Last night, I did some math with my kill a watt, and realized my Unraid server with a Sandy Bridge E5-1650 was adding around 12 dollars per month to my electric bill. I had a supermicro MBD-X10SL7-F-O mobe and an i3-4170 not being used, so I swapped motherboards to see what the i3 would draw on average.

 

 

The flash drive caught on my arm and got bent and destroyed. After figuring out which disk was parity, I reassigned all the drives, and started the array. All went well, and my data was still there. I also learned to keep the USB backup on a disk other than the array. I'm sure none of the drives are assigned to where they used to be other than the parity drive.

 

On the E5, I had 2 drives connected to the motherboard, and 8 on a LSI9211-8i HBA. I figured it'd be easier to just put the HBA into the i3 motherboard. Once all seemed to be running smoothly, I ran a non correcting parity check. I woke up to 39% complete with over 12000 sync errors.

 

I left for work, and started a correcting parity check. I'm currently at work, so I can't provide the diagnostics now. I did download them already however. I looked through the system log, and didn't see any disk errors. How likely is it that between the different motherboards, and different disk assignments, parity errors arose? The two drives not on the HBA are attached to 2 of the motherboard's integrated HBA ports. After the parity is done correcting, If I rerun it and there are no more errors, all should be well correct? Before this, I ran monthly parity checks and had no parity errors ever. 

 

I'm currently running 6.4.0-RC15 and was before I destroyed the flash drive. Thanks for help. 

34 minutes ago, david11129 said:

After figuring out which disk was parity,

How did you figured which drive was parity? If it was by starting all disks as data disks a few parity sync errors are expected on the 1st check.

16 minutes ago, johnnie.black said:

How did you figured which drive was parity? If it was by starting all disks as data disks a few parity sync errors are expected on the 1st check.

That's exactly what I did. I assigned them all as data disks, saw which one was unmountable, did a new config and set the unmountable one to parity.

 

Does 12 thousand or so count as a few though? It was around 40 percent complete when I stopped it, and started a correcting parity check.

Edited December 14, 20178 yr by david11129

Was the parity drive originally assigned as Parity 2 and now assigned as Parity (or vice versa)?

A little more than I usually see when doing this, usually under 1000, but it's not an absurd value, so if there are no disk related errors in the log I wounld'b be worried, as long a next check is 0 you're probably fine.

Just now, Squid said:

Was the parity drive originally assigned as Parity 2 and now assigned as Parity (or vice versa)?

That would cause millions and millions of errors.

I've only ever ran single Parity, and I did not assign parity to parity 2 at all. Thanks everyone for you're help. I'll update if the second parity check has problems.

The parity check finished with around 13000 errors. It's currently running again now. Would it be of any benefit to upload my diagnostics? Or just wait until this check is done?

Post them in the end if there are errors on the 2nd check.

Looks like we're good so far. I woke up to it being 80 percent done without errors. Thanks for the help! At least now someone will know it's normal to have parity errors due to this.

If manually mount individual array disks, it's an advantage to mount them read-only.