Hey all,

 

I've started having issues where the unraid UI, access to shares in windows and dockers are all freezing intermittently at the same time.

 

Initially I looked around and it seems like it could be a ram issue, which makes sense, and I'm running a full memtest now, but it's not currently coming up with any errors (if it does I'll close this off).

 

I've restarted the host, Scrubbed the docker container, disabled the mover temp, checked the shares and cache configurations. I just can't seem to find the culprit. The server is under load when this occurs, but it doesn't make sense that it would cause this much freezing. I've regularly put it under load before and never had issues like this.

 

Just wondering if there's any thoughts on what might be causing it. Diagnostics attached.

paulflixhost-diagnostics-20230614-1438.zip

Enable the syslog server and post that after a crash to see if it catches something.

Cool, will do - For clarity these freezes only last around 5-10 seconds, where nothing is useable, nothing loads. After that it pops back to life before it intermittently freezes again.

 

Will update shortly.

 

 

1 minute ago, mrpauljames said:

After that it pops back to life before it intermittently freezes again.

Ahh, in the case it won't help, instead try boot in safe mode and leave all docker containers disabled, then start enabling one by one and re-test.

OK So booted in safe mode, restarted dockers, same issue - as soon as anything intensive starts, like an extraction or moving of a file from cache to the array it seems to freeze up the entire system.

 

I've got SABNZB and Unmanic setup and if I run either of those, if they're running processes it causes everything to stutter and freeze for up to 10 - 15 seconds. including with unraid UI. I've got the downloads and transcoding going to the cache drive and the paths configured to use cache instead of user to speed things up. Maybe I need to resinstall the OS, just not sure at this point. If I leave these dockers disabled, everything seems to run ok, but I'm not 100% confident.

 

Is there anything in logs that I might be able to look at to see what's going on?

 

As I've said, I've run memtest and that came back OK. Is the CPU knackered?

 

Driving me a bit up the wall lol

If it works for you use case you could try using a disk share (or updating to v6.12 was use an exclusive share) to bypass shfs, this has been known to help some users, but basically those docker containers would need to be limited to a single array disk or pool.

That's the thing, they don't touch the array - only the cache pool. It has been running fine previously, it's only just started having these issues over the last few days, which is what seems so strange. 

18 minutes ago, mrpauljames said:

That's the thing, they don't touch the array - only the cache pool.

But are you using /mnt/user or /mnt/cache for the container paths?

/mnt/cache

Bump

18 hours ago, mrpauljames said:

/mnt/cache

In that case sorry but no more ideas.

Any else this can be escalated to or am I just stuck at the moment. If not I'll rebuild, recreate all my dockers and settings from scratch (rather avoid if possible)

 

Over the years there have been a few users that see this behavior, suspect it could be hardware or some hardware combination related, but was never able to pin down, some users had success using different cache devices or filesystem.

OK. to give it a punt I've ordered all new sata cables and an additional SSD to add as a new pool just for downloads and transcode, so it's not sitting with the other docker app folders, I'll try a fresh install and see what happens. Fingers cross, thanks for all your help so far.

 

OK, so after working on it over the weekend the issues are now fixed. I didn't exactly troubleshoot this in a systematic way so I can know for sure what the issue is what I did:

 

- Replaced all SATA Cables, just to ensure that they weren't the issue.

- Added an additional SSD as a cache dedicated for heavy usage (downloads / transcoding).

- Fresh install of unraid, keeping the array data, but rebuilding all the docker containers effectively from scratch.

 

After all of this, the system now seems rock solid - not more freezing (touch wood).

 

Cheers to @JorgeB for your help.