[Unraid 6.9.2 - Out Of Memory errors detected on your server]

Hello All,

Fix common problems has found OOM errors on my server. PFA the log output, I enabled mcelog via nerdpack as well. I'm very new to unraid and I'm uncertain if this is a normal CPU startup error or if I have faulty RAM. 

 

I got the error from Fix Common Problems on 17-07-2022 04:30

 

tortuga-diagnostics-20220717-1055.zip

[Would SSD Trim and the Mover ever interfere with each other?]

Would SSD TRIM and the mover ever interfere with each other? Currently they're set at the same time. 

[Unraid 6.9.2 - appdata best practices & weird mover issue]

Hello All.

I primarily use my server to host media with all the *arr applications. I have 2 issues. 

1) I don't think my server is setup optimally (due to point 2). I'm particularly worried my shares are not setup to use the cache drive correctly. 

2) When my mover runs, weird things happen. 

 

In regards to point 2. Last night I was downloading a movie from Radarr. It moved to Deluge and was in the process of getting downloaded when my mover ran. This morning when I check I can see the download is completed. (I also checked my /downloads/complete folder and it's there). However Radarr and Plex still do not see it. So it looks like Radarr did not move it / hardlink it from completed items to my /movies folder. Why is that? 

 

I suspect the issue may be because I set my shares to use my cache drive incorrectly?  I've read some conflicting things on the cache drive and how it should be configured for each share. I believe the appdata should be set to Cache - Only, and I've had it on Cache - Prefer. 

 

How can I fix this? 

 

 

 

tortuga-diagnostics-20220413-0948.zip

[v6.9.2 MCE Detected on your server - Fix Common Problems]

I get this error every time I reboot my server. Running fix common problems I get the below error:

 

Quote

Machine Check Events detected on your server. Your server has detected hardware errors. You should install mcelog via the NerdPack plugin, post your diagnostics and ask for assistance on the unRaid forums. The output of mcelog (if installed) has been logged.

 

I've now installed MCE via Nerdpack and collected the Diagnostics logs. Please find them attached. Below is a snippet.

 

Oct 16 12:24:09 Tortuga kernel: mce: [Hardware Error]: Machine check events logged
Oct 16 12:24:09 Tortuga kernel: mce: [Hardware Error]: CPU 0: Machine Check: 0 Bank 6: ee0000000040110a
Oct 16 12:24:09 Tortuga kernel: mce: [Hardware Error]: TSC 0 ADDR fef20080 MISC 3880000086 
Oct 16 12:24:09 Tortuga kernel: mce: [Hardware Error]: PROCESSOR 0:a0653 TIME 1634401431 SOCKET 0 APIC 0 microcode e0
Oct 16 12:24:09 Tortuga kernel: Performance Events: PEBS fmt3+, Skylake events, 32-deep LBR, full-width counters, Intel PMU driver.
Oct 16 12:24:09 Tortuga kernel: ... version:                4
Oct 16 12:24:09 Tortuga kernel: ... bit width:              48
Oct 16 12:24:09 Tortuga kernel: ... generic registers:      4
Oct 16 12:24:09 Tortuga kernel: ... value mask:             0000ffffffffffff
Oct 16 12:24:09 Tortuga kernel: ... max period:             00007fffffffffff
Oct 16 12:24:09 Tortuga kernel: ... fixed-purpose events:   3
Oct 16 12:24:09 Tortuga kernel: ... event mask:             000000070000000f

 

Any ideas what's wrong? I will run the memtest tonight, but kinda looks like CPU support issue.

tortuga-diagnostics-20211016-1244.zip

[Version: 6.9.2 - Lost all docker settings/config]

27 minutes ago, JonathanM said:

I don't think hardlinks work like that, I could be wrong but I think moving the source file breaks the hardlink. So if you need to use hardlinks I think you will have to set cache: no so the downloads go directly to the array.

 

I don't use hardlinks, so maybe there is a better way of doing things that I'm not aware of. The user share system complicates hardlinks greatly.

 

@JonathanM If you don't use hardlinks, do you have your media copied from your downloads (which is on your cache drive) over to your media share (on your array). Then how do you go about deleting the original file from your cache? Is their a setting in somewhere for once the file is copied to your array the original gets deleted?

[Version: 6.9.2 - Lost all docker settings/config]

3 minutes ago, JonathanM said:

Shouldn't matter, /mnt/user/downloads will show all content of /mnt/disk1/downloads and /mnt/cache/downloads, so that shouldn't be the issue. However, why not set the downloads to cache only?

@JonathanM I'd do this by setting the Downloads share to "Cache - Prefer" correct?

