Hello-

I'm using heartbeat 3.0.2 with DRBD on Ubuntu 10.04, to manage MySQL. When I 
fail from the primary (mysqlha-lucid1) to the secondary (mysqlha-lucid2) node 
(either manually with hb_takeover or by powering off the primary), the DRBD 
device gets mounted properly on the secondary, but MySQL doesn't start, despite 
logging on the secondary that suggests it is running the init script:

May 19 16:15:06 mysqlha-lucid2 ResourceManager[1539]: info: Running 
/etc/init.d/mysql  start

FYI, the MySQL binaries and data all live on the DRBD device. Strangely, when I 
fail back to the primary, MySQL starts without a problem. The configs from 
mysqlha-lucid2 are below:

ha.cf:
autojoin none
logfacility local0
keepalive 500ms
deadtime 10
warntime 5
initdead 30
# IP of other node's interface on the crossover
ucast eth1 192.168.1.111
auto_failback off
node mysqlha-lucid1
node mysqlha-lucid2
ping 10.103.11.1
respawn hacluster /usr/lib/heartbeat/ipfail

haresources:
mysqlha-lucid1 10.103.11.113 drbddisk Filesystem::/dev/drbd1::/db01::ext3 mysql 
postfix

Incidentally, you'll notice that I also have postfix listed in haresources, and 
it starts just fine on both nodes. Strangely, doing this same thing on Ubuntu 
8.04 with heartbeat 2.1.4 seems to work perfectly. 

Does anyone have any suggestions on why mysql may not be starting on 
mysqlha-lucid2?

Joel 
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to