Closed kajla closed 6 years ago
If your client application does not close the socket, or wait for processing the response, such errors occur. This is a clear indication of something going wrong, nothing I would call "spam".
Kind regards, Michael
Oh, I see. I used Thruk to SLA reports. It needs a LiveStatus. Do you think it doesn't close this socket? What can I do with it?
Thank you in advance.
I'd investigate on the time such queries take, extract what the application fires and debug that. You may enable Icinga 2's debug log to see more, but the rest really is network (sockets) and the remote application.
Cheers, Michael
Hi,
I have same problem in Icinga 2.6.2, with Thruk 2.18.
My icinga2.log conatins a lot of Livestatus query errors:
[2018-01-09 12:36:57 +0100] critical/LivestatusQuery: Cannot write query response to socket.
I haven't socket and thread pool error in the log, just this line. I have got same result in the debug.log.
My Thruk collected some information about the problem: [ERROR][Thruk] No Backend available [2018/01/08 10:10:16][ERROR][Thruk] on page: https://*********/thruk/cgi-bin/status.cgi?host=all&servicestatustypes=28&_=1515399710886 [2018/01/08 10:10:16][ERROR][Thruk] All: ERROR: failed to connect - Connection refused. (*****:6557) [2018/01/08 10:10:16][ERROR][Thruk] Error in: /thruk/cgi-bin/status.cgi [2018/01/08 10:10:16][ERROR][Thruk] failed to connect - Connection refused at /usr/share/thruk/lib/Monitoring/Livestatus/Class/Lite.pm line 380. at /usr/lib64/thruk/perl5/Plack/Util.pm line 142. eval {...} called at /usr/lib64/thruk/perl5/Plack/Util.pm line 142 Plack::Util::run_app('CODE(0x2da2040)', 'HASH(0x3bf7db0)') called at /usr/lib64/thruk/perl5/Plack/Handler/FCGI.pm line 143 Plack::Handler::FCGI::run('Plack::Handler::FCGI=HASH(0x10ecc48)', 'CODE(0x2da2040)') called at /usr/share/thruk/script/thruk_fastcgi.pl line 24
My httpd error.log content: broken pipe at /usr/share/thruk/lib/Thruk/Backend/Pool.pm line 30. Thruk::Backend::Pool::ANON('PIPE') called at /usr/lib64/thruk/perl5/Plack/Handler/FCGI.pm line 161 eval {...} called at /usr/lib64/thruk/perl5/Plack/Handler/FCGI.pm line 161 Plack::Handler::FCGI::run('Plack::Handler::FCGI=HASH(0x207dc48)', 'CODE(0x3d33050)') called at /usr/share/thruk/script/thruk_fastcgi.pl line 24 [Mon Jan 08 08:53:28.588033 2018] [fcgid:warn] [pid 76590] mod_fcgid: cleanup zombie process 112971 ERROR: got signal TERM while handling request, possible timeout in https://*******/thruk/cgi-bin/history.cgi?entries=100&start=2017-05-01+00:00:00&end=2018-01-09+00:00&archive=&host=****&service=Filesys$ at /usr/share/thruk/lib/Thruk.pm line 543.
My Thruk version is same; 2.18. We have different Icinga2 version with the same error. I think it's a Thruk issue. Am I correct, @dnsmichi ?
Thank you.
I'd say so.
Dear Developers,
I recently noticed that the Icinga2 LiveStatusQuery spam errors every minute to the icinga2.log file.
Current Behavior
Context
Spam errors to the log file. Socket file:
srw-rw---- 1 icinga icingacmd 0 Jan 8 13:07 /var/run/icinga2/cmd/livestatus=
Your Environment
Version used (
icinga2 --version
): r2.8.0-1Operating System and version: SUSE Linux Enterprise Server 12 SP3
Enabled features (
icinga2 feature list
): api checker compatlog ido-mysql livestatus mainlog notificationIcinga Web 2 version and modules (System - About): businessprocess | 2.1.0 director | 1.4.2 doc | 2.5.0 fileshipper | 1.0.1 monitoring | 2.5.0 translation | 2.5.0
Config validation (
icinga2 daemon -C
):/etc/icinga2/features-enabled/livestatus.conf file contents:
Thank you in advance.