If I do this, wouldn't my media content live exclusively on the cache drive? Radarr/Sonarr HardLink to the Downloads folder. So from my understanding if I set Downloads to Prefer Cache, my media would live exclusively on the cache and not be backed up to the array. Radarr/Sonarr would store the HardLinks to the array, but being as that is just a link to the Downloads my media still wouldn't be getting backed up. (Please teach me if I'm wrong).

[Version: 6.9.2 - Lost all docker settings/config]

@JonathanM mover ran again last night and I lost all my Queued downloads for Deluge. I have complete, and incomplete folders setup and they're under the downloads directory. Could the issue be a result of my incomplete directory being moved to the array?

 

 

 

[Version: 6.9.2 - Lost all docker settings/config]

47 minutes ago, JonathanM said:

Mover won't overwrite existing files, so if starting the containers created new files with the same names as before on the cache pool, mover won't fix it.

 

The quickest correct way to fix it is using mc at the console command line, and move the files from the array disk to the cache pool. DO NOT go into the /mnt/user folder for this specific operation, it will result in data loss. You want to move the appdata and system folder from /mnt/disk1 to /mnt/cache, overwriting all. The docker and VM services must be DISABLED when you do this, not just stopped, you must not have a tab for docker or vm in the GUI. You also need to set appdata and system shares back to cache prefer.

 

After the files are moved back, you can enable the docker and VM services in the settings, and things should be back to normal.

 

 

 

Urgh @JonathanM thanks so much! I had to look up what Midnight Commander was and how to use it. In case someone down the line makes the same mistake as me use this for reference: https://linuxcommand.org/lc3_adv_mc.php

 

Things are back to normal! Thanks Mods for the knowledge share :).

 

TLDR - Read the fine print on "Cache Prefer vs Cache Yes" for your shares. It affects mover settings as well.

[Version: 6.9.2 - Lost all docker settings/config]

8 minutes ago, ImEhCanadian said:

Are you referring to how I have the shares set to "Cache - Yes" instead of Cache - Prefer? Or where is this set?

 

I was trying to set it up so Appdata lives on the Cache drive, and backups are stored on the array.

 

I thought the only difference between Cache - Yes and Cache - Prefer was that Cache prefer stores the files on the cache until it's full, then moves to the Array. My cache drive didn't even reach 75% so I thought having this set to Yes it'd perform the exact same as prefer.

 

 

Oh Jeez 😢. I can't read.... just read now that "when mover invokes" prefer moves them from array to cache... I'm an idiot. Thank you for the help @trurl.

 

So to be clear, the problem that I faced is because my AppData got moved from cache to the array? I just set the shares AppData, System, and Domains to cache Prefer so that should fix the issue in the future right?

Can I restore my old appdata? Now that I set them back to Prefer, can I just invoke my mover again and it'll be fixed??

[Version: 6.9.2 - Lost all docker settings/config]

10 minutes ago, trurl said:

Why do you have appdata, domains, system shares set to be moved to the array? If these are on the array, dockers/VMs won't perform as well due to slower parity array, and they will keep array disks spunup since they always have open files.

Are you referring to how I have the shares set to "Cache - Yes" instead of Cache - Prefer? Or where is this set?

 

8 minutes ago, trurl said:

Also noticed in syslog that you have CA Backup configured to backup /mnt/cache/appdata. But your appdata setting has moved some of appdata to the array so those won't be included in the backup.

I was trying to set it up so Appdata lives on the Cache drive, and backups are stored on the array.

 

5 minutes ago, trurl said:

The page you linked says:

which is not what you did. You set them to cache=yes

I thought the only difference between Cache - Yes and Cache - Prefer was that Cache prefer stores the files on the cache until it's full, then moves to the Array. My cache drive didn't even reach 75% so I thought having this set to Yes it'd perform the exact same as prefer.

 

[Version: 6.9.2 - Lost all docker settings/config]

Hello Unraid Community!

 

I've just recently started with unraid and have been trying to learn as much as I can. I set up unraid a week ago and all was working fine until I stopped all my dockers, manually ran the mover, then manually ran the parity check.

 

The next morning my docker containers do not contain any of the saved settings. Plex no longer points to my movie/tv shares, deluge doesn't boot, and sonarr/radarr do not have any of the content in them. (Prior it was all setup and working fine).

 

I believe the problem is with my shares and how they utilize the cache drive. When I ran the mover maybe it just shoved the config off? But I'm not really sure what I did wrong. Any help/insights is appreciated.

 

 

 

I tried following this guide as best I could for the setup: http://www.exzacklyright.com/2020/02/how-to-setup-and-install-unraid-plex.html

 

 

Insert image from URL

tortuga-diagnostics-20211015-0720.zip