Not all dockers starting after restart

What is the output of

cat /var/lib/docker/unraid-autostart

 

root@Unraid-Server:~# cat /var/lib/docker/unraid-autostart
duplicati
DDclient
pihole
SteamCache-DNS
SteamCache
SteamCacheSnipproxy
mariadb
nextcloud
radarr
Sonarr
heimdall
hydra
JDownloader2
nzbget
PlexMediaServer
plexrequests
Xeoma
Storj
Storj1
Storj2
Storj3
Storj4
Storj5
Storj6
Storj7
Storj8
Storj9
Storj10

Edited May 4, 2018 by nuhll

You are using the "CA Docker Autostart Manager" to start the containers, not the standard unRAID startup sequence.

 

No, i had it, but i removed it.

 

edit:

doubelchecked it, plugin is not installed and it doesnt show up under dockers.

Edited May 4, 2018 by nuhll

Your syslog tells otherwise

Apr 30 13:27:36 Unraid-Server root: +==============================================================================
Apr 30 13:27:36 Unraid-Server root: | Installing new package /boot/config/plugins/ca.docker.autostart/ca.docker.autostart-2018.03.22-x86_64-1.txz
Apr 30 13:27:36 Unraid-Server root: +==============================================================================
Apr 30 13:27:36 Unraid-Server root: 
Apr 30 13:27:36 Unraid-Server root: Verifying package ca.docker.autostart-2018.03.22-x86_64-1.txz.
Apr 30 13:27:36 Unraid-Server root: Installing package ca.docker.autostart-2018.03.22-x86_64-1.txz:
Apr 30 13:27:36 Unraid-Server root: PACKAGE DESCRIPTION:
Apr 30 13:27:36 Unraid-Server root: Package ca.docker.autostart-2018.03.22-x86_64-1.txz installed.

Please reboot your system and make fresh diagnostics.

Upps, sorry, yea, that was the wrong one, i didnt realized that the diagnostic take so much time and closed page before it was downloading, i guess...

unraid-server-diagnostics-20180504-2033.zip

Same thing happened to me @bonienl with MariaDB, and I don't use autostart manager at all.   I'll post diags later.

 

And, autostart manager happens after emhttp does its starts and if the container is listed in the dynamix autostart file, then dynamix starts it.

The log shows all containers as succesfully started with docker attaching them to the network interface.

 

Are it always the same containers not starting or is this random?

 

It startet to happen first after i installed the autostart plugin. (but like i said, i removed it)

Edited May 4, 2018 by nuhll

today were 20+ updates to docker, again 4 docker didnt start.

Storj2

Storj7

Storj8

Storj9

So it seems its not docker specific, manual it startet just fine.

 

Could it be because theres no delay between starts of dockers? (or too low)

 

 

unraid-server-diagnostics-20180505-1423.zip

Edited May 5, 2018 by nuhll

Since I didn't have my diags from when this happened to me, and I'm unable to replicate my one-off issue under any circumstance I've thrown at it, I've got to tear apart yours:

 

Right here:

8 minutes ago, nuhll said:

today were 20+ updates to docker, again 4 docker didnt start.

 

is a key.

 

Your storj containers did start with the array.  (And this is the point at which @bonienl and @limetech will stop caring, and I will start ?

 

The update plugin stopped them

May  5 11:02:12 Unraid-Server Docker Auto Update: Stopping Storj2
.
May  5 11:03:54 Unraid-Server Docker Auto Update: Stopping Storj7
.
May  5 11:04:14 Unraid-Server Docker Auto Update: Stopping Storj8
.
May  5 11:04:34 Unraid-Server Docker Auto Update: Stopping Storj9

Ran the updates

May  5 11:05:14 Unraid-Server Docker Auto Update: Installing Updates for heimdall hydra mariadb nextcloud nzbget plexrequests radarr Storj Storj1 Storj2 Storj3 Storj4 Storj5 Storj6 Storj7 Storj8 Storj9 Storj10

And then restarted them

May  5 11:12:15 Unraid-Server Docker Auto Update: Restarting Storj2
.
May  5 11:12:17 Unraid-Server Docker Auto Update: Restarting Storj7
.
May  5 11:12:18 Unraid-Server Docker Auto Update: Restarting Storj8
.
May  5 11:12:19 Unraid-Server Docker Auto Update: Restarting Storj9

So far as I can tell they did start up successfully.

Additionally, I find it strange that docker.txt in the diags has a file length of zero bytes.  Never really seen that before if docker is running.

Yes, i know it looks like everything is working, thats why i post here. There seem to be a bug somewhere.

 

If you dont refresh the ui, it also looks like all dockers are running, but when u refresh, u see them red.. 

 

It should not be anything about the auto update, because it also happend after restart, without any update involved. (but different dockers)

Edited May 5, 2018 by nuhll

Then you should check the logs of the containers in question and see if they've exited for some reason.  Best I can say right now is that after the start of the array, they were running properly, because the autoupdate only stops a container that is running.

I guess i can only see them when i dont restart the docker and first look into log, or is there a other way?

 

But if its a problem with the dockers itself, why are it always others (atleast some) shouldnt it be always the same dockers, if there would be a problem with them?

Maybe there is a race condition with the storj* containers, do they all share the same appdata (or another) volume folder mapping?

all the same share, but different directorys like

/storj  /storj

/storj  /storj1

/storj  /storj2

 

while /storj/ is the share

and so on

Hey nuhll, I think the issue was the docker config from Jcloud. If you are using the non monitor version of storj, these dockers were always behaving erratically for me. The monitor version which was the one he created later works as expected for me.

On 5/25/2018 at 9:42 AM, MrChunky said:

Hey nuhll, I think the issue was the docker config from Jcloud. If you are using the non monitor version of storj, these dockers were always behaving erratically for me. The monitor version which was the one he created later works as expected for me.

 

I dont know. I use the same repo since start.

 

Since i added the restart cmd, it seems to work tho.

Changed Status to Closed