[Top] [All Lists]

Re: [PATCH] xfstests: replace xfs_check with xfs_repair -n

To: sekharan@xxxxxxxxxx
Subject: Re: [PATCH] xfstests: replace xfs_check with xfs_repair -n
From: Eric Sandeen <sandeen@xxxxxxxxxxx>
Date: Wed, 17 Apr 2013 11:23:29 -0700
Cc: XFS mailing list <xfs@xxxxxxxxxxx>
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <1366221810.3762.32341.camel@xxxxxxxxxxxxxxxxxx>
References: <1366216695.3762.32325.camel@xxxxxxxxxxxxxxxxxx> <516ED4A0.9040706@xxxxxxxxxxx> <1366221810.3762.32341.camel@xxxxxxxxxxxxxxxxxx>
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.8; rv:17.0) Gecko/20130328 Thunderbird/17.0.5
On 4/17/13 11:03 AM, Chandra Seetharaman wrote:
> Hi Eric,
> Thanks for the quick feedback.
> On Wed, 2013-04-17 at 09:58 -0700, Eric Sandeen wrote:
>> On 4/17/13 9:38 AM, Chandra Seetharaman wrote:
>>> Replace the usage of "xfs_check" with "xfs_repair -n" as xfs_check
>>> is planned to be depracated.
>> Hm, I thought the plan was to keep xfs_check around for xfstests
> I didn't think the plan was to keep xfs_check, may be I misunderstood.
> My understanding was that we wanted to deprecate xfs_check, but first we
> have to make xfstests not use xfs_check.
>> use, for now; as Dave said in the earlier thread:
>>> xfstests also still needs to run xfs_check. That means we also need
>>> either an override flag an make $XFS_CHECK_PROG have it set
>>> appropriately or add an internal xfs_db wrapper that runs the
>>> xfs_check functionality appropriately. The second is probably the
>>> better option...
>> but that's not what this patch does...
> The usages of xfs_check in xfstests looked simple and straight forward.
> Besides, I thought we should do what we suggest our users to do :),
> hence replaced xfs_check with "xfs_repair -n".

Dave or others can chime in too, but I think we still want xfs_check
(xfs_db) as a verifier against xfs_repair.

> Does this patch break something or technically incorrect ?

We used to explicitly run both xfs_repair and xfs_check to get two
distinct verification passes; the patch removes part of that, so I'd
say yes, it does "break" things a little.

> Do you think I should instead use 
>     xfs_db -F -i -p xfs_check -c "check" <dev>

Right, if the xfs_check script itself is going away, I think we still
want to invoke "xfs_check" behavior one way or another in xfstests to
keep current xfs verification levels for now.


> Please advise.

<Prev in Thread] Current Thread [Next in Thread>