Hi Antony,
that's a known issue with sysvinit. See
https://dev.icinga.com/issues/13567 for more details.
Best
Tobias
On 2016-12-16 16:13, Antony Stone wrote:
Hi.
I have several machines being monitored by Icinga2 in a master -
satellite -
client hierarchy. They all run Debian Wheezy 7.11 and Icinga is
installed
from the debmon.org repository. So far they've all been running Icinga
2.5.4
Yesterday I got notifications that 2.6.0 was now available, so I
upgraded one
of the machines to check that it worked okay before upgrading
everything
across the whole network.
It didn't go entirely smoothly (specifically, npcd stopped because the
package
upgrade didn't automatically apply the IDO database schema update, so I
was
left without performance data until it was spotted that pnp4nagios
wasn't
updating as usual. This failure was remarkably silent.), so I held
back on
upgrading other machines until we're confident about the performance of
2.6.0
Today I'm noticing that the 2.6.0 machine (which is a satellite,
receiving its
updates from a 2.5.4 master) is stopping Icinga2 completely after the
configuration is changed on the master and icinga on the master is
reloaded.
Here's what I see in /var/log/icinga2/icinga2.log on the satellite
machine
(hostnames have been redacted for confidentiality) after doing
"/etc/init.d/icinga2 reload" on the master:
[2016-12-16 13:37:27 +0000] warning/TlsStream: TLS stream was
disconnected.
[2016-12-16 13:37:27 +0000] warning/JsonRpcConnection: API client
disconnected
for identity 'icinga2.master'
[2016-12-16 13:37:27 +0000] warning/ApiListener: Removing API client
for
endpoint 'icinga2.master'. 0 API clients left.
[2016-12-16 13:37:28 +0000] information/ApiListener: New client
connection for
identity 'icinga2.master' from [xx.214.yyy.106]:32914
[2016-12-16 13:37:28 +0000] information/ApiListener: Sending config
updates for
endpoint 'icinga2.master'.
[2016-12-16 13:37:28 +0000] information/ApiListener: Syncing runtime
objects
to endpoint 'icinga2.master'.
[2016-12-16 13:37:28 +0000] information/ApiListener: Finished syncing
runtime
objects to endpoint 'icinga2.master'.
[2016-12-16 13:37:28 +0000] information/ApiListener: Finished sending
config
updates for endpoint 'icinga2.master'.
[2016-12-16 13:37:28 +0000] information/ApiListener: Sending replay log
for
endpoint 'icinga2.master'.
[2016-12-16 13:37:28 +0000] information/ApiListener: Replayed 4
messages.
[2016-12-16 13:37:28 +0000] information/ApiListener: Finished sending
replay
log for endpoint 'icinga2.master'.
[2016-12-16 13:37:28 +0000] information/ApiListener: Updating
configuration
file: /var/lib/icinga2/api/zones/Archive//.timestamp
[2016-12-16 13:37:28 +0000] information/ApiListener: Updating
configuration
file: /var/lib/icinga2/api/zones/Archive-MCR//.timestamp
[2016-12-16 13:37:28 +0000] information/ApiListener: Updating
configuration
file: /var/lib/icinga2/api/zones/Customer//.timestamp
[2016-12-16 13:37:28 +0000] information/ApiListener: Updating
configuration
file: /var/lib/icinga2/api/zones/Customer//_etc/hosts.conf
[2016-12-16 13:37:28 +0000] information/ApiListener: Updating
configuration
file: /var/lib/icinga2/api/zones/global-templates//.timestamp
[2016-12-16 13:37:28 +0000] information/ApiListener: Restarting after
configuration change.
[2016-12-16 13:37:29 +0000] information/Application: Got reload
command:
Starting new instance.
[2016-12-16 13:37:30 +0000] information/Application: Received request
to shut
down.
[2016-12-16 13:37:30 +0000] information/Application: Shutting down...
[2016-12-16 13:37:30 +0000] information/CheckerComponent: Checker
stopped.
...and that's it. Icinga2 has stopped, and no more logging (or
monitoring)
occurs until I manually restart it on the satellite machine.
There doesn't appear to be any problem with the configuration update,
because
once I manually restart Icinga2 on the satellite, everything runs as
expected.
So, is this an incompatibility between a 2.5.4 master and a 2.6.0
satellite
(so I can fix it by upgrading the master), or is it a problem with
2.6.0 (so
I'd be better off downgrading my satellite and returning to a
fully-2.5.4
network)?
Is there anything else I can investigate or quote from log files to
track down
why the satellite "Received request to shut down" in the log file
above?
Thanks,
Antony.
--
Regards
Tobias
_______________________________________________
icinga-users mailing list
icinga-users@lists.icinga.org
https://lists.icinga.org/mailman/listinfo/icinga-users
