Recurring downtime is not sent to livestatus backend

[mkempken]

Description
Recurring downtime is not sent to Nagios livestatus backend while adding a manual downtime is working fine

Thruk Version
Version 2.34 + mk-livestatus-1.2.8p18

To Reproduce

1. setup multiple livestatus backends and schedule recurring downtimes

Expected behavior
Recurring downtime is send to the specific livestatus backend + nagios.log + thruk.log show corresponding messages:
"... cmd: COMMAND [1588153985] SCHEDULE_SVC_DOWNTIME ..."
"... cmd: COMMAND [1588153985] SCHEDULE_HOST_DOWNTIME ..."

Screenshots

[mkempken]

Short update, I just saw the my livestatus version is quite old. I will try mk-livestatus-1.5.0p24 now, may be it's related to that old version

[mkempken]

Same behavior with the latest version mk-livestatus-1.5.0p24.

• Manual and scheduled downtimes work fine
• Recurring downtimes are not send to the livestatus backend

This is the setup:

[sni]

A few things you could check:

• is the cronjob created correctly (as webserver user)
• anything in the system log regarding that cronjob, got it executed?
• anything in the thruk log?
• anything in the nagios.log?

[mkempken]

• is the cronjob created correctly (as webserver user)

/var/spool/cron/crontabs/www-data
THRUK_CRON=1
# downtimes
 3 15  *  *  * cd /usr/share/thruk && /bin/bash -l -c '/usr/bin/thruk downtimetask "2"' >/dev/null 2>>/var/lib/thruk/cron.log
 5 12  *  *  * cd /usr/share/thruk && /bin/bash -l -c '/usr/bin/thruk downtimetask "1"' >/dev/null 2>>/var/lib/thruk/cron.log
# general
20,50 * * * * cd /usr/share/thruk && /bin/bash -l -c '/usr/bin/thruk maintenance' >/dev/null 2>>/var/lib/thruk/cron.log

• Syslog of the thruk instance
  May 7 09:26:23 632135f64e03 crontab[8064]: (www-data) LIST (www-data)
  May 7 09:26:23 632135f64e03 crontab[8067]: (www-data) REPLACE (www-data)
• Thruk Log
  -> No entries for this event
• Nagios Log
  -> Nothing

BUT I've created a file to be used for cron tasks '/var/cache/thruk/thruk_cron.tmp' and changed the owner to www-data, when I save the recurring downtime entry in thruk the file disappears.

When I run the cron manually I get logs in /var/lib/thruk/cron.log

mesg: ttyname failed: Inappropriate ioctl for device

[sni]

You could edit the crontab manually with crontab -u www-data -e (as root) and change the >/dev/null 2>>/var/lib/thruk/cron.log into >>/var/lib/thruk/cron.log 2>&1 and if that still doesn't bring anything useful. Replace /usr/bin/thruk with /usr/bin/thruk -v to get verbose output.
For testing you could make it run every minute by using * * * * *.
You should see something like:

May  7 11:07:01 mo CRON[659033]: (sven) CMD (cd /opt/projects/git/Thruk && /bin/bash -l -c 'script/thruk downtimetask "5"' >/dev/null 2>>./var/cron.log)

in /var/log/syslog on debian or whereever that file is on other systems.

[mkempken]

Thanks, I will try that and let you know. So, I recreated my container, here is the outcome.

• Cronjob for www-data is created properly in /var/spool/crontab/www-data
• If I run the cronjob manually like cd /usr/share/thruk && /bin/bash -l -c '/usr/bin/thruk downtimetask "1"' >/dev/null 2>>/var/lib/thruk/cron.log the expected output in the corresponding nagios.log can be found

[1589368277] EXTERNAL COMMAND: SCHEDULE_HOSTGROUP_HOST_DOWNTIME;c0001;1589368277;1589371877;1;0;0;(cron);automatic downtime
[1589368277] HOST DOWNTIME ALERT: c0001-license;STARTED; Host has entered a period of scheduled downtime
[1589368277] HOST DOWNTIME ALERT: c0001-node1;STARTED; Host has entered a period of scheduled downtime
[1589368278] HOST DOWNTIME ALERT: c0001-node2;STARTED; Host has entered a period of scheduled downtime
[1589368278] HOST DOWNTIME ALERT: c0001-node3;STARTED; Host has entered a period of scheduled downtime
[1589368278] HOST DOWNTIME ALERT: c0001-node4;STARTED; Host has entered a period of scheduled downtime
[1589368278] HOST DOWNTIME ALERT: c0001-node5;STARTED; Host has entered a period of scheduled downtime

Last question, when is the downtime task executed by Thruk, when I click on "Save" or when cron is executed? My expectation was that the command is sent to Nagios when I click on "Save".

[sni]

The downtime task is triggered by cron. When you hit save it only creates the new crontab, nothing else. Nagios does not know anything about recurring downtimes.
If this is a container, make sure the cron daemon is actually running because this is usually not the case.