[Top] [All Lists]

Re: [PATCH] xfstests: v2 xfs directory unbalance assert test

To: Mark Tinguely <tinguely@xxxxxxx>
Subject: Re: [PATCH] xfstests: v2 xfs directory unbalance assert test
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Tue, 24 Sep 2013 09:30:12 +1000
Cc: xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20130917192538.230164044@xxxxxxx>
References: <20130917192538.230164044@xxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Tue, Sep 17, 2013 at 02:25:23PM -0500, Mark Tinguely wrote:
> This tests triggers an assert in the XFS directory unbalance code.
> This test originally written by Brian Foster and suggestions
> from Micheal Semon.
> Signed-off-by: Mark Tinguely <tinguely@xxxxxxx> 

A recent comment about the _cleanup funtion being removed made me
look at this again....

Where's the trap function to catch a ctrl-c and clean up the test
when that happens?

i.e. at minimum, the trap and cleanup function is used to ensure the
test is marked as failed if it is interrupted, and in this case it
probably needs to kill all the fsstress processes and wait for them
to complete, as well as remove

>  put into auto group and "long" group so people can opt out:
>   this test almost an hour on my test box.

The group for this is called "stress" - there's no need to introduce
a new group for the same test criteria.


Dave Chinner

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