Documentserver: Daily server restarts by logrotate

[L3o-pold]

Every day in the /var/log/supervisor/supervisord.log I see that the server restart.
Using onlyoffice/documentserver-de:5.4.1.39

2020-01-27 06:35:54,187 WARN received SIGTERM indicating exit request
2020-01-27 06:35:54,188 INFO waiting for metrics, spellchecker, gc, docservice, converter to die
2020-01-27 06:35:54,190 INFO stopped: spellchecker (terminated by SIGTERM)
2020-01-27 06:35:54,190 INFO stopped: metrics (exit status 0)
2020-01-27 06:35:54,191 INFO stopped: converter (terminated by SIGTERM)
2020-01-27 06:35:54,195 INFO stopped: docservice (terminated by SIGTERM)
2020-01-27 06:35:54,234 INFO stopped: gc (terminated by SIGTERM)
2020-01-27 06:35:59,271 CRIT Supervisor running as root (no user in config file)

2020-01-26 06:32:48,147 WARN received SIGTERM indicating exit request

2020-01-25 06:40:51,297 WARN received SIGTERM indicating exit request

cat /etc/apt/apt.conf.d/20auto-upgrades

APT::Periodic::Update-Package-Lists "0";
APT::Periodic::Unattended-Upgrade "0";

Any idea?

[ShockwaveNN]

Does server restart at same time? Does your system have enought ram and swap?

[L3o-pold]

It restart every day around 6AM. Yeah enough RAM and swap, seems more like an automatic shutdown but I don't see why.

[ShockwaveNN]

So it's just a supervisor restart? Container doesn't restart itself?

[L3o-pold]

Exactly.

[ShockwaveNN]

Maybe remove unattended-upgrades altogether to rule it out? I have no other idea (

[L3o-pold]

Should be reproductible if you use the docker image, once a day.
I'll try to solve this but if you can reproduce and/or have any other idea, could be nice.

[ShockwaveNN]

Seems this is logrotate server restarts
You can modify settings in /etc/logrotate.d/ds.conf or disable by removing this file

[L3o-pold]

Nice find. I'll give it a try.

[ShockwaveNN]

@L3o-pold I'm still not sure what restarting server by logrotate is good idea on our side, but currently we'll not change it

[L3o-pold]

@ShockwaveNN Yeah I'm not sure reloading the server, is a good idea, cause it'll break connections...
Could you send me a link where this logrotate is configured in the docker image or server repo please?

[ShockwaveNN]

Could you send me a link where this logrotate is configured in the docker image or server repo please?

I'm not sure what do you mean. It configure in /etc/logrotate.d/ds.conf file as I mention earlier

[L3o-pold]

How is this file created?

[ShockwaveNN]

/etc/logrotate.d/ds.conf created while installing deb
It's located at https://github.com/ONLYOFFICE/document-server-package/blob/release/v5.5.0/common/documentserver/logrotate/ds.conf.m4 but sorry, this repo is private (

[L3o-pold]

Ok that why I can't find it. Feel free to close this (I don't recommend it because I don't see how nobody haven't see this issue) or not.

[ShockwaveNN]

@L3o-pold I'll leave issue open, because I'm not sure that it's a good idea to restart service daily, but currently do not how to elegant solve this problem

And I move this issue to DocumentServer repo, it has nothing to do with Docker

[hahaxd35]

maybe i can help:

@ShockwaveNN Yeah I'm not sure reloading the server, is a good idea, cause it'll break connections...

Run service nginx reload or /etc/init.d/nginx reload

If you have pending requests, then there will be lingering nginx processes that will handle those connections before it dies, so it's an extremely graceful way to reload configs.

but currently do not how to elegant solve this problem
mayby you have a look @ rsylog:
http://jensd.be/372/linux/write-to-and-manage-syslog-messages-with-logger-and-rsyslog

after logrotate you only need to reload rsylog (by default after logrote)

[L3o-pold]

I don’t think that an ideal solution because even using graceful stop on Nginx, document server are using websocket and it’ll disconnect the user.

[Githopp192]

i think .. i'm running into the same issue -->

2020-02-18 06:28:16,759 WARN received SIGTERM indicating exit request
2020-02-18 06:28:16,939 INFO waiting for metrics, spellchecker, gc, docservice, converter to die
2020-02-18 06:28:18,823 INFO stopped: gc (terminated by SIGTERM)
2020-02-18 06:28:21,345 INFO waiting for metrics, spellchecker, docservice, converter to die
2020-02-18 06:28:21,414 INFO stopped: spellchecker (terminated by SIGTERM)
2020-02-18 06:28:21,414 INFO stopped: docservice (terminated by SIGTERM)
2020-02-18 06:28:21,415 INFO stopped: converter (terminated by SIGTERM)
2020-02-18 06:28:22,730 INFO stopped: metrics (exit status 0)

/etc/crontab -->

m h dom mon dow user command

25 6 * * * root test -x /usr/sbin/anacron || ( cd / && run-parts --report /etc/cron.daily )

Three minutes later OnlyOffice will stop working .. every day .. same time