Parity check extremely slow (< 1MB/s)

[tim86]

Hello guys,

 

in the last couple of months I have become increasingly frustrated with Unraid and the parity algorithm. While it does SOMETIMES work to create or validate parity, most of the time (especially if I haven't done it in a while) the process is just completely broken. I have tried to debug what causes the errors that sometimes appear in the syslog, but I don't know what to do anymore. Obviously, checking 18TB of parity drive would take half a year with this speed, which is kinda stupid. Listening to the system I can also hear drives stopping and restarting (SATA issues?).

 

I have now replaced all SATA cables twice and the problem is still occuring. Any ideas how to debug?

tower-syslog-20220927-1433.zip

[trurl]

attach diagnostics to your NEXT post in this thread

[JorgeB]

Please post the complete diagnostics.

[tim86]

Sorry, I was stupid. I was clicking through everything so much, I accidentally downloaded the wrong one. 

It's attached now.

tower-diagnostics-20220927-1647.zip

[trurl]

Your appdata, domains, system shares have files on the array. You want these shares all on fast pool (cache) so Docker/VM performance won't be impacted by slower array, and so array disks can spin down since these files are always open. Probably you have Dockers/VMs accessing array disks during parity check because of this.

 

Why do you have 50G docker.img?

[trurl]

You're having connection problems on disk3 so this is going to be the main reason for your parity check speed.

[tim86]

Just now, trurl said:

Your appdata, domains, system shares have files on the array. You want these shares all on fast pool (cache) so Docker/VM performance won't be impacted by slower array, and so array disks can spin down since these files are always open. Probably you have dockers accessing array disks during parity check because of this.

I have recently replaced the cache disk, which is why they are currently all on the array. I also kinda don't trust the cache disk in terms of inconsistency. I did have some problems with that about 2 years ago. But because of all these stability problems I have moved basically all of my productive Docker containers off of Unraid for now anyway, so the impact shouldn't be too bad.

 

2 minutes ago, trurl said:

 

Why do you have 50G docker.img?

I remember running out of space with the default settings a couple of years ago, so I increased it. Given that my cache disk is large enough, I didn't see why it is a problem.