahoe-0.7.0~16 X-Git-Url: https://git.rkrishnan.org/%5B%5E?a=commitdiff_plain;h=083ab7252c8730942953fcb5b5e0146b32bc4880;p=tahoe-lafs%2Ftahoe-lafs.git mutable: improve logging: mark events with level=log.WEIRD and log.UNUSUAL --- diff --git a/src/allmydata/mutable.py b/src/allmydata/mutable.py index 200d1ed3..37d45980 100644 --- a/src/allmydata/mutable.py +++ b/src/allmydata/mutable.py @@ -210,10 +210,10 @@ class Retrieve: num = self._node._client.log("Retrieve(%s): starting" % prefix) self._log_number = num - def log(self, msg): + def log(self, msg, **kwargs): prefix = self._log_prefix num = self._node._client.log("Retrieve(%s): %s" % (prefix, msg), - parent=self._log_number) + parent=self._log_number, **kwargs) return num def log_err(self, f): @@ -383,7 +383,7 @@ class Retrieve: except CorruptShareError, e: # log it and give the other shares a chance to be processed f = failure.Failure() - self.log("WEIRD: bad share: %s %s" % (f, f.value)) + self.log("bad share: %s %s" % (f, f.value), level=log.WEIRD) self._bad_peerids.add(peerid) self._last_failure = f pass @@ -446,7 +446,7 @@ class Retrieve: self._used_peers.add(peerid) self._last_failure = f self._bad_peerids.add(peerid) - self.log("WEIRD: error during query: %s %s" % (f, f.value)) + self.log("error during query: %s %s" % (f, f.value), level=log.WEIRD) def _check_for_done(self, res): if not self._running: @@ -476,7 +476,8 @@ class Retrieve: def _problem(f): self._last_failure = f if f.check(CorruptShareError): - self.log("WEIRD: saw corrupt share, rescheduling") + self.log("saw corrupt share, rescheduling", + level=log.WEIRD) # _attempt_decode is responsible for removing the bad # share, so we can just try again eventually(self._check_for_done, None) @@ -502,15 +503,14 @@ class Retrieve: # no more queries are outstanding. Can we send out more? First, # should we be looking at more peers? - self.log("UNUSUAL: need more peers: " - "N=%s, peerlist=%d peerlist_limit=%d" % + self.log("need more peers: N=%s, peerlist=%d peerlist_limit=%d" % (self._total_shares, len(self._peerlist), - self._peerlist_limit)) + self._peerlist_limit), level=log.UNUSUAL) if self._total_shares is not None: search_distance = self._total_shares * 2 else: search_distance = 20 - self.log("UNUSUAL: search_distance=%d" % search_distance) + self.log("search_distance=%d" % search_distance, level=log.UNUSUAL) if self._peerlist_limit < search_distance: # we might be able to get some more peers from the list peers = self._node._client.get_permuted_peers(self._storage_index, @@ -518,8 +518,9 @@ class Retrieve: self._peerlist = [(p[1],p[2]) for p in islice(peers, search_distance)] self._peerlist_limit = search_distance - self.log("UNUSUAL: added peers, peerlist=%d, peerlist_limit=%d" - % (len(self._peerlist), self._peerlist_limit)) + self.log("added peers, peerlist=%d, peerlist_limit=%d" + % (len(self._peerlist), self._peerlist_limit), + level=log.UNUSUAL) # are there any peers on the list that we haven't used? new_query_peers = [] for (peerid, conn) in self._peerlist: @@ -531,8 +532,8 @@ class Retrieve: # k - max(known_version_sharecounts) + some extra break if new_query_peers: - self.log("UNUSUAL: sending %d new queries (read %d bytes)" % - (len(new_query_peers), self._read_size)) + self.log("sending %d new queries (read %d bytes)" % + (len(new_query_peers), self._read_size), level=log.UNUSUAL) for (peerid, conn) in new_query_peers: self._do_query(conn, peerid, self._storage_index, self._read_size, @@ -541,7 +542,7 @@ class Retrieve: return # we've used up all the peers we're allowed to search. Failure. - self.log("WEIRD: ran out of peers") + self.log("ran out of peers", level=log.WEIRD) e = NotEnoughPeersError("last failure: %s" % self._last_failure) return self._done(failure.Failure(e)) @@ -587,7 +588,7 @@ class Retrieve: sharedata = validator(peerid, root_hash, shnum, data) assert isinstance(sharedata, str) except CorruptShareError, e: - self.log("WEIRD: share was corrupt: %s" % e) + self.log("share was corrupt: %s" % e, level=log.WEIRD) sharemap[shnum].discard(shareinfo) if not sharemap[shnum]: # remove the key so the test in _check_for_done @@ -714,12 +715,12 @@ class Publish: num = self._node._client.log("Publish(%s): starting" % prefix) self._log_number = num - def log(self, msg, parent=None): + def log(self, msg, parent=None, **kwargs): prefix = self._log_prefix if parent is None: parent = self._log_number num = self._node._client.log("Publish(%s): %s" % (prefix, msg), - parent=parent) + parent=parent, **kwargs) return num def log_err(self, f): @@ -863,7 +864,7 @@ class Publish: # found it partial_peerlist.append( (permutedid, peerid, conn) ) return partial_peerlist - self.log("WEIRD: we aren't in our own peerlist??") + self.log("we aren't in our own peerlist??", level=log.WEIRD) return partial_peerlist def _do_query(self, conn, peerid, peer_storage_servers, storage_index): @@ -905,9 +906,9 @@ class Publish: # self._pubkey is present because we require read-before-replace valid = self._pubkey.verify(prefix, signature) if not valid: - self.log("WEIRD: bad signature from %s shnum %d" % + self.log("bad signature from %s shnum %d" % (shnum, idlib.shortnodeid_b2a(peerid)), - parent=lp) + parent=lp, level=log.WEIRD) continue share = (shnum, seqnum, root_hash) @@ -950,8 +951,8 @@ class Publish: alleged_privkey_s = self._node._decrypt_privkey(enc_privkey) alleged_writekey = hashutil.ssk_writekey_hash(alleged_privkey_s) if alleged_writekey != self._writekey: - self.log("WEIRD: invalid privkey from %s shnum %d" % - (idlib.nodeid_b2a(peerid)[:8], shnum)) + self.log("invalid privkey from %s shnum %d" % + (idlib.nodeid_b2a(peerid)[:8], shnum), level=log.WEIRD) return # it's good