My comments are below.
From: dm-devel-bounces redhat com [mailto:dm-devel-bounces redhat com] On Behalf Of Mike Christie
Sent: Thursday, March 30, 2006 2:45 PM
To: device-mapper development
Cc: Ryan O'Hara
Subject: Re: Fwd: [dm-devel] IO error on DM device
Jonathan E Brassow wrote:
> Another observation is that we are not seeing any IO errors when the
> same test is executed on SLES9 SP3/SP4.
[Murthy] our other team has seen the error on SLES9 too. The Qlogic driver is same as RHEL ones, so we have to see similar behaviour.
I think Ryan's post is correct, but I wonder why the SLES9 is not returning errors. Was the SLES statement referring to the scsi errors you see in the log like this:
Feb 1 11:47:45 apwtest52 kernel: SCSI error : <3 0 0 1> return code
or the dm-multipath and block layer messages
Feb 1 11:48:40 apwtest52 kernel: device-mapper: dm-multipath:
Failing path 8:32.
Feb 1 11:48:40 apwtest52 kernel: Buffer I/O error on device dm-0,
logical block 116940 u
Do you have the same dm-multipath settings (queue if no path, etc)? [Murthy] yes
Is this is the qlogic qla2xxx driver? [Murthy] yes
Are there qlogic guys on the list that would know if there are any differences between the SLES and RHEL drivers or are we going to have to dig in there? Are there differences between the SLES and RHEL dm multipath drivers? Does this occur with the upstream kernel?
dm-devel mailing list
dm-devel redhat com