disk 5 keeps going Missing (Disabled) from my Array. Cannot identify if this is an Issue with my HBA h220 controller card, SATA cable or Bad disk?  - smart reports show no errors on the disk. rebuilt array 3 times and the disk keeps eventually going offline within a few days time.

 

In my attached logs, you will see that the SAS controller suddenly goes "non operational" followed by a write error.

 

Not sure how to properly diagnose. Appreciate any guidance received. Thank you much.

 

unraid2-diagnostics-20190503-1000.zip

Edited May 3, 20197 yr by Maz

Looks like a problem with the HBA, make sure it's well seated and sufficiently cooled, you can also try a different PCIe slot if available.

Thanks for the feedback. I have 2 PCIe slots, so I moved the H220 (HBA-9207) card to the secondary PCIe slot and re-seated cables. Its re-building the parity again now... we shall see if this makes any difference.  The Primary PCIe slot is generally intended for graphic cards so not sure if that can cause any issues having the HBA card installed there.... My secondary PCIe slot is X4 vs X16 so less bandwidth.

 

On a second note, my H220 SAS controller card heat-sink it very hot to the touch. I know these things are Known to run hot, but I am worried the heat might also be a factor. Is it common practice to install fans on these controller cards or to consider re-applying heat-sink thermal paste?  I have good airflow in the case so its not that...

You should have some airflow directly over the heatsink, these HBAs are made for servers where low airflow is not problem.

2 hours ago, johnnie.black said:

You should have some airflow directly over the heatsink, these HBAs are made for servers where low airflow is not problem.

Thanks for the feedback - I went ahead and ordered a fan to permanently mount onto the HBA controller heat-sink.  (Noctua NF-A4X10)

 

I first tested by temporarily hovering a fan over the heat-sink of the HBA, and it made a big difference temperature wise.  My Newly Licensed Unraid build is in a tower configuration, so it likely does lack the same cooling as a server case where these cards are generally intended for.

 

Between moving the SAS card to a different PCIe slot and better cooling, I shall see it this eliminates the dropping of disks from my array.

 

Once system is stable, My next concern is loading the HBA card up with 8 drives (8i) ... Wondering if the LSI card produces more heat  with 8 drives connected vs. only 1 ... ... thx again.

Edited May 3, 20197 yr by Maz

25 minutes ago, Maz said:

Wondering if the LSI card produces more heat  with 8 drives connected vs. only 1 ... ... thx again.

I remember doing dome tests and 1 or 8 disks is about the same, since I use tower cases I use a large fan on the side on the case and that's enough to keep the HBAs / SAS expanders cool enough, a dedicated fan is also a good option, in some cases you might get away with just good case airflow, but having a fan blowing in or around the HBA is always better.