premature munin-update termination due to read timeout from node?
Hello,
I'm running a current Debian 7.2 with munin 2.0.6-4+deb7u1.
There are a couple of plugins which take some time to finish due to the nature
of what they do, which is around 17-19 seconds average. There is about half a dozen
of these plugins.
Unfortunaely munin-update terminates almost all the time after around 72
seconds, which is at least 50 seconds early to what a complete and successful run
takes. Seems it loses patience with reading from the node.
munin-node.log:
2013/11/07-06:51:05 [26174] Service 'tke_FOOBAR' timed out.
munin-update.log:
2013/11/07 06:56:02 [FATAL] Socket read from vs2.gothschlampen.com failed. Terminating process. at /usr/share/perl5/Munin/Master/UpdateWorker.pm line 394
2013/11/07 06:56:02 [ERROR] Error in node communication with vs2.gothschlampen.com/127.0.0.1:4949: [FATAL] Socket read from vs2.gothschlampen.com failed. Terminating process. at /usr/share/perl5/Munin/Master/UpdateWorker.pm line 394
In munin-node.conf I tried to increase the timeout like this:
$ grep ^time /etc/munin/*.conf
/etc/munin/munin-node.conf:timeout 270
Which is undocumented, and doesn't help, nonetheless is the solution that comes
up most often if you google this kind of problem.
I also modified the call to munin-update inside munin-cron by appending
"--timeout=250" and/or "-timeout 250", which unfortunately doesn't seem to work
either.
Is munin really this broken in this regard, or is it with my otherwise rather
plain vanilla setup?
Best regards
Thomas
Reply to: