[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: ext3 and data=journal bug



I think I saw somewhere that the 2.4.20 kernel should be patched for use with 
quotas.  Is that true?  If you remember offhand, where was it that you got 
the info in the vfsold quota?  That is what we are currently using.

Thanks,

Mark

On Thursday 01 May 2003 06:50 am, Matt Bernstein wrote:
> At 00:18 -0700 Andrew Morton wrote:
> >Mark A Basil <mbasil alabanza com> wrote:
> >> One of my other main concerns is quotas.  I hear ext3+quota = deadlocks
> >> in alot of cases.  Luckily I havn't seen any as of yet.  I'm hoping that
> >> I don't see any at all.  I'm assuming that using the data=journal mode
> >> is beneficial to quotas on a server as well.
>
> You should chattr +j the {,a}quota.{user,group} files if you're not using
> data=journal.
>
> >Yes, quota is still deadlocky apparently.  I thought it was OK actually.
> >Maybe it got broken some time in the past 18 months :(
>
> I've read it's better with vfsv0 (ie the 32-bit patches), and have
> occasionally experienced quota corruption under vfsold.
>
> >Stephen is working on it.   I assume the deadlocks are pretty obscure.
>
> Our moderately-high-load NFS/SMB server has uptimes of about 3 months
> before deadlocking with no excuse offered.. I guess that's what's biting
> us.
>
> Have been running 2.4.19-pre10-ac2 on it, but will be looking at 2.4.21
> when it's out, and would dearly love to be brave enough to try htree and
> external journals (but I'm fresh out of Brave Pills just now).





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]