Plex Rock-On Dead

I rebooted my server yesterday. At the time of the reboot, server uptime was 71 days. Upon rebooting, the Plex server would not restart. The rockon page indicated that mkdir failed trying to make “shared_media” subdirectory in /mnt2/rockon_configs/btrfs/subvolumes/c1c0ff04b02da56d0feb3df5de5da949ea7f0622ecc1c07268bc58f9b397d733/data/

I attempted to create the folder manually and chowned it over to plex_user, and it still didn’t work. A second reboot of the server also didn’t fix it.

My troubleshooting time is very, very limited due to a 2 month old and a jealous 2 1/2 year old. Rather than try to root cause this, I decided to uninstall and reinstall the docker. Uninstallation removed the Plex Rock-On from the installed page, but did not add it back as an uninstallable option. So I decided to install Emby, but that failed to install.

At bash, command docker ps -a showed that Plex was still installed, so I issued command docker rm plex-linuxserver.io. This removed it from the list when I perform docker ps -a, but it still did not add it back to the Rock-On installable section.

I was able to get Plex working with a manual install from the RPM, but now Plex has lost all of my manual changes to Movies and TV Shows that identified incorrectly, or those which I manually adjusted sorting.

So, two questions:

  1. How do I get Plex to show up as an available Rock-On?
  2. How do I get my Plex customizations to my manual installation?

Edit: I may be slow to respond due to my limited time, but I will read every response and attempt to act on them.

@Noggin Hello again.

As a quick note re the Plex Rock-on not showing up in available even after having been uninstalled, did you try the “Update” button on the Roco-ons page?

Also the “docker rm” command removes a docker container / instance but maybe there is still an image / class that is holding up proceedings, they would be removed via “docker rmi”. So what is the output of:

docker images

However I think the main concern here and maybe the root cause of these issues (ie Emby also not installing) may be the rock-on root share itself, ie could you have pool trouble and it has mounted as read only for example?

Is rockon_configs your equivalent of the rock-ons-root as per: "The Rock-ons root"
http://rockstor.com/docs/docker-based-rock-ons/overview.html#the-rock-ons-root
in the docs? This also fits in with the current plex rock-on image doing an auto update whenever it’s started, maybe an issue with the rock-ons-root share caused this auto update to fail and consequently to the rock-on failing to start in the first place.

Future maintenance is going to be reduced by sorting the Plex Rock-on issue, where, once sorted, you can simply re-select your prior plex-config and plex-data share counterparts:
http://rockstor.com/docs/docker-based-rock-ons/plex-media-server.html
and consequently inherit your prior configs.

docker images showed plex as one of the images, so I used docker rmi linuxserver/plex and it appears to have been successfully removed. However, it still doesn’t show up in the available Rock-Ons if I hit the update button.

But I noticed something else, it appears that none of my rock-ons are working, even after restarting the service.

[root@holocron ~]# docker images
REPOSITORY           TAG                 IMAGE ID            CREATED             VIRTUAL SIZE
linuxserver/sonarr   latest              583641616531        9 months ago        347.4 MB
linuxserver/nzbget   latest              9b1f29921432        12 months ago       443.5 MB
kylemanna/openvpn    latest              04391d078597        12 months ago       11.03 MB
postgres             latest              b57a32e15dec        12 months ago       265.8 MB
bittorrent/sync      latest              aa7975f7a339        12 months ago       147.3 MB
pschmitt/owncloud    latest              a44ef8508cd0        12 months ago       478.9 MB
busybox              latest              307ac631f1b5        14 months ago       1.113 MB
[root@holocron ~]# docker ps
CONTAINER ID        IMAGE                COMMAND                  CREATED             STATUS                          PORTS                           NAMES
3d981c67efe9        linuxserver/sonarr   "/init"                  9 months ago        Up 52 seconds                   0.0.0.0:8989->8989/tcp          Sonarr
7a7d4234cacc        linuxserver/nzbget   "/sbin/my_init"          9 months ago        Up 4 seconds                                                    nzbget
abc07603e4e2        pschmitt/owncloud    "/usr/bin/run.sh"        11 months ago       Restarting (1) 21 seconds ago   80/tcp, 0.0.0.0:8084->443/tcp   owncloud
3bf14a6d8241        postgres             "/docker-entrypoint.s"   11 months ago       Restarting (1) 22 seconds ago   5432/tcp                        owncloud-postg
res
05f0838a100e        kylemanna/openvpn    "ovpn_run"               12 months ago       Up 52 seconds                   0.0.0.0:1194->1194/udp          openvpn

