[Linux-ha-dev] LRMd exits with no warning (+ logging preferences)

[Alan Robertson]

Andrew Beekhof wrote:
> I'm still getting the same behavior as the other day...
> 
> Dec 10 12:58:07 c001n09 heartbeat[3352]: info: Respawning client 
> "/usr/lib/heartbeat/lrmd":
> Dec 10 12:58:07 c001n09 heartbeat[3352]: info: Starting child client 
> "/usr/lib/heartbeat/lrmd" (0,0)
> Dec 10 12:58:08 c001n09 heartbeat[4006]: info: Starting 
> "/usr/lib/heartbeat/lrmd" as uid 0  gid 0 (pid 4006)
> Dec 10 12:58:08 c001n09 heartbeat[3352]: WARN: Exiting 
> /usr/lib/heartbeat/lrmd process 4006 returned rc 127.
> Dec 10 12:58:08 c001n09 heartbeat[3352]: ERROR: Client 
> /usr/lib/heartbeat/lrmd "respawning too fast"
> 
> 3 points:
> - (If you ask me) running the LRM without the stonithd running should be 
> supported
> 
> - If something is bad enough that the LRM needs to exit, its bad enough 
> that the logs should say why.
> 
> - If this is a permanent error of which there is no hope of recovery, 
> ever.  Then please consider exiting with rc = 100.  That way HA doesn't 
> waste time re-spawning the LRMd over and over.
> 
> On a related note, there really needs to be a way for the LRMd to use 
> the same logging preferences as everyone else does.  Currently this is a 
> little hard given that it doesn't (any pretty much has no need to) 
> connect to HA itself.
> 
> Perhaps a common cl_ function for determining the logging preferences 
> directly from ha.cf is required?  (Yes I can write it too, I first 
> wanted to know what people think of the idea).

I don't want to have other people reading ha.cf if at all possible.  In the 
future, we want to make the logging go through the logging daemon, so then 
logging preferences would automatically be shared.

In this case, it seems less necessary to read ha.cf.


-- 
     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