Another Slow Mover question

[Effoc]

Hello folks i have today an array of 17 disks and a pool of ZFS i download to.
But latley have the ZFS pool started to get full for some reason so i started a mover session by myself and i noticed it onley writes to my array in 2.5 -> 8,2 MB sometimes 12Mb at maximum for some sec.

And i have no ideas left af read every guide on Write cash, Turbo write and reconstruct write and i cant figure out whats wrong now anymore.
Please can someone smarter than me see if you can find anything

Best regards a confused Swede

ph2-diagnostics-20231109-1523.zip

[JorgeB]

Which share are you moving to? Note that most free allocation method can be much slower, also if they are small files the mover itself will be slow.

[Effoc]

Im moving From ZFS/media/film to array/media/film
And most of the files are around 50gb in size.
Changed setting to high water now.

[JorgeB]

Any better?

[Effoc]

Do i need to restart unraid to know the difference or will it affect direct?
its still the same and i have restarded mover but still 2.0 ->8MB/s sometimes at 14MBs

Attached new file after the changes

 

 

ph2-diagnostics-20231109-1753.zip

Edited November 9, 2023 by Effoc

[JorgeB]

40 minutes ago, Effoc said:

Do i need to restart unraid to know the difference or will it affect direct?

Not but turbo write is disabled now, enable it.

[Effoc]

Tryed restart the server and now is the tranfer speed 3.9MBs to 8.1MBs
Mover is moving to disk 5 (sdad) ib the array

 

One thing that hits me now is the HDD has writen 4Kn on it if that has to do something whit it

 

Made some math now if i tranfer at 3MBs it gonna take me 30days to tranfer 1 Terabyte and i have 6 Tb
 

 

Edited November 9, 2023 by Effoc
added picture and info

[JorgeB]

It doesn't look like a mover problem to me, transfer something directly to the array to see if it's the same.

[Effoc]

Dit that yesterday. Transfered from my pc direct to disc 5 and same speeds

[JorgeB]

So it's a general speed issue, unrelated to the mover, start by running the disk and controller tests using the diskspeed docker.

[Effoc]

But this test only test read speed

 

[JorgeB]

Yes, but test all disks and also do the controller bandwidth tests.

[Effoc]

18 minutes ago, JorgeB said:

Yes, but test all disks and also do the controller bandwidth tests.

Doing one now just wanted to provide some info as fast as possible.

 

Edited November 10, 2023 by Effoc

[JorgeB]

Everything looks normal, only thing I see that raises a concern are the controllers, the LSI in IT mode is fine, though the firmware should be updated, but there's also an LSI MegaRAID controller, there could be something with the write cache or config that is slowing down the writes, and please note that RAID controllers are not recommended in general.

[Effoc]

Starnge it have worked fine always before.
But i can try to erase everything and separate everything so all disks on the array is on the same raid card

[JorgeB]

If you can start over start with a small array, even just one disk on a different controller, no parity initially, and retest.

[Effoc]

Will do its 1 year of harvesting files but wtf need to find out whats wrong

[Effoc]

when i restar the server i get this message
 

[Effoc]

Removed all the HDDs from the other raid card and now its full steam ahead
 

[Effoc]

Im so confused now and this is like watching paint dry. Now im back at 2mb -> 12mb again
And it have been like this for 2 houers now

[JorgeB]

Doesn't seem very likely but possibly one of the disks has slow sectors, temporarily disable turbo write and check if it speeds up.