[SOLVED] SAS drive issues - Too many wrong and/or missing disks!

[abynez]

Running 6.8.3

Supermicro X9DRH-7F motherboard

LSI 9207-8i HBA 

 

Longtime user who never had any problems with SATA disks except a bad cable so I got cocky. I fell to temptation and picked up some secondhand SAS drives for cheap and threw them in the server. I precleared (3) disks with dual cycles before using them and they all passed. Last week I started getting the email warnings about read errors. Then I got a notification about a disabled disk. I backed up everything vital and started poking around. Now I have one disk missing (disk 9), one disk disabled (disk 8 ) and another disk that is greenball but SMART shows predictive failure (disk 7) It appears to me that I just got a bunch of dying HDDs but they're all connected to a new LSI 9207-8i and new breakout cable that I've never used before and so the cause of this is suspicious to me. Unfortunately I don't have another cable to test with right now BUT I did buy (2) replacement SATA drives that I'm more than happy to replace the SAS drives with. Microcenter only had two left in stock ¯\_(ツ)_/¯ I figure if I can safely recover my server then I'll play with the SAS drives separately.

 

Attached here are my diagnostics from 12/22 which is from before I reboot. It was a few days after the initial problems so I worry the logs may not capture the original causes. Also attached are diagnostics from 12/23 which is from after I reboot.

 

Disks 1-6 are original and not an issue. 

Disk 7 has 1.28 TB of data written which I'm not worried about losing if necessary

Disk 8 has 169 GB of data written which I'm not worried about losing if necessary

Disk 9 had nothing written to it except (I believe) a 21 GB docker img file. 

 

I've spent the past two hours reading similar posts in the forum and got some leads but I worry about proceeding without checking with the more knowledgeable community. Help me obi-wan kenobis. 

viento-diagnostics-20201222-1555.zip viento-diagnostics-20201223-1057.zip

[JorgeB]

There's no SMART for one them, the other two don't look very good, and it's not cable related.

 

You can always do a new config with the new SATA drives, re-sync parity, then try to copy the data from the SAS drives using UD.

[abynez]

Yea no SMART for disk 9, it looks completely dead. I figured it would be bad drives but thought you all could interpret the logs and reports better than I could. Thanks for taking a look. 

 

Sorry to be a bother but could you give me more detailed instruction for the new config and resyncing parity steps? That sounds like a scary move . Once that's done I can try UD to transfer data like you suggested.

[JorgeB]

-Tools -> New Config -> Retain current configuration: All -> Apply

-Unassign all the SAS disks

-Assign the new SATA disks

-Start array to begin parity sync (new disks will need to be formatted)

-Once the sync finishes try to copy the data from the old disks.

[abynez]

Thank you!

 

I had to wait for the disks to preclear before starting but it worked great. New config allowed me to restart the server with disks 1-6 data in tact and usable. After parity-sync completed I tried to mount the disks in UD. Disk 7 mounted and I was able to recover all 1.28 TB of data off of it. Disk 8 wouldn't mount and Disk 9 is completely dead. I ended up giving up on the 169 GB of data from Disk 8 but it's no biggie. Nothing critical. Thank you again and Merry Christmas, Happy belated Hannukah, Happy Kwanza, Happy New Year, Happy almost end to 2020, etc 😃