public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Adrian Bridgett <adrian.bridgett@iname.com>
To: Ingo Molnar <mingo@elte.hu>
Cc: linux-raid@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [patch] multipath RAID personality, 2.4.10-pre9
Date: Tue, 18 Sep 2001 20:39:46 +0100	[thread overview]
Message-ID: <20010918203946.A12814@wyvern> (raw)
In-Reply-To: <20010916150806.E1541@turbolinux.com> <Pine.LNX.4.33.0109170113010.3960-100000@localhost.localdomain>
In-Reply-To: <Pine.LNX.4.33.0109170113010.3960-100000@localhost.localdomain>

On Mon, Sep 17, 2001 at 01:16:56 +0200 (+0000), Ingo Molnar wrote:
[snip]
> > [...] Also, it is my understanding that with some multipath hardware,
> > if you read from the "backup" path it will kill access to the primary
> > path (this can be used when more than one system access shared disk
> > for failover).  As a result, we should always read from the "primary"
> > path for each disk unless there is an error.
> 
> yes, and this is being done currently, only the primary path is used.

Do you have plans to change this?  I know that the SDD software for AIX load
balances between paths (and I think EMC's Powerpaths do for AIX too), DMP
(from Veritas) for Solaris doesn't.

In fact that brings up another point - which path do you use by default?  If
you have the SAN situation where you have a farm of servers each with two FC
cards two two FC switches and then to two ports on a storage array, you
don't want everything going to the first switch.  Just picking one path to
use at random would be preferable (unless you want to swap every other
servers cables around).

One thing I know SDD does is that it doesn't use a path which has started
working again until it's been working for a little while.  I had a quick
newbie grok through the code and couldn't see anything like that.  Just an
idea.

Just doing a quick read I came across this comment:

+ * TODO: now if there are 2 multipaths in the same 2 devices, performance
+ * degrades dramatically because position is multipath, not device based.
+ * This should be changed to be device based. Also atomic sequential
+ * reads should be somehow balanced.

shouldn't that read "more than one multipath to the same device"?

I presume it's not limited to 2 multipaths?

Thanks

Adrian

Email: adrian.bridgett@iname.com
Windows NT - Unix in beta-testing. GPG/PGP keys available on public key servers
Debian GNU/Linux  -*-  By professionals for professionals  -*-  www.debian.org

  reply	other threads:[~2001-09-18 19:50 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2001-09-14 10:01 [patch] multipath RAID personality, 2.4.10-pre9 Ingo Molnar
2001-09-16 21:08 ` Andreas Dilger
2001-09-16 23:16   ` Ingo Molnar
2001-09-18 19:39     ` Adrian Bridgett [this message]
2001-09-18 20:59       ` Ragnar Kjørstad
2001-09-18 21:22       ` Brian Beattie
2001-09-24  8:30       ` Ingo Molnar

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=20010918203946.A12814@wyvern \
    --to=adrian.bridgett@iname.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-raid@vger.kernel.org \
    --cc=mingo@elte.hu \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox