Interpreting Badblocks Output

[cyberjock]

That's just SO freakin' weird. That drive must be really f'ed up!

 

[NannerMuffin]

Yeah....crazy. It could have been another poorly seated cable, unlikely, but I wasn't willing to risk it. Glad I decided to RMA. I'll rest easy knowing all the disks are healthy.

 

[paylesspizzaman]

Just throwing this out there, you dont have a pool on those disks yet correct?
I had the same issue you did and found out that it was the .system writing to the disks behind my back. I destroyed the pool and reran badblocks and the compare errors totally went away.

Will having a pool on drives really screw up badblocks? I read in the guide "make sure you don't have a pool on the drives" but I thought that was because it would destroy your saved data. 3 of my 16 drives have between 400-500 errors during the second badblocks pass. I had a test pool setup on 3 drives that I did not destroy prior to running badblocks. I thought about the pool maybe causing problems after I saw the errors, but my pool was on da0-2 and the errors are on da8-10. However, I did change some hardware, so I bet the pool is no longer on da0-2. I know what I'm doing tonight when I get home...

 

[nick779]

Will having a pool on drives really screw up badblocks? I read in the guide "make sure you don't have a pool on the drives" but I thought that was because it would destroy your saved data. 3 of my 16 drives have between 400-500 errors during the second badblocks pass. I had a test pool setup on 3 drives that I did not destroy prior to running badblocks. I thought about the pool maybe causing problems after I saw the errors, but my pool was on da0-2 and the errors are on da8-10. However, I did change some hardware, so I bet the pool is no longer on da0-2. I know what I'm doing tonight when I get home...

If you have an active pool, and especially if your system dataset is located on the drive, badblocks wont be accurate because the system will write over a block that badblocks already wrote to.

Ive heard that you can do a non destructive badblocks test, but imo its not even worth doing at that point.

 

[paylesspizzaman]

If you have an active pool, and especially if your system dataset is located on the drive, badblocks wont be accurate because the system will write over a block that badblocks already wrote to.

Ive heard that you can do a non destructive badblocks test, but imo its not even worth doing at that point.

Thank you for the reply. I looked last night and all three drives with errors were indeed part of the test pool I had created. I had not saved anything on that pool and the .sys file goes to a usb drive for now. I find it very unlikely that I have 3 bad drives, that I just happened to randomly choose to put in a 3 drive pool. So obviously, freenas still saves something to the pool. I destroyed the pool last night and I'm going to restart the badblocks test tonight. Sucks that I wasted 5 days on the test though.

 

[nick779]

Thank you for the reply. I looked last night and all three drives with errors were indeed part of the test pool I had created. I had not saved anything on that pool and the .sys file goes to a usb drive for now. I find it very unlikely that I have 3 bad drives, that I just happened to randomly choose to put in a 3 drive pool. So obviously, freenas still saves something to the pool. I destroyed the pool last night and I'm going to restart the badblocks test tonight. Sucks that I wasted 5 days on the test though.

Jesus 5 days? a badblocks test should only take around a day or two for even the biggest of drives.

Use Tmux and run the tests simultaneously.

 

[paylesspizzaman]

I did tmux, but I'm running them on sixteen 4TB drives at once. I only have a 4core xeon D cpu, if that makes a difference. When I started the first test, my estimate was 3 days, but once all the tests were running, it seemed to slow down. Do you think something is wrong?

 

[nick779]

I did tmux, but I'm running them on sixteen 4TB drives at once. I only have a 4core xeon D cpu, if that makes a difference. When I started the first test, my estimate was 3 days, but once all the tests were running, it seemed to slow down. Do you think something is wrong?

Nevermind, I thought I read you were using 3 drives :P