I am running unRaid 5.0. I am in the final process of replacing a 2 TB drive with a 4 TB drive. I mention this only for context as the process went fine.

 

As the final step, I started a parity check just to make sure that all had gone well. That parity check is about 60% complete, but it started throwing parity errors on a drive unrelated to the replacement drive.

 

I went to look at at the syslog (attached) and discovered what looks to me like memory errors at the start of the syslog. Lines 63,65,67,69,70,72 and 74 all show in red in the syslog, and have the word '(Errors)' at the end of the line.

 

(Curiously when printing out the syslog via unmenu it doesn't show the '(Errors)' at the end of the line.)

 

What I'd like to know is if these indicate there is a problem with my ram memory, and if so, is the first step in correcting the problem to run a mem test after the parity check finishes?

 

I'd also be curious to know if it is possible that the memory problems are the source of the parity errors that I got during this check, starting at line 1347. They are all on disk 8 which appears ok to me (smart report attached).

 

 

Thank you in advance,

 

hwilker

 

syslog-2015-12-16.zip

smart_report.txt

SMART shows 2 pending sectors.

SMART shows 2 pending sectors.

 

Thanks. Missed that. Any insight on the memory errors in the log?

No memory errors, those are false positives.  That color coding and syslog line categorizing was added a long time ago, and has not been kept up to date, so is a little unreliable at times.

 

The only problems I saw in the syslog are related to issues with Disk 8 (WDC_WD20EARS-00MVWB0_WD-WMAZA4383175).  Unfortunately, you are running an old version with early driver support for that LSI card, so the errors are too cryptic to interpret.  Once you've upgraded to v6, you should get better error descriptions.  However as Trurl said, the SMART report shows 2 bad sectors, probably the cause of the issues.  You might try a SMART long test to verify.  And unfortunately, you can't trust your parity drive now for rebuilding data drives, until this is cleared up.

"No memory errors, those are false positives.  That color coding and syslog line categorizing was added a long time ago, and has not been kept up to date, so is a little unreliable at times."

 

Thanks for the info. They certainly looked like a problem.

 

"Once you've upgraded to v6, you should get better error descriptions.  However as Trurl said, the SMART report shows 2 bad sectors, probably the cause of the issues.  You might try a SMART long test to verify.  And unfortunately, you can't trust your parity drive now for rebuilding data drives, until this is cleared up."

 

Does this make sense as a plan.

1. Let the parity check complete (I'm at 86%) just to see if there are further issues. (FYI, it's a non-correcting check)

2. I have a hot spare. Use it to replace and rebuild disk 8.

3. Run another parity check after replacing the disk.

4. (Long overdue) upgrade to V6.

 

Thanks for the help.

 

 

 

 

 

"No memory errors, those are false positives.  That color coding and syslog line categorizing was added a long time ago, and has not been kept up to date, so is a little unreliable at times."

 

Thanks for the info. They certainly looked like a problem.

 

"Once you've upgraded to v6, you should get better error descriptions.  However as Trurl said, the SMART report shows 2 bad sectors, probably the cause of the issues.  You might try a SMART long test to verify.  And unfortunately, you can't trust your parity drive now for rebuilding data drives, until this is cleared up."

 

Does this make sense as a plan.

1. Let the parity check complete (I'm at 86%) just to see if there are further issues. (FYI, it's a non-correcting check)

2. I have a hot spare. Use it to replace and rebuild disk 8.

3. Run another parity check after replacing the disk.

4. (Long overdue) upgrade to V6.

 

Thanks for the help.

Since it's a non-correcting check, that does sound like a good plan - *IF* you are confident in the parity drive prior to this.  The fact that you were running a parity check would seem to imply that you weren't completely confident in it?.?

"The fact that you were running a parity check would seem to imply that you weren't completely confident in it?"

 

No. I was confident in it. It's just that I had just finished replacing another drive. I needed another drive on my PC, and so replaced a healthy 2TB from my array with a 4TB to buy some extra capacity. That went fine and didn't indicate any errors at all.

 

But out of an abundance of caution, whenever I've replaced a drive, I've always followed it up with a non-correcting parity check after replacing it. Just seemed like good form. I guess it paid off this time.

 

FYI I probably should have mentioned that the parity errors I got were exactly 256. I don't know the sector size but that looks suspiciously like the size of something, either a sector or perhaps a byte at the same offset in 256 sectors. I don't have the expertise to know but such a distinct numbe seems to imply to me that those sectors really are the problem. The disk is also rather old. It's been out of warranty for over 3 years which makes it 6-8 years old. Not terribly surprising that it might be bad. Swapping it out is fine with me if it's likely to solve the problem. I'd only do a Long Smart report if I were desperate to save the disk. I'm not, as I'm slowly getting rid of all my 2 TB drives and repurposing them on PC's if they're healthy.

 

Thanks for you help.

Just a follow up. I did the procedure outlined. I replaced the problematic disk and rebuilt with a new disk. Then did a non-correcting parity check, receiving no errors. So all appears well. Now on to installing V6, which everyone says is simple but looks intimidating to me 

 

Thanks for the assistance