[Linux-ha-dev] Shutdown process
alanr at unix.sh
Sat Jun 19 08:25:00 MDT 2004
Lars Marowsky-Bree wrote:
> On 2004-06-18T18:51:33,
> Andrew <lists at beekhof.homeip.net> said:
>>The second is slightly harder and I dont really have an answer yet.
>>One option might be to have the CRMd start and stop the LRM... there
>>was originally talk that it would be around before the cluster started
>>but that seems to have been dropped which might make this option
>>Other than that I'll keep thinking.
> Another option would be to have heartbeat, lrmd and crmd be started &
> stopped via regular lsb init scripts; then on shutdown, the init
> subsystem would take care of the proper ordering, like it does for a lot
> of other stuff.
> For runtime monitoring, they would register with apphbd.
> Alas, this would make shutting down "just the cluster stuff" harder than
> just a "/etc/init.d/heartbeat stop".
> If not, we need to copy this concept into heartbeat, and start/stop the
> other layers in the right order too.
This latter idea is very similar in some respects to something I've talked
about before in OCF. (But I've forgotten what I called it... :-()
Alan Robertson <alanr at unix.sh>
"Openness is the foundation and preservative of friendship... Let me claim
from you at all times your undisguised opinions." - William Wilberforce
More information about the Linux-HA-Dev