Open Stack

On Wednesday, July 02, 2014 1:06 PM, Pete Zaitcev <zaitcev at redhat.com> wrote:

Thanks for the detailed explanation.

Let me clarify the behavior of swift.

(1) Use ext4 on devices.
(2) Corrupt the data on (1)'s filesystem
(3) Move corrupt files to lost+found without a trace by ext4's fsck
(4) Cannot recognize (3) by Swift's auditors so hashes.pkl is not updated.

Is above sequence correct?
If it's correct, I understand we better to use xfs.

Thanks in advance,
Hisashi Osanai

> -----Original Message-----
> From: Pete Zaitcev [mailto:zaitcev at redhat.com]
> Sent: Wednesday, July 02, 2014 1:06 PM
> To: Osanai, Hisashi/小山内 尚
> Cc: OpenStack Development Mailing List (not for usage questions)
> Subject: Re: [openstack-dev] Swift: reason for using xfs on devices
> 
> On Wed, 2 Jul 2014 00:16:42 +0000
> "Osanai, Hisashi" <osanai.hisashi at jp.fujitsu.com> wrote:
> 
> > So I think if performance of swift is more important rather than
> scalability of it, it is a
> > good idea to use ext4.
> 
> The real problem is what happens when your drives corrupt the data.
> Both ext4 and XFS demonstrated good resilience, but XFS leaves empty
> files in directories where corrupt files were, while ext4's fsck moves
> them to lost+found without a trace. When that happens, Swift's auditors
> cannot know that something was amiss and the replication is not
> triggered (because hash lists are only updated by auditors).
> 
> Mr. You Yamagata worked on a patch to address this problem, but did
> not complete it. See here:
>  https://review.openstack.org/11452
> 
> -- Pete