[DRBD-user] uuid, label or equivalent in drbd.conf

Admin drbd-info at internode.com.au
Fri Apr 7 09:14:30 CEST 2006

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


Hi.

I'm testing various failure modes with drbd svn 2141 + 2.6.16 + vserver
2.1.1rc15 amd64
I'm not sure why the svn version numbers differ between what it said when I 
did an svn update and what /proc/drbd says. I essentially copied the svn 
tree, removed the .svn directories turned that into a tarball which I pointed 
the 0.7.17 ebuild at (gentoo machines)

With a sun v20z (LSI Logic / Symbios Logic 53c1030 scsi controller with
hardware raid disabled), I'm having a problem with scsi /dev nodes being
swapped around.

I'm using the following config

           machine1                           machine2
    sda1 - md1 - sdb1               sda1 - md1 - sdb1 /boot
    sda2 - md2 - sdb2               sda2 - md2 - sdb2 swap
    sda3 - md3 - sdb3               sda3 - md3 - sdb3 /
               sda4 --------drbd1 --------- sda4
               sdb4 --------drbd2 --------- sdb4

The idea is to maximise available disk space (in a machine that can only take
2 disks) whilst retaining resiliency in the case of disk failure and reducing
io contention. I'm running vservers on drbd1 and drbd2 with heartbeat
controlling everything.

The problem is that booting without the first disk (normally sda) plugged
in/functional results in sdb being transposed to sda and the drbd.conf gets
messed up (since it refers to sda4 and sdb4).

The linux software raid has been configured to use UUIDs so it works as
expected.

Is there a way to tell drbd to identify a block device some way other than
by /dev/<node>? I'm assuming that evms volume labels should be the way to go
but I can't get that to work (see below).

Reinserting the disk and rebooting resulted in a full resync of drbd1 (the
machine1 halves of drbd1 and drbd2 were broken whilst the machine
was booted with only the second disk inserted).

machine1 ~ # /etc/init.d/drbd start
 * Starting DRBD ...Can not open lower device '/dev/sdb4': No such device or 
address
Command '/sbin/drbdsetup /dev/drbd2 disk /dev/sdb4 internal -1 
--on-io-error=detach' terminated with exit code 20
                                                      [ ok ]
machine1 ~ # cat /proc/drbd 
version: 0.7.17 (api:77/proto:74)
SVN Revision: 2125 build by root at localhost, 2006-04-05 10:34:58
 0: cs:Unconfigured
 1: cs:StandAlone st:Secondary/Unknown ld:Consistent
    ns:0 nr:0 dw:0 dr:0 al:0 bm:0 lo:0 pe:0 ua:0 ap:0
 2: cs:Unconfigured

I've tried forcing the correct device node for sda and sdb with udev rules by
scsiid, but it seems that the mpt driver makes the choice when initialising
and doesn't like being overridden.

I've also tried turning sda4 and sdb4 into evms volumes, but for whatever
reason it isn't making them appear in /dev/evms (possibly something to do
with the reset of the disk being controlled by mdraid). Perhaps I'll have to
go with initrd images (which I'd rather avoid).

Thanks for any insights.
Admin




More information about the drbd-user mailing list