My initial thought is memory aswell, but I haven't had the time to check it yet.

I think that if the problem was related to libraries, I would see more of it, and not a random meltdown months appart.


On 04/06/2011 06:17 PM, Marco A P D'Andrade wrote:

Usually SIGPIPE have three causes:
1. linked library problems, after reinstall or upgrade
2. Memory problems - try to run one full test on your server memory


2011/4/6 Anders Synstad < <>>

    Hello all MON users,

    Last night, my MON server went a bit crazy, spamming down with
    "Caught SIGPIPE" alarms.

    According to the graphs I got of the MON server, there was no
    unusual activity related to number of processes, load, memory or io.

    The server hasn't logged anything unusual. I see no indications of
    running out of fildescriptors, or oomkiller going on a frenzy.

    After a reboot, server was back to normal again.

    Has anyone experienced MON going nuts and just sending "Caught
    SIGPIPE" on every probe it forks out? I've had this happen once or
    twice before during the last 2 years this server has been active.

    Or if anyone got any ideas what could be causing this. I know the
    basics of the "Caught SIGPIPE" error itself, but I'm struggling with
    finding out what has caused it.

    Anders Synstad
    Basefarm AS

    mon mailing list <>

mon mailing list

Reply via email to