Data Corruption bug in GlusterFS 3.0.2

[nutznboltz]

Impact: data corruption when using GlusterFS with certain options set.
Patch: https://bugs.launchpad.net/ubuntu/+source/glusterfs/+bug/666494/+attachment/1711751/+files/glusterfs_3.0.2-1ubuntu1.debdiff
Reproducing: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=639
Regression potential: none known

As an interim fix I did a backport of GlusterFS 3.0.4 from Maverick to Lucid
https://launchpad.net/~nutznboltz/+archive/glusterfs3
so my systems are safe.

This is the upstream Changelog message:

commit c55134fc0c5ebe952de12757d5c53d463a9c21b5
Author: Anand Avati <email address hidden>
Date: Thu Feb 18 10:28:06 2010 +0000

    server_writev: add proper iobuf into iobref

    this typo results in corruption when write-behind is loaded on server side

    Signed-off-by: Anand V. Avati <email address hidden>
    Signed-off-by: Anand V. Avati <email address hidden>

    BUG: 639 (Data corruption with write-behind on server side)
    URL: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=639

ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: glusterfs-server (not installed)
ProcVersionSignature: Ubuntu 2.6.32-25.45-generic 2.6.32.21+drm33.7
Uname: Linux 2.6.32-25-generic i686
Architecture: i386
Date: Mon Oct 25 16:24:24 2010
InstallationMedia: Ubuntu 10.04 LTS "Lucid Lynx" - Release i386 (20100427.1)
ProcEnviron:
 LANG=en_US.utf8
 SHELL=/bin/bash
SourcePackage: glusterfs

[Stefano Rivera (stefanor)]

Hi, you subscribed ubuntu-sponsors, but there's no clear action for us, so I've unsubscribed sponsors for now. Please resubscribe when there's an upload requiring sponsorship.

What exactly are you requesting? An SRU patch for the version in lucid? If so please nominate this bug for lucid. Can you prepare a debdiff / bzr branch for SRUing?

[nutznboltz (nutznboltz-deactivatedaccount)]

Debdiff with patch from upstream.

[Stefano Rivera (stefanor)]

Thanks. This just needs some tweaks before it can be uploaded:

> +## Description: add some description
> +## Origin/Author: add some origin or author
> +## Bug: bug URL

Would you mind updating those to actually have content?
Can you also provide a test case (as required by https://wiki.ubuntu.com/StableReleaseUpdates ). I understand that this might not be possible for this bug, but that should then be noted.

> +glusterfs (3.0.2-1ubuntu1) lucid; urgency=low

Can you change that version number to 3.0.2-1ubuntu0.1 and target it at lucid-proposed?

> + * debian/patches/1-fix-data-corruption-bug.patch: Fixes LP 666494

The correct syntax to close a bug is LP: #666494. "Fixes" is not required.

> + -- nutznboltz <email address hidden>

Can you provide a real name?

Normally adding a patchsystem to Debian packages is not the recommended approach, but as this is an SRU, it seems like a good enough method to me.

[nutznboltz (nutznboltz-deactivatedaccount)]

debdiff updated as requested.

[nutznboltz (nutznboltz-deactivatedaccount)]

Didn't see the part about marking test cases as "TEST CASE:" until a moment ago.
Updated debdiff regarding that.

[nutznboltz (nutznboltz-deactivatedaccount)]

Another slight improvement to the comments in the patch. The "impact:" field has been renamed to "Description:".

[Stefano Rivera (stefanor)]

The test case and regression potential belong in the description of this bug, not the patch.

> * debian/patches/1-fix-data-corruption-bug.patch: LP #666494

It's still missing a colon after the LP, and it could use a description. I'd do this:
* 1-fix-data-corruption-bug.patch: Backport typo bugfix patch. LP: #666494

> -- nutznboltz <email address hidden>

I'd prefer a real name.

> +## Description: data corruption when using GlusterFS with certain options set.
> +## Origin/Author: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=639
> +## TEST CASE: http://bugs.gluster.com/cgi-bin/bugzilla3/show_bug.cgi?id=639
> +## Regression potential: none known

I'd remove the second two lines of that, and remove "/Author", you are giving the origin not the author.

[nutznboltz (nutznboltz-deactivatedaccount)]

Thanks for helping me with my first debdiff ever, Stefano Rivera. :)

Here's the updated debdiff again.

[Stefano Rivera (stefanor)]

No problem. Hope to see more debdiffs from you soon :)

Uploaded to lucid-proposed, pending SRU team review. (Subscribed Ubuntu-sru).

[Martin Pitt (pitti)]

SRU ack, please get it uploaded.

[nutznboltz (nutznboltz-deactivatedaccount)]

Just a link for my benefit to see recent lucid-proposed uploads:
http://www.ubuntuupdates.org/package_logs?type=sources&vals=325

[nutznboltz (nutznboltz-deactivatedaccount)]

Another link for my own benefit
http://people.canonical.com/~ubuntu-archive/pending-sru.html

[Martin Pitt (pitti)]

Accepted glusterfs into lucid-proposed, the package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

[Martin Pitt (pitti)]

Any testers of the proposed lucid package?

[nutznboltz (nutznboltz-deactivatedaccount)]

We stopped using GlusterFS as it was deemed too much of a risk.

[Martin Pitt (pitti)]

I removed the proposed package, as there has been no feeback in over three months, and we need to clear the queues for 10.04.2.