All of lore.kernel.org
 help / color / mirror / Atom feed
From: Benjamin Marzinski <bmarzins@redhat.com>
To: device-mapper development <dm-devel@redhat.com>
Subject: Re: RDAC path checker status change messages
Date: Fri, 26 Jun 2009 11:07:36 -0500	[thread overview]
Message-ID: <20090626160736.GK3172@ether.msp.redhat.com> (raw)
In-Reply-To: <Pine.LNX.4.64.0906251314330.4229@e-smith.charlieb.ott.istop.com>

On Thu, Jun 25, 2009 at 01:28:40PM -0400, Charlie Brady wrote:
>
> On Thu, 25 Jun 2009, Charlie Brady wrote:
>
>> On Thu, 25 Jun 2009, Charlie Brady wrote:
>>
>>>  I think I need to retract that. The messages appear to be printed when the
>>>  status is not changing, at least for the "path down" messages.
>>>
>>>  I've patched scsi_dh_rdac.c so that SUN/LCSM100_I is included in
>>>  rdac_dev_list[], and restarted iscsi and multipathd. I've now provoked a
>>>  path failure via iptables. Here are the logs I see (with unpatched
>>>  multipathd) - the duplicate message suppression from the rdac path checker
>>>  is not working:
>>
>> OTOH, on path recovery, the wrong message is displayed ("path down" 
>> instead of "path up") but the message is not repeated:
>
>> From multipathd/main.c, there are three conditions where the messages are 
> printed:
>
> ...
>                         if (newstate != pp->state) {
>                                 int oldstate = pp->state;
>                                 pp->state = newstate;
>                                 LOG_MSG(1, checker_message(&pp->checker));
> ...
>                         }
>                         else if (newstate == PATH_UP || newstate == PATH_GHOST) {
>                                 LOG_MSG(4, checker_message(&pp->checker));
> ...
>                         else if (newstate == PATH_DOWN)
>                                 LOG_MSG(2, checker_message(&pp->checker));
> ...
>
> I think only the first might be required, and only the first will be seen 
> if logging verbosity < 2 (default is 2).
>
> So there are two path checker problems I can see (in 
> multipath-tools-0.4.7.rhel5.13):
>
> - the rdac checker leaves the message set to "path down" when path
>   transitions back to up or ghost.

Any your patch fixes that.

> - multipathd prints the message on every PATH_DOWN except the first when
>   verbosity is >= 2, and on every PATH_UP or PATH_GHOST except the first
>   when verbosity is >= 4.

Why is this a problem? 

On verbosity = 1, you only get messages when you change states.
On verbosity = 2, you additionally get repeated messages for failed
paths, which are often something that the sysadmin should be doing
something about.
On verbosity = 4, you get a message whenever the path checker runs,
which is helpful for debugging, which is the only time you would ever
want verbosity = 4.

This seems sensible to me.  If you don't like the repeated path down
messages, you can change your verbosity level to 1.  Right now that
involves editting the init script, but in future releases, you can
change it in /etc/multipath.conf.

-Ben

>
> ---
> Charlie
>
> --
> dm-devel mailing list
> dm-devel@redhat.com
> https://www.redhat.com/mailman/listinfo/dm-devel

  reply	other threads:[~2009-06-26 16:07 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-06-23 19:50 RDAC path checker status change messages Charlie Brady
2009-06-24 18:15 ` Chandra Seetharaman
2009-06-24 20:19   ` Charlie Brady
2009-06-24 21:01     ` Benjamin Marzinski
2009-06-25 16:57     ` Charlie Brady
2009-06-25 17:02       ` Charlie Brady
2009-06-25 17:28         ` Charlie Brady
2009-06-26 16:07           ` Benjamin Marzinski [this message]
2009-06-25 18:29     ` Chandra Seetharaman
2009-06-25 18:40       ` Charlie Brady
2010-11-03 21:09 ` Mike Snitzer
2010-11-03 22:16   ` Christophe Varoqui

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20090626160736.GK3172@ether.msp.redhat.com \
    --to=bmarzins@redhat.com \
    --cc=dm-devel@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.