[dm-devel] dm mpath: delay retry activate_path on SCSI_DH_RETRY

Chandra Seetharaman sekharan at us.ibm.com
Tue Nov 2 19:56:57 UTC 2010


Yes, I would like to see this go in.

Thanks for cleaning it up.

chandra
On Tue, 2010-11-02 at 15:32 -0400, Mike Snitzer wrote:
> On Tue, Jun 09 2009 at  4:54pm -0400,
> Chandra Seetharaman <sekharan at us.ibm.com> wrote:
> 
> > No functional change.
> > 
> > Repost with changes made to apply cleanly on top of the patch
> > (http://marc.info/?l=dm-devel&m=124424663327710&w=2) that changed
> > dm-mpath.c
> > 
> > -----------------
> > From: Chandra Seetharaman <sekharan at us.ibm.com>
> > 
> > SCSI Device Handlers return SCSI_DH_IMM_RETRY if we could retry
> > immediately and SCSI_DH_RETRY in cases where it is better to retry
> > after some delay.
> > 
> > Currently we retry immediately irrespective of SCSI_DH_IMM_RETRY and
> > SCSI_DH_RETRY. This patch adds a user configurable attribute
> > pg_init_delay_secs which specifies the number of seconds to delay
> > before retrying scsi_dh_activate, when it returned SCSI_DH_RETRY.
> > 
> > Default for this attribute is set to 2 seconds.
> > 
> > Signed-off-by: Nikanth Karthikesan <knikanth at suse.de>
> > Signed-off-by: Chandra Seetharaman <sekharan at us.ibm.com>
> 
> I'm auditing dm-devel's patchwork and would just like to confirm that
> others would still like to see this go in?
> 
> Patch in question is captured here:
> https://patchwork.kernel.org/patch/29088/
> 
> Mike C.,
> An earlier version of this patch got your "Looks ok.":
> https://patchwork.kernel.org/patch/23946/
> Can I translate that to your "Acked-by:" if/when I rebase and repost?
> 
> Please advise,
> Mike





More information about the dm-devel mailing list