[OpenAFS] Crash testing OpenAFS

ted creedon tcreedon@easystreet.com
Thu, 11 Aug 2005 21:34:46 -0700


Thanks for no flames.

Yes, I have something that can be used but it won't fit on anything but a
250GB disk drive.

Looks to me like there a problem related to data transfer volume. (I.e.
MIT's experience).

I'd be glad to send 2 or 3 250GB firewire or ide drives at no cost to
competent developers who would actually use the data for testing.

One note, my /usr/vice partition actually crashed every time (/dev/sda2),
since no other partition on that disk crashed and it can be resurrrected
with a mkfs.ext2 /dev/sda2 it should not be a hardware problem.

Send Fedex addresses and specify what you want, IDE or firewire.

You pick the lucky recipients...


-----Original Message-----
From: openafs-info-admin@openafs.org [mailto:openafs-info-admin@openafs.org]
On Behalf Of Derrick J Brashear
Sent: Thursday, August 11, 2005 9:17 PM
To: openafs-info@openafs.org
Subject: Re: [OpenAFS] Crash testing OpenAFS

On Thu, 11 Aug 2005, ted creedon wrote:

> A simple cp -rpv /afs/.bigcell/foo /afs/.home-ted-doris.fam/bar hangs 
> the system so badly Linux won't even "halt".

If you have magic sysrq turned on, hit alt sysrq t, and ideally collect the
resulting backtrace and put it somewhere.

> If there's a problem with windows file names, (Jeff has stated so) it 
> needs to be documented. And fixed.

I don't know the details of what you speak. But I'd guess either "it's
simple and we can fix it" or "it's complicated and either no one has fixed
it or no one has decided to pay someone else to fix it"

> There needs to be a set of tests run on every release before its 
> thrown over the threshold.

No argument. Having a better framework than the one in src/tests would help,
but it is at least a start. Have something we should be running?


OpenAFS-info mailing list