detect hanging gamin/fam server

[tcampbel]

16:43:13 < HCoyote> desai:  so, at some point we encountered this in the past and the symptom threw an error that looked like "Hit event timeout without getting any 
                    events; GAMIN/FAM problem?" 
16:43:26 < desai> yeah, we can detect it upon startup pretty easily
16:43:27 < HCoyote> in this case, i can't find an error that would indicate a gamin problem.
16:43:28 < desai> oooh
16:43:35 < desai> we could add a watchdog
16:43:44 < desai> that would be pretty easy, and would catch cases like this
16:43:49 < HCoyote> is there some sort of feedback loop then?
16:44:03 < desai> just keep a timestamp in the gamin driver for the last event seen

add a method to have bcfg2-server detect if gam_server has hung and let admin know that it needs to be restarted.

[somekool <…>]

encounter this error for the first time. it seems bcfg2 is already trapping this error. it should just mentionned the required fix more clearly. that is to kill gam_server process

[https://www.google.com/accounts/o8/id?id=AItOawnSjgovXZr-_V3vGkvMSR0pc5LDykRc1Nc]

We used to run into this problem a lot with Bcfg2 1.0, but I haven't seen it with either 1.1 or 1.2, so this may have been fixed already.

[solj]

Closing this since we now support pyinotify which should be _much_ more reliable.