bzr crashes on commit with exceptions.SystemErrorr

Bug #882056 reported by Kaja Podlaska Christiansen
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Bazaar
New
Undecided
Unassigned

Bug Description

bzr reports:
aborting commit write group: SystemError('../Objects/stringobject.c:4271: bad argument to internal function',)
bzr: ERROR: exceptions.SystemError: ../Objects/stringobject.c:4271: bad argument to internal function

Revision history for this message
Kaja Podlaska Christiansen (kaja) wrote :
Revision history for this message
Martin Packman (gz) wrote :

See the existing bug for more info, but in short this is a bug that should be fixed in Python 2.6.6 which is available in newer Ubuntu versions.

Revision history for this message
Kaja Podlaska Christiansen (kaja) wrote :

I've attempted the same bzr commit on a machine running Maverick and Python 2.6.6. Commit crashed, thou this time at the other end (i.e. receiving server) and with a different error message:

bzr: failed to report crash using apport:
     OSError(13, 'Permission denied')
bzr: ERROR: bzrlib.errors.ErrorFromSmartServer: Error received from smart server: ('error', 'Objects/stringobject.c:3565: bad argument to internal function')

The file I've been trying to commit is very large: 2.6G directory tree with about 100000 files. The commit basically eats up all the memory and then crashes. So, Python 2.6.6 doesn't seem to be an answer. What is the solution then? Not use bzr on huge files, or add lots of swap, or wait for another Python fix?

Thanks/Kaja

Revision history for this message
Martin Packman (gz) wrote :

What server is that operating against? The server also needs the Python fix in 2.6.6 and launchpad seems to be running the older 2.6.5 version too.

That said, versioning 2.6GB of data may well run out of memory even if you get past this issue. Large binary files especially are best not committed to version control, they can't be usefully merged or diffed and storage is optimised for those uses,

Revision history for this message
Kaja Podlaska Christiansen (kaja) wrote : Re: [Bug 882056] Re: bzr crashes on commit with exceptions.SystemErrorr

> What server is that operating against? The server also needs the Python
> fix in 2.6.6

Yes, thanks, didn't check the server: this is a RedHat server with an older
Python.

> That said, versioning 2.6GB of data may well run out of memory even if
> you get past this issue. Large binary files especially are best not
> committed to version control

Ok, thanks again.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.