bzr crashed with ShortReadvError in _seek_and_read(): readv() read 0 bytes rather than 174 bytes at 0 for "469df00e67f848246f876663fb299e10.rix"
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
bzr (Ubuntu) |
New
|
Undecided
|
Unassigned |
Bug Description
crashed during commit operation
ProblemType: Crash
DistroRelease: Ubuntu 11.10
Package: bzr 2.4.0-0ubuntu2
ProcVersionSign
Uname: Linux 3.0.0-9-generic i686
Architecture: i386
BzrDebugFlags: set()
BzrVersion: 2.4.0
CommandLine: ['/usr/bin/bzr', 'commit', '-m', 'Started baobab tests']
CrashDb: bzr
Date: Mon Aug 22 11:55:02 2011
ExecutablePath: /usr/bin/bzr
FileSystemEncoding: UTF-8
InstallationMedia: Ubuntu 11.04 "Natty Narwhal" - Beta i386 (20110418)
InterpreterPath: /usr/bin/python2.7
Locale: en_US.UTF-8
PackageArchitec
Platform: Linux-3.
ProcCmdline: /usr/bin/python /usr/bin/bzr commit -m Started\ baobab\ tests
ProcEnviron:
SHELL=/bin/bash
PATH=(custom, no user)
LANG=en_US.UTF-8
PythonVersion: 2.7.2
SourcePackage: bzr
Title: bzr crashed with ShortReadvError in _seek_and_read(): readv() read 0 bytes rather than 174 bytes at 0 for "469df00e67f848
UpgradeStatus: No upgrade log present (probably fresh install)
UserEncoding: UTF-8
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare
Hi,
This is a dupe of bug 413430 which basically says a file has been truncated - typically this happens because of an OS error or disk error or a machine crash.
In 2.4.0 we started flushing files to disk by default after they're written, which was intended to reduce the frequency of this. So, it's possible this problem was introduced by an earlier client and just discovered by bzr 2.4.0, or perhaps our check is not enough, or perhaps something else.
Do you recall if your machine recently crashed or abruptly stopped or had any similar problem?
(fdatasync is not going to be enough on its own; continued incidents of this mean perhaps we need built in recovery.)