All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Markus Baertschi" <markus@markus.org>
To: "linux-lvm@sistina.com" <linux-lvm@sistina.com>
Subject: Re: [linux-lvm] Devices seen twice - how does LVM handle ?
Date: Wed Sep 10 13:24:01 2003	[thread overview]
Message-ID: <E19x9cD-0006C5-SW@server2.hostpoint.ch> (raw)
In-Reply-To: <200309101538.54429.krienke@uni-koblenz.de>

Rainer,

I'm using fiber cards coming with a load-balancing driver and a SAN-Storage
box (IBM FastT 700) box. Unfortunately the storage-controlles got reconfigured
to accommodate more devices on other ports. This broke the configuration for me.
As it happens I found out a couple of months later when the new devices were
in production (We reboot out servers only when necessary). I can not fix the
configuration immediately and am trying to find out how to survive best
until I can fix the real problem it.

At the moment I'm pulling a fiber to have only one path as the installed LVM
version (0.9beta ?) just crashes if is sees two devices pointing to the 
same disk.
The most recent version (1.0.7) does not have this problem. It allows me
the keep both fibers connected. But I don't know how it handles this situation
and started this discussion to find out.

Markus


>It did not follow this discussion right from the start, so I might have mis=
>sed=20
>something...

>We are running 3 fileservers with raid devices attached by fibrechannel. Si=
>nce=20
>each divice can be reached by 2 seperate io paths we are using a multipath=
>=20
>setup based on the md interface (on top of the multipath md devices we=20
>finally run a software mirror (raid1)). This works just fine for our server=
>s=20
>(running suse8.2) .  At the beginning we also had trouble with lvm1 since i=
>t=20
>was able to see the volume on different levels of abstraction: lvm scanned=
>=20
>the real disks, it also scanned the md multipath metadevice and this probab=
>ly=20
>led to the bad behaviour that whenever creating new logical volumes the=20
>complete configuration was invalid and so all data lost.=20

>We then changed to lvm2 since it has the ability to specify which devices t=
>o=20
>scan for lvm metadata and since then everything works fine. =20

>Rainer



--
  Markus Baertschi             Phone: ++41 (21) 807 1677
  Bas du Rossé 14b             Fax  : ++41 (21) 807 1678
  CH-1163, Etoy                Email: markus@markus.org
  Switzerland                  Homepage: www.markus.org

  reply	other threads:[~2003-09-10 13:24 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-09-10  6:49 [linux-lvm] Can't install LVM 1.0.7 on Redhat Advanced Server 2.1 Please help Gene
2003-09-10  8:05 ` [linux-lvm] Devices seen twice - how does LVM handle ? Markus Baertschi
2003-09-10  8:21   ` Luca Berra
2003-09-10 11:29     ` Markus Baertschi
2003-09-10 11:58     ` Rainer Krienke
2003-09-10 13:24       ` Markus Baertschi [this message]
2003-09-10  8:15 ` [linux-lvm] Can't install LVM 1.0.7 on Redhat Advanced Server 2.1 Please help Luca Berra
2003-09-10  8:30   ` Gene

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=E19x9cD-0006C5-SW@server2.hostpoint.ch \
    --to=markus@markus.org \
    --cc=linux-lvm@sistina.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.