aborted pending locks left behind

Bug #109169 reported by Martin Pool
2
Affects Status Importance Assigned to Milestone
Bazaar
Fix Released
High
Martin Pool

Bug Description

Bazaar lockdirs can get pending.* directories left behind if clients are interrupted. The user can just delete them but maybe Bazaar should remove them automatically if they are more than N days old, or on some other reasonable criteria.

Revision history for this message
John A Meinel (jameinel) wrote :

It creates a lot more if the lock is currently held. At present, it tries every 1s to lock the target, and if it is already locked, it will leave a pending lock file.

With the standard timeout of 5m an old lock can generate 300 pending.* directories.

I think that we could

1) Have a failed attempt cleanup after itself

or

2) Have a failed attempt reuse the 'pending.*' directory. It might re-use the 'info' file, but as that contains a timestamp, we may prefer to update it.

Revision history for this message
Robert Collins (lifeless) wrote :

It should do 1) imo.

Revision history for this message
Martin Pool (mbp) wrote :

Making failed attempts clean up should be pretty trivial and handle most cases. Collecting any other leftovers would be useful.

Changed in bzr:
importance: Undecided → High
status: Unconfirmed → Confirmed
Revision history for this message
John A Meinel (jameinel) wrote :

well doing 2 means a bit less overhead, since you don't have a "mkdir()" and then "rm pending/info; rmdir pending".

Certainly it should clean up when all is done.

But 1) is easier to do, so it would be a reasonable starting location.

Revision history for this message
Bubba Siggler (bud3) wrote : Re: [Bug 109169] Re: aborted pending locks left behind

Hi john, I don't have permission to edit the invertory.py evethrow I
tried gvim to do so it was a great try in so. what the step to clean
up

best regards

bubba

On 4/27/07, John A Meinel <email address hidden> wrote:
> well doing 2 means a bit less overhead, since you don't have a "mkdir()"
> and then "rm pending/info; rmdir pending".
>
> Certainly it should clean up when all is done.
>
> But 1) is easier to do, so it would be a reasonable starting location.
>
> --
> aborted pending locks left behind
> https://bugs.launchpad.net/bugs/109169
> You received this bug notification because you are a member of Bazaar
> Developers, which is the registrant for Bazaar.
>

--
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)

iD8DBQFFkcN7yXWcajQQndYRAgbqAKCMyXN9Jx4g0X7jocg+aUSFz0x4LwCgrURW
eGtqLjpzQVYa9+gzpCRtB84=
=zrpM
-----END PGP SIGNATURE-----

Martin Pool (mbp)
Changed in bzr:
assignee: nobody → mbp
status: Confirmed → Fix Committed
Martin Pool (mbp)
Changed in bzr:
status: Fix Committed → Fix Released
Revision history for this message
Peter Balazovic (peter-balazovic) wrote :

Other

Run command: bzr pull ..\code\DEV\ver06
bzr: ERROR: [Error 5]

Revision history for this message
Martin Pool (mbp) wrote :

Hi Peter,

I'm pretty sure this bug has been fixed; please report a new one with
your problem. You may find
<http://www.chiark.greenend.org.uk/~sgtatham/bugs.html> useful.

Martin

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.