xfs
[Top] [All Lists]

Re: stress scripts patches

To: "Marcelo E. Magallon" <marcelo.magallon@xxxxxxxxxxx>, tes@xxxxxxxxxxxxxxxxxxxxxxx, ivanr@xxxxxxxxxxxxxxxxxxxxxxx
Subject: Re: stress scripts patches
From: "Nathan Scott" <nathans@xxxxxxxxxxxxxxxxxxxxxxxx>
Date: Wed, 1 Nov 2000 09:17:02 -0400
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: "Marcelo E. Magallon" <marcelo.magallon@xxxxxxxxxxx> "stress scripts patches" (Oct 31, 8:21pm)
References: <20001031202111.A26772@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
hi Marcelo,

On Oct 31,  8:21pm, Marcelo E. Magallon wrote:
> Subject: stress scripts patches
> 
> Hi,
> 
>     attached is a patch for the stress test scripts.  The first one (002)

great - thanks.

>  works arround a bug in bash (in case /bin/sh is bash) where it doesn't
>  discard trailing whitespace when comparing two numbers.  lstat64 will
>  always output a line containing "Links:", so $x should never be empty.
> 

ok, fixed.

>  The second one (004) adds a "_require_scratch" to the test, otherwise
>  it fails when run as patch of a batch because the scratch partition is
>  still mounted from the previous test.
> 

fixed.

>  The last one was the product of me trying to figure out if a failure
>  (see other email) was because of the script or a real failure.  It just
>  moves the RESTORE_DIR and RESTORE_SUBDIR (this one is not really
>  needed) to the top of the substitution list, otherwise you end up with
>  something like "SCRATCH_MNT/some/random/directory".
> 

I'll leave this for Tim/Ivan to apply while they're looking
into the other dump/restore failures.

cheers.

-- 
Nathan

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