Parity Sync stuck

[dezai]

Hi Guy´s,

 

sorry but i have new problems with my server.

Last weekend my server locked completly down.

 

I couldn´t get into the webui/ssh into the server.

 

So i killed him and he started normaly after that and he started a parity sync.

And now he is stuck at this sny at 5%.

 

Attatched the diagnostics - can you please help me with that?

I´ve upgraded the storage pool last week (started with a completly new pool with new drives and copied everything from the backup server).

 

Everything is up and running - only the parity sync is stuck.

 

 

 

 

tower-diagnostics-20200416-2048.zip

[JorgeB]

Try pause/unpause, if it doesn't work reboot and start another check.

[dezai]

if i reboot the server it get stuck completly.

pause unpause is not possible.....

 

Is there anything else which is a problem in the diagnostics?

 

[JorgeB]

Then you'll need to force a reboot.

 

11 hours ago, dezai said:

Is there anything else which is a problem in the diagnostics?

Nothing I can see.

[dezai]

Ok one of my ssd´s is damaged

 

SDC can´t get formatted any more and spit out BTRFS Errors.

I remove this ssd ent test the server again.

 

RAM is after 6 hours memtest ok - no errors.

 

I order no one new Cache drive SSD for the server

 

A littlebit strange is, that i can use it in windows "normaly" - crystaldisk info is telling me that the drive is ok?

 

Edited April 18, 2020 by dezai

[dezai]

Ok Problems are stil lthere......attatched new diagnostics....tower-diagnostics-20200419-0831.zip

 

Parity Sync is stuck at 50% - drives are spindown - some docker containers went offline this morning.

I don´t know what i can do with this build - that is realy horrible 

[dezai]

can it be a cable problem?

 

After changing the ssd the other ssd spit out errors - no i changed to another sata cable and everything is fine after the reboot.

I give it some time again and keep you updated

[JorgeB]

There are still issues with the cache device, looks more like a connection issue, try replacing both cables, it can also be a compatibility issue with the controller.

 

That might not explain the parity check pausing, but don't any other issues logged, so try again after fixing the cache issue.

[dezai]

Sorry in my masseage 5 hours ago ther was no new diagnostics. Attatched the diagnostics from the past 6 hours.

 

It looks good at this moment.tower-diagnostics-20200419-1449.zip

[JorgeB]

Yes, everything looks fine so far.

[dezai]

so far.....the drive errors have gone but i get now new errors, i can´t get da diagnostics data and i can´t see the docker/vm´s in the dashboard.

But they are still running.

 

Errors in the syslog:

 

Apr 20 22:58:20 Tower kernel: CPU: 10 PID: 3278 Comm: kworker/10:2 Tainted: G O 4.19.107-Unraid #1 Apr 20 22:58:20 Tower kernel: Call Trace: Apr 20 23:12:14 Tower kernel: CPU: 5 PID: 26716 Comm: kworker/u64:5 Tainted: G W O 4.19.107-Unraid #1 Apr 20 23:12:14 Tower kernel: Call Trace: Apr 20 23:15:14 Tower kernel: CPU: 5 PID: 26716 Comm: kworker/u64:5 Tainted: G W O 4.19.107-Unraid #1 Apr 20 23:15:14 Tower kernel: Call Trace:

 

And the server locked up again today.

 

I´ve set this fix from spaceinvader a few weeks ago:

 

rcu_nocbs=0-15

 

For me as a linux noob it appears, that there is an issue with the cpu?

[JorgeB]

9 hours ago, dezai said:

For me as a linux noob it appears, that there is an issue with the cpu?

Could be the typical Ryzen issues, also see here.

[kAI53r]

On 4/16/2020 at 8:52 PM, dezai said:

Hi Guy´s,

 

sorry but i have new problems with my server.

Last weekend my server locked completly down.

 

I couldn´t get into the webui/ssh into the server.

 

So i killed him and he started normaly after that and he started a parity sync.

And now he is stuck at this sny at 5%.

 

Attatched the diagnostics - can you please help me with that?

I´ve upgraded the storage pool last week (started with a completly new pool with new drives and copied everything from the backup server).

 

Everything is up and running - only the parity sync is stuck.

 

 

 

 

tower-diagnostics-20200416-2048.zip 144.07 kB · 4 downloads

 

ONE POSSIBLE SOLUTION.

 

I had the same problem. My webGUI wasnt showing any progress. I was bulding a second parity + preclearing two new unassigned devices, which you know takes quite a while with high TB drives.

 

Apparently I use the "Dynamix System Statistics" Plugin and could see that there was a lot going on my storage read/write-wise eventhough the MainGui didnt show any activity.

 

I tested pausing the "parity sync" process and logging out of webGUI. What happend is that the parity sync continued and everthing worked normaly. I rebooted after all was done.

 

This might also work with the parity check.

[itimpi]

46 minutes ago, kAI53r said:

What happend is that the parity sync continued and everthing worked normaly

Do not understand this bit.  If you paused the parity sync it should not have completed until you logged in again and did a resume.  Are you sure it actually paused?

[kAI53r]

In my case the webGUI just froze, certain operations were not possible. The underlying process was still running and the information just got refreshed after this procedure.

 

So, just because it might look like its stuck, you should double check, because it might just be the webGUI thats causing issues.