Unraid loads old config on reboot

[lmanstl]

I am setting up a new unraid server and am still working through all the kinks. 3 of my SSDs refused to format no matter what HBA I attached it to so I just removed those.
 

I replaced them and then added them to the cache pool in place of the 3 broken ones. It worked correctly and I was able to format the pool. But the next time I rebooted, it says that the 3 drives are wrong and that it wants the 3 broken drives again. When I start the array it says that those 3 drives are disconnected and contents emulated. It also won’t let me reformat the pool.

 

I tried making the pool bigger and that didn’t work. I made the pool smaller and that worked but when I made it the right size again it pulled up the old config. I was able to format if I changed the file system to encrypted but a reboot still brings back the old config and then I converted back to regular and it let me format again but after a reboot it still has the old config. Deleting and remaking the pool also didn’t help. I can’t get unraid to forget the old broken drives.

[JorgeB]

Unassign all cache devices, start array, stop array, re-assign all cache devices, start array, that should do it.

[lmanstl]

20 minutes ago, JorgeB said:

Unassign all cache devices, start array, stop array, re-assign all cache devices, start array, that should do it.

You are amazing. I can’t believe it was that easy, the only thing I haven’t tried. I have spent like 3 days trying to fix this among other issues.

 

Thank you.

 

I don’t suppose you know how to fix an SSD that shows up perfectly fine everywhere but refuses to be accessed through an OS and refuses to format through the HBA BIOS. It is an HGST HUSMR1650ASS210.

Edited July 25, 2022 by lmanstl
Add another question

[JorgeB]

11 hours ago, lmanstl said:

I don’t suppose you know how to fix an SSD that shows up perfectly fine everywhere but refuses to be accessed through an OS and refuses to format through the HBA BIOS. It is an HGST HUSMR1650ASS210.

 

Please post the diagnostics after a format attempt.

[lmanstl]

52 minutes ago, JorgeB said:

 

Please post the diagnostics after a format attempt.

I can get the diagnostics in about 18 hours. The error message when trying to use fdisk to format in unraid terminal or Ubuntu is /dev/sdf: fsync device failed: Input/output error

 

When I try to format using the internal Cisco 12g SAS controller or the LSI 9300-8e SAS HBA I just get a message that says Failed. The cards also don’t seem to realize that the disk is an SSD like the others that are present instead listing it like a hard drive without self encrypting features.

 

My best guess is some sort of firmware issue but I don’t know what caused it or how to solve it. It happened on 3 drives at the same time but not the other 5 that were attached or the 17 that weren’t attached.

[lmanstl]

Here are the diagnostics.

 

This line in the syslog marks the start of the format operation on the bad SSD: Jul 26 15:36:23 Artemis emhttpd: shcmd (152): mkdir -p /mnt/test

 

I am not really expecting a fix, but if you recognize something that could cause this let me know so I can avoid breaking the rest of my drives.

 

Thanks again for helping.

artemis-diagnostics-20220726-1743.zip

[JorgeB]

Looks like a device problem, it's not giving a complete SMART report:

 

>> Terminate command early due to bad response to IEC mode page