For some unkown reason (to me at least) my ram keeps filling up to the point where I have to shut down VM:s or reboot my server. 
I´ve got no syslog-server running and I don´t understand why this is happening. 
Since I'm not fluent in logs, or even what to look for, I'm turning to you golden gentlemen's (and lady's) who have saved me so many times before. 

On a fresh reboot I've got between 69-72% ram utilization, but 48h later its up at 85%. 

Got 3 vms running and 17 dockers (normaly 19, but have still not gotten around to fix the fuckall nextcloud/Collabora thing) so they are offline. 
As I said, no syslog running in the background, and I really cant find what is filling up my ram over time. 
And before someone says it, I know I'm running kinda short on ram, and its next on my "to do/upgrade list". 

Diag was downloaded before I rebooted the machine. 

Since I got my UPS my "record" is 1 month 18 days, but had to shut down vms one after one as it filled up, and I tried solving it myself by looking through everything and googling but I lack the skills. 

 

 

(Fuuuuck me, I forgot to upload the diag, but here it is!)

 

belk-diagnostics-20220719-1824.zip

Edited July 19, 20223 yr by BeardElk

attach diagnostics to your next post in this thread.

 

setup syslog server

15 minutes ago, trurl said:

attach diagnostics to your next post in this thread.

 

setup syslog server

Diagnostics are attached. 
And why do I need to setup a syslog server? 
I get that I need to log as much as possible, but my main goal is to not have to reboot the server every week.

3 minutes ago, BeardElk said:

And why do I need to setup a syslog server? 

unRAID runs in RAM which includes the syslog.  Any reboot will wipe out the syslog and any useful information it may have captured regarding system issues.  The syslog server records this information elsewhere so it can be referenced after a reboot.

Just a note on using the forum. When you edit a previous post, nobody will know about it unless they intentionally look there. If a new post isn't made to the thread it might not even be visited again since we have already read it and we have a lot of other threads to deal with.

 

That is why I say

31 minutes ago, trurl said:

attach diagnostics to your NEXT post in this thread.

Sorry I didn't stress that before.

11 minutes ago, Hoopster said:

unRAID runs in RAM which includes the syslog.  Any reboot will wipe out the syslog and any useful information it may have captured regarding system issues.  The syslog server records this information elsewhere so it can be referenced after a reboot.

as explained at syslog server link

What indication do you have that RAM is filling up?

Im gonna start by apologizing for the lack of information on my og post.

I´ve started a syslog server now, and in a couple of days im gonna post the output of the syslog.

 

(The indication I got for my RAM filling up, is that my RAM is filling up after I start my server. If I boot with 16gb free, I´ve got 10gb free a day later. (Im exaggerating, english is a second language for me)

 

What are you looking at specifically that tells you?

On 7/23/2022 at 2:41 AM, trurl said:

What are you looking at specifically that tells you?

Memory utilization on the diag page. (edit, Dashboard, not diag, my thoughts were elsewhere when i typed that)

Edited July 28, 20223 yr by BeardElk

Here's an updated diag with a syslog over the past 6 days. 
I hope you guys can help me! 

belk-diagnostics-20220728-2214.zip syslog-192.168.1.111.log

41 minutes ago, BeardElk said:

Memory utilization on the diag page.

Do you mean the Dashboard page?

3 hours ago, trurl said:

Do you mean the Dashboard page?

Yeah sorry, my head was in another thought when i typed that😅 (Im from Sweden, and Im autistic, sometimes I just "wander off" and keep typing without checking what I just wrote..)

Edited July 29, 20223 yr by BeardElk

Anyone?

So, it was plex-docker that were filling up the ram, and I fixed it with a simple script to run every other day to just reboot plex-docker! 

That's not a fix, it's a workaround because you didn't figure out how to fix it.

 

The usual cause of filling rootfs (which is in RAM) is an application writing to a path that isn't mounted storage, such as a container with a host path that isn't a disk or user share. In the case of plex, transcode path would be a likely suspect.

 

Post docker run for plex, and also post new diagnostics just before that script runs.

 

You can monitor rootfs usage yourself from the command line with this:

df -h /

 

On 9/6/2022 at 3:46 PM, trurl said:

That's not a fix, it's a workaround because you didn't figure out how to fix it.

 

The usual cause of filling rootfs (which is in RAM) is an application writing to a path that isn't mounted storage, such as a container with a host path that isn't a disk or user share. In the case of plex, transcode path would be a likely suspect.

 

Post docker run for plex, and also post new diagnostics just before that script runs.

 

You can monitor rootfs usage yourself from the command line with this:

df -h /

 

 

 

So, I'm running binhex-plexpass docker, and I've followed everything in his guide to setup transcoding, I've got Key-1 /transcode, then added /transcode - path /tmp/plex-ramdisk (made a ramdisk to limit the amount it can use), and all that is working fine, the only thing that "solved" my problem was the "custom user script" to reboot it every morning. 

 

Here is docker run and diag :

 

docker run
  -d
  --name='Plex'
  --net='host'
  -e TZ="Europe/Berlin"
  -e HOST_OS="Unraid"
  -e HOST_HOSTNAME="BelK"
  -e HOST_CONTAINERNAME="Plex"
  -e 'TRANS_DIR'='/transcode'
  -e 'NVIDIA_DRIVER_CAPABILITIES'='all'
  -e 'NVIDIA_VISIBLE_DEVICES'=''
  -e 'UMASK'='000'
  -e 'PUID'='99'
  -e 'PGID'='100'
  -l net.unraid.docker.managed=dockerman
  -l net.unraid.docker.webui='http://[IP]:[PORT:32400]/web/index.html'
  -l net.unraid.docker.icon='https://raw.githubusercontent.com/binhex/docker-templates/master/binhex/images/plex-icon.png'
  -v '/mnt/user/Media/':'/media':'rw'
  -v '/mnt/user/Backup/Backup/Annan Media/Filmer/':'/marvel':'rw'
  -v '/mnt/user/Backup/Backup/Annan Media/Musik Videos/':'/musikvideor':'rw'
  -v '/tmp/plex-ramdisk':'/transcode':'rw'
  -v '/mnt/user/appdata/binhex-plexpass':'/config':'rw'
  --device='/dev/dri' 'binhex/arch-plexpass'

remove actual plexpass

belk-diagnostics-20220924-2201.zip

Edited September 24, 20223 yr by ChatNoir
I edited the plex pass out of the dockerrun