Different disks with disk error and getting disabled - Help needed

[jcarre]

Hello, I posted a couple of days ago regarding my server acting up after nearly a year working flawlessly. I fixed that error with a new installation of postgres, but I am experiencing major issues with disk failure.

I decided to fully restart the server, first time after the last update, and one parity drive gave write errors and was disabled (this was yesterday). I though nothing of it, and tried to fix the problem with my memory getting full.

A couple of minutes ago I attempted to add the disabled parity drive back into the array, starting a new parity operation. Shortly after, disk7 shows write error and is also disabled.

The parity operation was automatically stopped, and I am worried about losing data.
I know that having two disks suddenly dying is not normal and that something else is probably going on.
The last parity calculation was done on the 11th of January without issue. I set it to do it every 3 months.

What is the best course of action?
All my disks are connected with an HBA (LSI 9201-16i). I am worrying about it being the failure point.
The server has been working fine for nearly a year, but it lacks an exhaust fan. Maybe the card is failing due to old thermal paste and bad airflow?

The server is on a 4u case mounted on a rack with other computers nearby. The case has 3 Noctua industrial fans blowing air from the front, controlled with a plugin with the disk temperatures. This makes the server very quiet when not in use, but may get the card very hot most of the time?. I really don't know.

 

Please see attached my syslog + diagnostics.
Thank you very much.

 

ivpiter-diagnostics-20240309-1959.zip ivpiter-syslog-20240309-1858.zip

[JorgeB]

It's not logged as a disk problem and SMART for both disks looks OK, disk7 has a large number of UDMA CRC errors, but they don't look recent, still power/connection issue would be the main suspect.