[Zenoss-dev] Re: watchdog problems after upgrade from 2.1.2 to 2.1.7.0

[Eric Newton]

pallium wrote:
> I've done the above removal of "--watchdog" from $ZENHOME/bin/zenfunctions however it appears that watchdog is still during zenhub and having a similar issue.
> 
> This only occured when adding 150+ devices via command line and renaming each right after the modeling was completed.
> 
> exerpt from zenhub.log:

if you remove --watchdog, shutdown everything:

	$ pkill -f /z

Then start up zeo, and zenhub:

	$ zeoctl start
	$ zenhub start

Then do a ps -elf for me.  You shouldn't see any watchdog gunk in any of
the command lines.  If you do, please let me know.

> 
> 2008-01-29 13:23:17 ERROR watchdog: (11, 'Resource temporarily unavailable')
> Traceback (most recent call last):
>   File "$ZENHOME/Products/ZenUtils/Watchdog.py", line 215, in run
>     self._runOnce()
>   File "$ZENHOME/Products/ZenUtils/Watchdog.py", line 159, in _runOnce
>     conn, addr = sock.accept()
>   File "$ZENHOME/lib/python2.4/socket.py", line 161, in accept
>     sock, addr = self._sock.accept()
> error: (11, 'Resource temporarily unavailable')

I know you are not crazy... others have seen this behavior, too.
Somehow two processes end up trying to listen to a child.  I have traced
through the code and I don't yet understand how this is possible.

In the 2.1.71, the watchdog will be turned off for everything but
zenwin, zenwinmodeler and zeneventlog (the windows collectors).

-Eric