Hello,

 

Something seems to have happened during the process of upgrading parity drive(s) and I'm not what caused the issue or how to proceed. I have also included my diagnostics file for review. My existing array had one 8tb WD red as single parity, a second 8tb WD red and one 4 tb WD red as storage for a total of 12 tb of storage with just one parity drive backing them up.

 

I bought four 20 tb Seagate Exos x22 drives (refurbed) from serverpartdeals to upgrade to dual parity and increase the storage capacity of the array. I added the four drives to my server, then used the Unassigned Devices Preclear plugin to perform pre-read, pre-clear, post-read on all four drives, this process completed successfully.

 

I then stopped the array, assigned one of the 20 tb drives as parity and another one of the 20 tb drives as parity 2. I selected maintenance mode, and then started the array. The parity build process began and ran just fine over night. 

 

I had to work a 24 hour shift starting the next morning, and expected to come home and find my parity rebuild process complete. But, nope. The two parity drives showed red X's next to them, and in the columns next to them in the Array List, the first of the two parity drives showed something like 1250 errors and the the second one something like 600 errors. Down below in the Array Operation area, the parity status showed something like 12.1 tb, 62%, 0 errors, however things seemed to be stopped. There was both a Resume and Cancel button directly above, I clicked resume, it greyed out but then came back active, I tried clicking it again but nothing happened. I also noticed the two drives I had selected to use as parity and parity 2 now also showed up at the bottom in an area named Historical Devices. It seems that somehow Unraid is identifying them differently than before? The long device names of the drives are the same, but the "sd_letter" names have changed.

 

Also of note, I recently migrated my Unraid server to new hardware. Everything seemed to work fine after the transition, and I successfully completed a parity check on my old array setup while on the new hardware before attempting to upgrade my parity drive(s). The successful preclear of the four Exos drives was also done on the new hardware. However, if you see something in my diagnostics that you think my be related to something other than drives themselves, please let me know.

 

Have the new Exos drives I've received failed? Or is it just a communication issue due to loose/failing drive cables or case backplane? Again I successfully checked parity of the old array on the new equipment and precleared the four new disks on the new equipment, so it seems like if it was a bad/loose cable/backplane something would have failed then? However, I have no real knowledge of how to read or translate the Unraid diagnostics so any assistance in isolating and resolving this issue would be greatly appreciated! 

 

Thanks in advance!

 

 

epicenter-diagnostics-20240109-1252.zip

Unfortunately you rebooted before getting diagnostics so can't see anything in syslog before that. But it looks like you are still having connection problems with multiple disks.

2 hours ago, trurl said:

Unfortunately you rebooted before getting diagnostics so can't see anything in syslog before that. But it looks like you are still having connection problems with multiple disks.

How would you advise to track down the cause of these connection issues? Thank you!

Check connections, all disks, both ends, SATA and power, including splitters. Reseat controller card. Ideally no more than 4 disks per power cable.

 

All connections should sit squarely on the connector with no tension in cables that might disturb the connection. Don't bundle data cables.

Upon taking a closer look, I got to thinking that maybe my hard drives were getting power hungry during this process, and possibly trying to use more power than was available to them, due to the way I had them connected, and therefore causing them to drop out intermittently.

 

I have a Rosewill L4412U server case, with 12 hot swap bays that are arranged in 3 cages of 4 bays. Each cage has 2 molex power connectors on the back and 4 sata data connectors. I got this case second hand from a friend, and he had the 2 molex connectors on each cage connected to a Y-adapter that reduced the two connections to a single sata power connector. I went ahead and just used those and connected each cage to an individual sata power cable from the power supply.

 

After having the issues above, I decided to go ahead and remove those adapters, and connected a dedicated molex power cable to each of the cages. This seems to have resolved the issue, as I have now successfully upgraded my parity drives without any errors.

 

Thanks for your help trurl!