Hi

I've had some trouble about the fiabiliy of the return of 
/etc/init.d/corosync status,
meaning that sometimes I got "started" whereas it was in fact "stopped" .
So I 've done more systematic tests and this leads to a systematic problem
about this status.

The problem is easy to reproduce :
Set the Pacemaker at state "stopped" with :
/etc/init.d/corosync stop
on one window on this node , launch the loop :
while true
do
/etc/init.d/corosync status
done

it should display a list of :
....
corosync is stopped
corosync is stopped
corosync is stopped
corosync is stopped
....

Now, from another window on same machine , or via ssh ,
just launch once :
/etc/init.d/corosync status
and you'll see :
....
corosync is stopped
corosync is stopped
corosync (pid 20303 20296) is running...
corosync is stopped
corosync is stopped
....

I know , this is a stupid test :-(      but nethertheless , that mean 
that the status target
of corosync is not completely reliable, due to "status $prog" in the 
init.d corosync script,
and I do this test because from time to time I got a bad status of corosync.
It seems to be more a problem of "status" (initctl executable)  than of 
corosync itself, but
that means that the status should perhaps be got in another way ?

Any opinion ?
Thanks
Alain Moullé


_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to