Hi, I have an unraid server running and since novembre last year, the drives connected to my HBA go missing. They usually show up after boot and then after a random amount of time they go missing.
it could take hours or days. After I reboot everything shows up as expected.

I have 6 drives and 3 of them connected to the HBA. They survive the parity check of 18TB with no problem.

At first I thought it's the drive, then I thought it was overheating, so I replaced the card and the cable, repasted it and installed a 40mm fan on it's heatsink and a 120mm fan directly into it. I live in Canada which basement temperature of 18c

this is the card in question : DELL PERC H200 HBA FW:P20 LSI 9211-8i IT Mode

https://www.ebay.ca/itm/124518680935?var=425395225711
Unraid version: 7.0.1

AMD Ryzen 9 3950X

32GB ram

Gigabyte B450 AORUS PRO WIFI-CF , Version F66d
RTX 3060

PSU 1200W
 

Attach Diagnostics ZIP to your NEXT post in this thread.

Also setup syslog server

Last issue happened yesterday, I setup an automatic reboot at 4am.  I attached the diagnostics and syslog 

syslog.log jaximus-diagnostics-20250403-1812.zip

I noticed in your syslog it says mpt3sas version 43.XXX.XXX.00 loaded with your 2008-based controller. With a 2308 controller in my server, t says mpt3sas version 43.100.00.00 loaded, with the same unraid version 7.01. That could be a sign something isn't entirely right. I'd run a memtest.

8 hours ago, hamidflacko said:

Last issue happened yesterday

Do you know the aprox time? DOn't see any drives going missing in the log.

8 hours ago, Wody said:

I noticed in your syslog it says mpt3sas version 43.XXX.XXX.00 loaded with your 2008-based controller. With a 2308 controller in my server, t says mpt3sas version 43.100.00.00 loaded, with the same unraid version 7.01. That could be a sign something isn't entirely right. I'd run a memtest.

I will run it again and see if something happens, also is there a way I can bump the version of the controller? 

6 hours ago, JorgeB said:

Do you know the aprox time? DOn't see any drives going missing in the log.

I'm not sure because it happens without any notification, I unfortunately, I'll be monitoring it more closely and upload a new log/diag with better timestamps if I can 

On 4/4/2025 at 3:39 PM, hamidflacko said:

I will run it again and see if something happens, also is there a way I can bump the version of the controller?

I got a 9211-8i today, and it also said version 43.100.00.00. What that means is somehow there is some corruption happening, because it displays XXX instead of a number. Figuring out where that comes from would and fixing that would probably be the solution. It could be your USB has issues, or your memory, or the CPU.

 

As for the version of the controller, it is already running 20.00.07, which is the last version released for it.

It happened again tonight at 2:20 am exactly, I don't see anything in the logs either. What can I do about the USB to fix the issue? Also could it be the motherboard? Today I didn't want to recognized any of the drives from the HBA until I changed it's position on the board. FYI when booting I could see that the HBA was initialized correctly and all of the drives were showing up, but once in unraid they were gone  

jaximus-diagnostics-20250405-1255.zip syslog-10.10.0.100 (5).log

What do you mean exactly by "the drives stop showing"?

In the Main section where I can see the drives of my array, they're marked red MISSING 

OK, I see the error now, but there isn't anything in the persistent syslog:

 

Apr  5 08:38:02 Jaximus kernel: mpt2sas_cm0: SAS host is non-operational !!!!

 

HBA issues, make sure it's well seated and sufficiently cooled, you can also try a different PCIe slot.

I tried moving it the main PCIe slot and it seems to be working so far(Uptime 1 day 6 hours ), I also found this  thread I'be adding it too and see how it does! 

append initrd=/bzroot pcie_aspm=off

 

Edited April 7, 20251 yr by hamidflacko

I replaced my GPU from RTX 3060 to an ARC A750 this morning and after 2 hours the same thing happened again. Issue still seem to be present even with the tweaks. I restarted the server and everything is back to normal, I pushed the HBA in place to check if there's a connection issue, we'll see.