[rabbitmq-discuss] Rabbitmq 2.8.2 with heartbeat/DRBD fail

Thu May 17 03:43:49 BST 2012

Hi,

i build a simple highly available environment in rabbitmq 2.7.1/Erlang
R14B03 with heartbeat 3.0.4/DRBD 8.3.11,and it's work fine and stable,
last  week i upgrade rabbitmq to 2.8.2 , when i test heartbeat
takeover,it's can take ip and drbd disk to another node,but mq service
can't start ,i try to use "/etc/init.d/rabbitmq-server start" ,mq
service can start success, i check /etc/init.d/rabbitmq-server
script,it's only different in 3 point,

in 2.7.1
RABBITMQ_PID_FILE=$PID_FILE setsid $DAEMON

in 2.8.2
START_PROG="runuser rabbitmq --session-command"

[ -f /etc/default/${NAME} ] && . /etc/default/${NAME}

RABBITMQ_PID_FILE=$PID_FILE $START_PROG $DAEMON

i check the initial proecss,when heartbeat takeover,mq service will
check status first,and run start_rabbitmq function,and i add logger
print some checkpoint to minitor start script, i found it didn't run
to start_rabbitmq function,

i try to use 2.7.1 start script,and 2.8.2 can manual start success,but
heartbeat takeover still start fail,

my os:centos5.5 x64

my ha.cf :

debug 1
debugfile /var/log/ha-debug
logfile /var/log/ha-log
logfacility local0
udpport 694
keepalive 1
deadtime 10
initdead 80
ucast eth1 10.19.100.4
respawn hacluster /usr/lib/heartbeat/ipfail
ping 172.19.1.252
node mq1
node mq2
auto_failback off

my haresource:

mq1 172.19.253.99/16/eth0 \
drbddisk::drbd1  Filesystem::/dev/drbd1::/media/drbd1 \
rabbitmq-server

Any help would be appreciated!
thanks for help!!