[SOLVED] Monthly check - 3 Errors detected, no unclean shutdown (logs included); new errors and SMART issue a month later

[JorgeB]

3 hours ago, Idaho121 said:

1) Could this be what's causing the parity errors?

Unlikely, and if yes it would be visible on the syslog.

[Idaho121]

Alright, as I should have done earlier, I ran a full MemTest and found errors. Have filed for an RMA, I'll test the new RAM before I load the box up.

 

So process from here - is this right?

1) Start server and array, save config file

2) Power down

3) Install new RAM

4) Run MemTest for a few days to be sure it's good

5) If it isn't, cry, then RMA again, repeating steps 2-5 until we're good

6) Restart server, start array

7) Run a non-correcting parity check.

8a) No errors, then this was just a RAM issue on the parity check

8b) If there are errors that are the same as my last check, run a correcting check

8c) If there are errors but they're different than my last check, run a new check to make sure I get the same errors

9) Once I get the same errors on consecutive checks, run a correcting check

 

Am I missing anything, or getting anything wrong?

 

Also, two other questions:

1) Should I be worried that I have corrupted files? I haven't noticed anything. I usually rip/record on one machine before moving it to my server.

2) Should I be worried about the drive with the read errors, or just have a backup ready to go because it'll fail eventually, but it's not a problem yet?

 

Thanks, as always!

[JorgeB]

It's possible the first check will find some errors, if yes just correct them, next checks should always return 0 errors.

 

It's always possible some data was corrupted while memory was bad, consider using ECC in the future.