|
In article <d6e10810606282123v50ab13c3i5462b55dc02477cb@xxxxxxxxxxxxxx> you
wrote:
>
> [root@SMSCONV11 init.d]# ps -ef | grep Re
> root 1299 1239 0 12:04 ? 00:00:00 /bin/sh
> /usr/lib/heartbeat/ResourceManager takegroup interopOB2
> root 1343 1299 0 12:04 ? 00:00:00 /bin/sh
> /usr/lib/heartbeat/ResourceManager takegroup interopOB2
> ****************************************
>
> And then it never carry on to start interopOB1 and ldirectord on
> SMSCONV11. I previously have the same problem with heartbeat version
> 1.2.3, but someone told it has been fixed in 1.2.4, since I can't get
> 1.2.4's rpms for RH EL3, so I upgraded to heartbeat 2.0.2.
>
> Is it a known bug for heartbeat 2.0.2 as well? How can I fix it or if
> I have configured something wrongly?
It is very important that your resource, interopOB2, behaves in the
expected maner. In particular, it should exit after it
has started the resource. It seems that it is not doing this.
Heartbeat runs the resources one by one (in reverse order) and
waits for each one to finish. It looks like interopOB2 is not
returning, and thus heartbeat is left waiting forever, and never
proceeds to starting the next resource.
In your case I suspect that you need to add some code to
your resource so it detaches itself from the controlling terminal,
or in otherwords daemonises itself. Also, please make sure that the
resource supplies stop satart and status actions, and that the
behaviour of those actions conforms with the LSB specification.
In particular these actions should be idempotent.
http://refspecs.freestandards.org/LSB_3.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html
--
Horms
H: http://www.vergenet.net/~horms/ W: http://www.valinux.co.jp/en/
|
