[tahoe-lafs-trac-stream] [Tahoe-LAFS] #1939: memory leak (during check --repair --add-lease)
Tahoe-LAFS
trac at tahoe-lafs.org
Wed Dec 3 21:47:46 UTC 2014
#1939: memory leak (during check --repair --add-lease)
----------------------------+------------------------
Reporter: killyourtv | Owner: killyourtv
Type: defect | Status: new
Priority: major | Milestone: undecided
Component: code | Version: 1.9.2
Resolution: | Keywords:
Launchpad Bug: |
----------------------------+------------------------
Comment (by zooko):
In contradiction to what kpreid said in comment:2, the flog that kv
attached (attachment:incident-2013-04-02--11-19-06Z-urhgqtq.flog.bz2) does
have some evidence that points to a similar problem as kpreid's original
report. It has this:
{{{
[CopiedFailure instance: Traceback from remote host -- Traceback (most
recent call last):
Failure: foolscap.ipb.DeadReferenceError: Connection was lost (to
tubid=ngqc) (during method=RIBucketReader:read)
]
}}}
and this:
{{{
2013-04-01_21:20:12.017651Z [110581]: UNUSUAL {'facility':'tahoe.encoder',
'failure':<foolscap.call.CopiedFailure
allmydata.util.pipeline.PipelineError>, 'format':'error while sending
%(method)s to shareholder=%(shnum)d',
'incarnation':('\xd0N\xca\xf7\xa7l\x8f\xce', None), 'parent':110314,
'shnum':5, ...} FAILURE:
[CopiedFailure instance: Traceback from remote host -- Traceback (most
recent call last):
Failure: allmydata.util.pipeline.PipelineError: <PipelineError
error=([Failure instance: Traceback: <class
'foolscap.ipb.DeadReferenceError'>: Calling Stale Broker
/usr/lib/pymodules/python2.7/allmydata/immutable/layout.py:237:_write
/usr/lib/pymodules/python2.7/allmydata/util/pipeline.py:89:add
/usr/lib/python2.7/dist-
packages/twisted/internet/defer.py:134:maybeDeferred
/usr/lib/pymodules/python2.7/foolscap/referenceable.py:415:callRemote
--- <exception caught here> ---
/usr/lib/python2.7/dist-
packages/twisted/internet/defer.py:134:maybeDeferred
/usr/lib/pymodules/python2.7/foolscap/referenceable.py:455:_callRemote
/usr/lib/pymodules/python2.7/foolscap/broker.py:477:newRequestID
])>
]
}}}
This is similar to the evidence in kpreid's log, which suggests that maybe
a foolscap rref becoming None led to the problem.
--
Ticket URL: <https://tahoe-lafs.org/trac/tahoe-lafs/ticket/1939#comment:16>
Tahoe-LAFS <https://Tahoe-LAFS.org>
secure decentralized storage
More information about the tahoe-lafs-trac-stream
mailing list