Container dies after a few days

[samcro1967]

I have noticed the container dies after running for a few days. Removing the container and adding it back gets it working again. Any ideas?

2023-03-03 10:59:55,476 INFO success: movary entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)2023-03-03 11:00:02,486 WARN received SIGQUIT indicating exit request
2023-03-03 11:00:02,487 INFO waiting for syslogd, nginxd, php-fpmd, crond, movary to die
2023-03-03 11:00:03,491 INFO stopped: movary (terminated by SIGTERM)
2023-03-03 11:00:05,493 INFO waiting for syslogd, nginxd, php-fpmd, crond to die
2023-03-03 11:00:05,494 INFO stopped: crond (terminated by SIGTERM)
2023-03-03 11:00:05,494 INFO reaped unknown pid 15200 (exit status 0)
2023-03-03 11:00:05,494 INFO reaped unknown pid 15201 (exit status 0)
[03-Mar-2023 11:00:05] NOTICE: Terminating ...
[03-Mar-2023 11:00:05] NOTICE: exiting, bye-bye!
2023-03-03 11:00:06,803 INFO stopped: php-fpmd (exit status 0)
2023-03-03 11:00:06,914 INFO stopped: nginxd (exit status 0)
[SYSLOG] syslog-ng[56]: syslog-ng shutting down; version='3.38.1'
2023-03-03 11:00:08,084 INFO stopped: syslogd (exit status 0)
error: failed switching to "root": operation not permitted
error: failed switching to "root": operation not permitted
error: failed switching to "root": operation not permitted

[leepeuker]

This is rooted somewhere in the base image (webdevops/php-nginx) we are using, see a discussion about it e.g. here

I am not sure how to solve this besides recreating the container. The current plan is to rework the docker image which should solve this problem.

[fscorrupt]

i have same problem with new installation: "error: failed switching to "root": operation not permitted" im using the nightly build, is there a fix/workaround?

[leepeuker]

Hey, the current fix is to recreate (not just restart!) the container. With docker compose a down + up should do it, without compose you have to stop + remove the container and then start it again.

If you use only managed volumes and not bind them manually from your host you should not get this problem.

[slimshizn]

Hey just noticed this is still a thing. I had to manually remove the container, and then docker-compose up -d movary to have it work again. Is this being worked on?

[samcro1967]

I am just running the script below every 30 mins via cron to look for the error, send an email, remove the container, and re-create the container:

``

!/bin/bash

ERROR_COUNT=$(docker logs movary --since 30m 2>&1 | grep 'operation not permitted' | wc -l) echo $ERROR_COUNT if [ "$ERROR_COUNT" != "0" ]; then echo "Movary operation not permitted, restarted movary" | sudo ssmtp user@mydomain.net docker rm -f movary cd /home/user/Documents/Docker && \ docker compose -f docker-compose.yml --env-file .env up -d movary fi

[leepeuker]

Is this being worked on?

Not currently sadly (I would love to, but did not find the time for it yet). Here is the main issue for this topic https://github.com/leepeuker/movary/issues/206 The rework of the docker image is the main blocker for version 1.0

[leepeuker]

@samcro1967 How many emails about errors do you get? For me the containers run without issues for weeks usually. would love to find out what is causing your problems

[samcro1967]

Looks like I have had 23 instances with the first being on March 7th, 2023. Seeing it about once a week, every once in awhile it is every 2 weeks.