[Top] [All Lists]

Re: xfs crash forensics

To: xfs@xxxxxxxxxxx
Subject: Re: xfs crash forensics
From: Stan Hoeppner <stan@xxxxxxxxxxxxxxxxx>
Date: Wed, 21 Apr 2010 11:11:43 -0500
In-reply-to: <20100421152751.24c833f2@xxxxxxxxxxxxxxxxxxxx>
References: <20100421131207.3c845ba9@xxxxxxxxxxxxxxxxxxxx> <4BCEE469.8040701@xxxxxxxxxxxxxxxxx> <20100421152751.24c833f2@xxxxxxxxxxxxxxxxxxxx>
User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv: Gecko/20100317 Thunderbird/3.0.4
Emmanuel Florac put forth on 4/21/2010 8:27 AM:
> Le Wed, 21 Apr 2010 06:41:29 -0500
> Stan Hoeppner <stan@xxxxxxxxxxxxxxxxx> écrivait:
>> Smells like a disk going bad.  What does SMART say about the disk
>> attached to port 11?
> surprisingly, absolutely nothing after the reboot. The disk just
> "cleaned up" all by itself. There are any registered alarms on the
> controller, too.

You need to dig for more information on drive scsi6.  The messages logged
appear to be saying that many sectors were replaced with spares and the
originals marked bad.  Additionally, there appears to have been a bus
timeout during the same time period.  This leads me to believe that drive is
faulty and should be replaced.  Use smartctl or other tools to grab the
SMART data from that drive.  I'm not sure exactly how to do so with drives
connected to a 3ware controller.  IIRC smartctl needs some extra switches
for 3ware cards.  Google is your friend here.

Please don't go on as if nothing happened and everything is fine now.  You
need to find out if that drive is indeed going bad, which appears, from
here, to be the case.


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