Hi all,
With my heartbeat errors dealt with, I am back thumping my head
against the wall trying to get DRBD8 working. Unfortunately, The website
for DRBD on Linux-HA.org is no help. The section on a DRBD config is a
'TODO' entry. Alas...
I have been documenting everything I've been doing and am more than
happy to update the website's config once I figure it out (or rather,
provide my docs to the site maintainer). Could you guys help me once
more to try and get this working?
In short, I can create the device, format the primary and sync to the
slave. Though once I down the master and mount the DRBD device on the
slave I get the no filesystem error, so obviously it's not actually
sync'ing.
Here are my notes detailing everything I have done:
-=-=-] Notes with mistake(s) [-=-=-
### On each nodes:
# Create the '/dev/drbd' device. This assumes the resource in question
# is named 'r0' in '/etc/drbd.conf'.
drbdadm create-md r0
# If you are asked: 'Valid meta-data already in place, recreate new?'
# enter:
yes
# Examples:
# drbdsetup /dev/drbd0 disk /dev/sda3 /dev/sda3 internal \
# --set-defaults --create-device --on-io-error=detach
# drbdsetup /dev/drbd0 disk /dev/cciss/c0d0p3 /dev/cciss/c0d0p3 \
# internal --set-defaults --create-device --on-io-error=detach
drbdsetup /dev/drbd0 disk /dev/<device> /dev/<device> internal \
--set-defaults --create-device --on-io-error=detach
# Make the directory you will mount the DRBD partition on
mkdir /ha
# On the master node:
drbdsetup /dev/drbd0 primary -o
/etc/ha.d/resource.d/drbddisk r0 start
# Now initialize the array:
drbdadm -- --overwrite-data-of-peer primary all
# Now on the master server, format the partition:
# change the 'mkfs...' command to match the FS you want. If it's not
# <which?> filesystems then DON'T mount the FS on the slave!
mkfs.ext3 /dev/drbd0
# On each terminal issue:
drbdadm -d adjust r0 # Does a 'd'ry-run and shows you what
# you need to do.
<run the commands given by the above call>
<Wait some time while the disks sync. Monitor by examining '/proc/drbd'>
# Watch the sync with this pseudo-script:
perl -e '$foo=1; while($foo == 1 ) { system "clear"; print "Status on:
"; system "echo `uname -n`"; system "cat /proc/drbd"; sleep 1; }'
# Now edit the '/etc/ha.d/haresources' and edit the entry to read
# something like:
vash 192.168.2.13 drbddisk::r0 Filesystem::/dev/drbd0::/ha::ext3
postgresql-8.1 apache2
# Each machine add the following lines to '/etc/rc.local'.
# If necessary set this file to 'chmod 755':
# The last line in '/etc/rc.local' must be 'exit 0'!
# Master:
drbdsetup /dev/drbd0 disk /dev/cciss/c0d0p3 /dev/cciss/c0d0p3 internal
--set-defaults --create-device --on-io-error=detach
drbdsetup /dev/drbd0 primary -o
mount /dev/drbd0 /ha
# Slave
drbdsetup /dev/drbd0 disk /dev/sda3 /dev/sda3 internal --set-defaults
--create-device --on-io-error=detach
drbdsetup /dev/drbd0 secondary
-=-=-] Notes with mistake(s) [-=-=-
Can anyone see/point out where I am going wrong? I can remount the
DRBD partition on the master and the data is still there... It just
seems to not want to sync to the slave. At this point, here is the
output of '/proc/drbd' on the:
# Master:
Status on: vash
version: 8.0.3 (api:86/proto:86)
SVN Revision: 2881 build by [EMAIL PROTECTED], 2007-06-21 15:05:15
0: cs:Connected st:Primary/Secondary ds:UpToDate/UpToDate C r---
ns:2222564 nr:0 dw:1694000 dr:2200083 al:1448 bm:2450 lo:0 pe:0
ua:0 ap:0
resync: used:0/31 hits:137428 misses:756 starving:0 dirty:0
changed:756
act_log: used:0/257 hits:422052 misses:1832 starving:0
dirty:384 changed:1448
# Slave:
Status on: minuet
version: 8.0.3 (api:86/proto:86)
SVN Revision: 2881 build by [EMAIL PROTECTED], 2007-06-21 15:04:42
0: cs:Connected st:Secondary/Primary ds:UpToDate/UpToDate C r---
ns:0 nr:2222564 dw:2222564 dr:0 al:0 bm:756 lo:0 pe:0 ua:0 ap:0
resync: used:0/31 hits:137428 misses:756 starving:0 dirty:0
changed:756
act_log: used:0/257 hits:0 misses:0 starving:0 dirty:0 changed:0
Thanks!!
Madison Kelly
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
