Hello. Over the past year or so, I've consistently been getting a very small number of parity check errors during my monthly scheduled party check (it does not write corrections to the array). I used to not see any parity check errors, even when using this LSI card which I have had installed since 2023. My assumption was that the problem was due to my LSI 9211-8i card overheating, or from the cable being bad. To fix this I replaced the SAS cable, applied new thermal paste to the heatsink of the LSI card, and strapped a 40mm noctua fan to the heatsink itself. This resulted in a reduction, but not an elimination of parity check errors, which to me further signaled that the LSI card was the problem. I ran another parity check, this time correcting the errors, and thought my problem was solved. However, on January 2nd, the parity check showed 1 new error, I ran another check to verify and also am getting 1 error.

Timeline if that helps...
2025-11-02 - Parity check completes showing 9 errors

2025-11-25 - I repaste the LSI card heatsink, add a noctua fan, and replace the SAS to SATA cable.

2025-11-28 - Parity check completes showing 4 errors

2025-12-02 - Parity check completes showing 4 errors. At this point, I am convinced I fixed the problem and these 4 errors are a result of the LSI card writing bad data due to heat/bad cable.

2025-12-03 - Parity check with correction writes fixes 4 errors. I assume this is the end of my problems.

2026-01-02 - Parity check completes showing 1 error.

2026-01-07 - Parity check completes showing 1 error.

Any tips on how to diagnose this before going out and buying new parts? I feel like it could still be the LSI card, a cable, or one of my drives, but I don't have enough information to make that determination.

tower-diagnostics-20260116-1341.zip

RAM would be the #1 suspect, you can run memtest, but it's only definitive if it finds errors. Since you have two DIMMs, try just one, then the other one.

Note that the first check after the issue is resolved can still find errors, so you need to run two consecutive checks. If the second one doesn't find errors, it should be resolved.

9 hours ago, JorgeB said:

RAM would be the #1 suspect, you can run memtest, but it's only definitive if it finds errors. Since you have two DIMMs, try just one, then the other one.

Note that the first check after the issue is resolved can still find errors, so you need to run two consecutive checks. If the second one doesn't find errors, it should be resolved.

Thanks, I’ll try and give this a shot this weekend.

Found one error on the second pass, I’m guessing this is all the evidence I need but I haven’t stopped it yet. I may swap the dimm to the other slot to make sure it’s not the motherboard.

You must never attempt to run any computer unless memory is working perfectly. Everything goes through RAM. The OS and other executable code, your DATA. EVERYTHING. The CPU can't do anything with anything until it is loaded into RAM.