All of lore.kernel.org
 help / color / mirror / Atom feed
From: Martin Schwenke <martin@meltin.net>
To: device-mapper development <dm-devel@redhat.com>
Cc: k-ueda@ct.jp.nec.com
Subject: Re: RHEL 5.4 multipath doesn't handle virtio block devices
Date: Tue, 30 Mar 2010 18:43:34 +1100	[thread overview]
Message-ID: <20100330184334.43cf7934@martins.ozlabs.org> (raw)
In-Reply-To: <4BB1A53A.4030806@suse.de>

On Tue, 30 Mar 2010 09:16:10 +0200, Hannes Reinecke <hare@suse.de>
wrote:

> megasas megasas megasas ...
> I posted a megasas HBA emulation for KVM a while back.
> Would solve your problem rather elegantly.
> But then you'd have to recompile KVM, at which point you could include
> the SCSI emulation as well ...

Right, that sounds good.  However, for reasons I won't go into, I need
to use the version of KVM that ships with RHEL 5.  So recompiling or
even using RPMs provided by a 3rd party isn't an easy option.  :-(

> Problem is, though, that you don't have a unique ID for virtio
> devices. EG scsi_id and friends don't work here.
> And AFAIR there is no feature flag which would allow you do set one.
> 
> How do you distinguish between, say, 2 virtio devices across 2 paths?

I write a unique identifier onto each image using dd and then as you
would expect, in the libvirt XML configuration, I make 2 devices point
to each image.

Then I use a custom getuid_callout setting, as follows:

  [root@strg001st068 ~]# cat /sbin/scsi_id_autocluster.sh
  #!/bin/sh
  # get scsi ID assuming disks setup by autocluster
  /bin/dd if=/dev/$1 bs=1 count=13 2> /dev/null | cat -v
  echo

The filesystem being used on those disks doesn't use the first
block...  ;-)

It works nicely...

peace & happiness,
martin

      reply	other threads:[~2010-03-30  7:43 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-03-30  1:16 RHEL 5.4 multipath doesn't handle virtio block devices Martin Schwenke
2010-03-30  5:07 ` Kiyoshi Ueda
2010-03-30  6:23   ` Martin Schwenke
2010-03-30  7:16     ` Hannes Reinecke
2010-03-30  7:43       ` Martin Schwenke [this message]

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=20100330184334.43cf7934@martins.ozlabs.org \
    --to=martin@meltin.net \
    --cc=dm-devel@redhat.com \
    --cc=k-ueda@ct.jp.nec.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.