Kernel Panic using LVM2 over DRBD 8.0.6

[Dave Buechler]

Hi all,

Using a stock CentOS 5.0 install, I have been attempting to establish
replication of a large LVM2 partition (approximately 1TB).
Unfortunately, after creating drbd0 and completing the initial sync, the
primary node will kernel panic referencing drbd_req.c, line 374 -- I
think.  Forgive me, I don't have the panic message in front of me.

My configuration is as follows:

a partition on sdb1 on both nodes, with a type 8e (LVM)
create metadata on /dev/drbd0
make a node primary with -o flag
finish syncing (or not - it doesn't appear to make any difference)
create physical volume on /dev/drbd0 -- kernel will panic.

Can anyone tell me what I may be doing wrong?  I've been researching the
problem without much luck.  Any insight would be very helpful.

Thanks for your time.

I'm attaching copies of my config files below.

--
Regards,
David A. Buechler
System Administrator,
CableAmerica Missouri



drbd.conf:

global {
    usage-count no;
}

common {
  syncer { rate 110M; }
}

resource r0 {

  protocol C;

  handlers {
    pri-on-incon-degr "echo o > /proc/sysrq-trigger ; halt -f";
    pri-lost-after-sb "echo o > /proc/sysrq-trigger ; halt -f";
    local-io-error "echo o > /proc/sysrq-trigger ; halt -f";
  }

  startup {
    degr-wfc-timeout 90;    # 2 minutes.
  }

  disk {
    on-io-error   detach;
  }

  net {
    sndbuf-size 256k;
    max-buffers     8192;
    unplug-watermark   32768;
    max-epoch-size  5;
    after-sb-0pri disconnect;
    after-sb-1pri disconnect;
    after-sb-2pri disconnect;
    rr-conflict disconnect;
  }

  syncer {
    rate 110M;
    al-extents 257;
  }

  on node1 {
    device     /dev/drbd0;
    disk       /dev/sdb1;
    address    172.16.1.1:7788;
    meta-disk internal;
  }

  on node2 {
    device    /dev/drbd0;
    disk      /dev/sedb1;
    address   172.16.1.2:7788;
    disk internal;
  }
}



lvm.conf:

devices {
    scan = [ "/dev" ]
    filter = [ "r|/dev/cdrom|", "r|/dev/sdb1|", "a|/dev/drbd0|" ]
    cache = "/etc/lvm/.cache"
    write_cache_state = 1
    types = [ "drbd", 16 ]
    sysfs_scan = 1
    md_component_detection = 1
}

log {
    verbose = 0
    syslog = 1
    overwrite = 0
    level = 0
    indent = 1
    command_names = 0
    prefix = "  "
}

backup {
    backup = 1
    backup_dir = "/etc/lvm/backup"
    archive = 1
    archive_dir = "/etc/lvm/archive"
    retain_min = 10
    retain_days = 30
}

shell {
    history_size = 100
}

global {
    umask = 077
    test = 0
    activation = 1
    proc = "/proc"
    locking_type = 1
    fallback_to_clustered_locking = 1
    fallback_to_local_locking = 1
    locking_dir = "/var/lock/lvm"
}

activation {
    missing_stripe_filler = "/dev/ioerror"
    reserved_stack = 256
    reserved_memory = 8192
    process_priority = -18
    mirror_region_size = 512
    mirror_log_fault_policy = "allocate"
    mirror_device_fault_policy = "remove"
}



--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

_______________________________________________
drbd-user mailing list
drbd-user@...
http://lists.linbit.com/mailman/listinfo/drbd-user

[N.J. van der Horn (Nico)]

Hi Dave,

Just a short guess: On node2 you have in drbd.conf  /dev/sedb1 as the
physical device, is that correct ?

