Re: [RFC] do you want jbd2 interface of ext3?

[Dmitry Monakhov <dmonakhov@openvz.org>]

Jan Kara <jack@suse.cz> writes:

> On Mon 22-02-10 08:55:53, Theodore Tso wrote:
>> As for quota --- quite seriously --- if you have mission critical users,
>> I'd suggest that they not use quota.  Dimitry has been turning up all
>> sorts of bugs in the quota subsystem, many of which are just as
>> applicable to ext3.  The real issue is that quota hasn't received as much
>> testing as other file system features --- in any file system, not just
>> ext4.
>   I don't agree with this. I know about quite a few large customers
> depending on quotas on their servers and they run on ext3 / reiserfs quite
> happily. Dmitry's patches touching the generic code were mostly cleanups,
> the fixes were just in the delayed allocation handling but that never
> gets executed for ext3 or reiserfs...
Stability is relative thing. I's quite depends on usecase.
For example after triggering bug on not empty orphan list on ext3_umount
i've started full orphan-list management code revision. And both
ext3/ext4 appears to be almost broken in case of errors.
But nobody seems never catch it in real life. But still
at that time i have triggered:
1) non empty orphan list on umount for both (ext3 and ext4)
2) on_disk linked list corruption  for both
3) data blocks beyond i_size 
4) bit-difference on fsck  for both
Currently i'm working on fixes. It takes week or so.
So at least i'll reduce "project_id quota" spam flow a bit.
>   I don't say there cannot be bugs and certainly quota code has less
> exposure than other more used filesystem parts. But I don't know about
> any serious quota issue on ext3 / reiserfs in last two years or so
> (except the one that was caused by Dmitry's fixes ;).
This time i'll try to give enough test coverage.
>
> 								Honza