[Linux-HA] heartbeat 2.0.8: hb + drbd stability

[greno at verizon.net]

I've been running heartbeat w/DRBD and I've been seeing some stability issues.  I use heartbeat to manage an IP, a DRBD device, a filesystem and an application that relies on the filesystem.  Everything works great for quite a while but every so often the application will show 'fail' in the GUI and that starts a chain of event where the filesystem goes 'not running' then the DRBD device goes 'not running' and finally the IP goes 'not running'.  A few minutes later the app goes 'unmanaged'.  Now I can clear this up by doing a 'cleanup' on the app and then everything automatically restarts and is once again fine for quite a while but later (several hours) when using the app the same sequence starts all over again.  What's puzzling is why heartbeat starts to take everything down.  Nothing depends on the app so why do these other resources disappear?  It's hard to show you the sequence in the GUI but if you've used it you know what it would look like to see this sequence happening.