Owncloud keeps restarting. I’m not worried with that one right now as it has always been problematic and I’m on the verge of reinstalling it. Sonarr and NZBget are not working, despite the fact that they’re listed as running. If I attempt to open a web page to them, whether by IP address or by hostname, I get a connection refused page.

I don’t think the pool is mounted as read only, I was able to manually create the folder.

I did make rockon_config as the main rockon config folder, but to be honest I don’t recall exactly what I’ve configured it for. When I set up plex, I made a plex_config, plex_media, and plex_transcode when I configured Plex. Similar for Owncloud, Emby (which didn’t install so I deleted this one), NZBget, and Sonarr.

I have 6x2TB drives a 2x6TB drives all in a RAID 1 configuration all in a single pool. The last automated scrub ran on the 1st of this month and I’ve manually started another one now.

The scrub did not find any issues and I did confrm that rockon_config is my root rockon config folder.

I noticed at some point during the scrub that the Rock-on service wouldn’t start. I rebooted the server again and the Rock-on service still won’t start. Kids are sleeping, so I’m strongly considering nuking this and starting fresh with a new installation of Rockstor. I just don’t want to have to reconfigure everything again.

Nevermind. 2 month old is awake.


2 month old is either asleep or pretending to be asleep now.  Not sure which.  He's evil, so probably pretending and waiting for me to get busy before making his move.

I made a new Rock-on root share and configured the Rock-on service to use it.  Now, the rock-on service starts, but it didn't really "reset" the rock-on stuff like I was hoping it would.  Docker ps -a and docker images show that nothing is installed, as I had expected, but the web interface is still showing all of my previously installed Rock-ons, except Plex, as installed but unable to start.

I’m in the middle of reinstalling Rockstor and found the following:

  1. Installation will not proceed with my RAID array connected. Disabling the SATA controller in BIOS is sufficient to get the installation going.
  2. My previous activation code no longer works. I’ve requested a new one.
  3. I attempted to reinstall the Sonarr docker using the previous sonarr_config directory. This allowed the docker to install, but I couldn’t access it via the web interface.
  4. I uninstalled the docker and created a new share, sonarr-config. I installed the docker using this as the configuration directory. Again, it installed but I couldn’t access it via the web interface.
  5. I created a new share for NZB Get and installed it using that. I COULD access this via the web interface.
  6. I’m reinstalling again hoping that all of the Dockers work this time. I’ll be creating new config shares for each of them this time.

@phillxnet I think you were spot on with something being borked in the rock-ons root. After getting the system up and running again, I started having more issues. I then opened a shell and found that in some directories I could create files and in others I could not. The places I could not create files (using “touch test” and then removing the file) appeared to be shares I set up for the dockers. I’m not sure if that is expected or not, but I imagine that as root I should have been able to create a file pretty much anywhere.

After finding that some places were read-only, I then rebooted. After rebooting, one of my 2TB drives was detached and the pool wasn’t mounted. A scrub from a couple days ago didn’t find any issues (or at least it didn’t report them in any way that I saw the issue) and when I manually pulled the SMART data I didn’t see any issues either. But, the drive is now detached. I haven’t yet determined if it is detached due to a cable / power issue or if the drive has failed. I’m leaning towards a drive failure as it has nearly 30,000 hours on it.