Latest ADM caused continuous RAID activity

[Norbyte]

Hi,

I've updated yesterday to ADM 3.4.5.RBJ3 and since that I have continuous RAID activity.
But there not running:
- RAID sync
- Disk Doctor
- SMART check
- RAID scrub
- AV (ClamAV) scan.

I have Camera recording to a non RAID volume (Single disk) but turning off Surveillance Center and cameras, the disk activity remains.
There are suspicious activities too.
When I login to web interface or just check the NAS via mobile app, both shows hign CPU percentage (above 70%), but it suddenly drops to 5-15.

Processes

activity.JPG (66.48 KiB) Viewed 2001 times

Also just after the login (Web/mobile) checking running processes suddenly some of them goes to sleep, like "watch_storage", "kworker", "md1_raid5"...

Tried just to restart the NAS, thinking it is affected by a stucked process, but the RAID activity immediately starts.
Interesting that, starteing (enabling) RAID scrubbing stops the noisy disk processes, but when thatfinishes, it returns.

My NAS is an Asustor AS6104T 3pcs 3TB WD RED in Raid5 and a single 2TB WD Purple. Memory is upgraded to 2x8GB.

Norbert

[SapoMalo]

Same here... on a 5002, a 602 and a 6102...

apparently solved (or in the worst case, much improved) after deleting the app
"IronWolf Health Management" (but still testing)

...that was also useless to me as I use WD Red.

Cheers

Luca

[orion]

I did not encounter this problem. However someone here found a solution: reboot NAS. https://forum.asustor.com/viewtopic.php?f=228&t=10696

[SapoMalo]

Well, in my case, rebooting was the first thing I tried... and then I let the NAS do "whatever it was doing"...

It did not help unfortunately...

Deleting the app apparently did, but maybe just coincidence... I can't tell...

[Norbyte]

I've run from terminal the "killall -9 thumbnail" which not really helped, but then I gave a try for "killall -9 watch_storage" several times.

This finally ended the HDD self destroy which ran - of course for temporary - the last 3 days continuously.
Very nice ADM update, grats for the developer. Without any status indicator...

Do you have idea how to recover the "lost" space eaten up by the thumb generation?
Under the 3 days from my ~5Tb movie, picture collection it made approx 150Gb something to somewhere (/volume1/.@thumbnail/finish and /volume1/.@thumbnail/jobqueue) with the size of 0 to 60MB.
The folder size command doesn't give back real size since it shows 4TB when running "/volume1 # du -hs .@thumbnail/"...
Volume2 isn't affected.

[Norbyte]

Today the convert, watch_storage and thumbnail processes came back.
I suppose they are scheduled to run, even when from ssh they are killed.
Had to kill them again.

Unfortunately the rollback to the previous ADM not possible. Doesn'tt Asustor plan to publish the prev version with newer build version to be possible the downgrade to the not shitty version?

[orion]

You can try to install a package that someone provided to stop this process:
https://forum.asustor.com/viewtopic.php?f=240&t=10717

[lmgaup]

I was on 3.4.5 and had that grinding. Te package orion lined to did fix it for me. Now I'm on 3.4.6.RCO3 and the grinding is back. I'l take any hints on what to try

edit: I can't at a quick glance find what is causing the grind. The previous culprits aren't there. There also are no high cpu usage than I have been used to having, as I have four docker instances running, neither of which gave this nasty grind on the disks before. Right now I kind of wish I could use iotop to see what is happening.

[orion]

I was on 3.4.5 and had that grinding. Te package orion lined to did fix it for me. Now I'm on 3.4.6.RCO3 and the grinding is back. I'l take any hints on what to try

edit: I can't at a quick glance find what is causing the grind. The previous culprits aren't there. There also are no high cpu usage than I have been used to having, as I have four docker instances running, neither of which gave this nasty grind on the disks before. Right now I kind of wish I could use iotop to see what is happening.

ADM 3.4.6 provides a setting to adjust the conversion task under ADM web -> Settings -> General -> Off peak hours. You can try to set it up.

[lmgaup]

ADM 3.4.6 provides a setting to adjust the conversion task under ADM web -> Settings -> General -> Off peak hours. You can try to set it up.

I hadn't noticed that. I will try changing there and also let it continue whatever it is doing. After a few hours I do think it sounds a bit less chuggy-sounding and closer to normal operation