Nico

Dave Buechler schreef:

> Hi all,
>
> Using a stock CentOS 5.0 install, I have been attempting to establish
> replication of a large LVM2 partition (approximately 1TB).
> Unfortunately, after creating drbd0 and completing the initial sync, the
> primary node will kernel panic referencing drbd_req.c, line 374 -- I
> think.  Forgive me, I don't have the panic message in front of me.
>
> My configuration is as follows:
>
> a partition on sdb1 on both nodes, with a type 8e (LVM)
> create metadata on /dev/drbd0
> make a node primary with -o flag
> finish syncing (or not - it doesn't appear to make any difference)
> create physical volume on /dev/drbd0 -- kernel will panic.
>
> Can anyone tell me what I may be doing wrong?  I've been researching the
> problem without much luck.  Any insight would be very helpful.
>
> Thanks for your time.
>
> I'm attaching copies of my config files below.
>  

--
Behandeld door / Handled by: N.J. van der Horn (Nico)
---
ICT Support Vanderhorn IT-works, www.vanderhorn.nl,
Voorstraat 55, 3135 HW Vlaardingen, The Netherlands,
Tel [phone], Fax [phone]


_______________________________________________
drbd-user mailing list
drbd-user@...
http://lists.linbit.com/mailman/listinfo/drbd-user

[Florian Haas-3]

On Monday 03 December 2007 07:50:27 Dave Buechler wrote:

> Hi all,
>
> Using a stock CentOS 5.0 install, I have been attempting to establish
> replication of a large LVM2 partition (approximately 1TB).
> Unfortunately, after creating drbd0 and completing the initial sync, the
> primary node will kernel panic referencing drbd_req.c, line 374 -- I
> think.  Forgive me, I don't have the panic message in front of me.
>
> My configuration is as follows:
>
> a partition on sdb1 on both nodes, with a type 8e (LVM)

AFAIK there's no need to set the 8e partition type here. 83 will do.

> create metadata on /dev/drbd0
> make a node primary with -o flag
> finish syncing (or not - it doesn't appear to make any difference)
> create physical volume on /dev/drbd0 -- kernel will panic.
>
> Can anyone tell me what I may be doing wrong?  I've been researching the
> problem without much luck.  Any insight would be very helpful.
>
> Thanks for your time.
>
> I'm attaching copies of my config files below.
>   on node1 {
>     device     /dev/drbd0;
>     disk       /dev/sdb1;
>     address    172.16.1.1:7788;
>     meta-disk internal;
>   }
>
>   on node2 {
>     device    /dev/drbd0;
>     disk      /dev/sedb1;

                     ^^^^^

Is that a typo in your config, or a cut & paste error?

Mind you, none of these things should really be causing a panic. But it would
be helpful for us to know if when you fix the typo, the problem disappears.
Likewise if that doesn't help, change the partition type from 8e to 83 and
see if that makes any difference. Let us know your results.

Cheers,
Florian

--
: Florian G. Haas
: LINBIT Information Technologies GmbH
: Vivenotgasse 48, A-1120 Vienna, Austria
_______________________________________________
drbd-user mailing list
drbd-user@...
http://lists.linbit.com/mailman/listinfo/drbd-user

[Lars Ellenberg]

On Sun, Dec 02, 2007 at 11:50:27PM -0700, Dave Buechler wrote:
> Hi all,
>
> Using a stock CentOS 5.0 install, I have been attempting to establish
> replication of a large LVM2 partition (approximately 1TB).
> Unfortunately, after creating drbd0 and completing the initial sync, the
> primary node will kernel panic referencing drbd_req.c, line 374 -- I
> think.  Forgive me, I don't have the panic message in front of me.

in _my_ source code, drbd 8.0.6,
the line 374 does only contain an if statement.
the line 376 is a BUG_ON statement.

so maybe you want to be more precise about what version
you actually are using, and what that "panic" looks like?
or is is the BUG_ON statement?

could it be that the cause of the problem happen much earlier, and is
about some memory shortage, which you ignored?

what is in the kernel log,
from the start of the drbd until the "panic"?

>   net {
>     sndbuf-size 256k;
>     max-buffers     8192;
>     unplug-watermark   32768;
>     max-epoch-size  5;

???
that is a very small max-epoch size.
you don't want to do that.
but that is unrelated.

--
: commercial DRBD/HA support and consulting: sales at linbit.com :
: Lars Ellenberg                            Tel [phone]  :
: LINBIT Information Technologies GmbH      Fax [phone] :
: Vivenotgasse 48, A-1120 Vienna/Europe    http://www.linbit.com :
__
please use the "List-Reply" function of your email client.
_______________________________________________
drbd-user mailing list
drbd-user@...
http://lists.linbit.com/mailman/listinfo/drbd-user