Register for the iXsystems Community to get an ad-free experience and exclusive discounts in our eBay Store.

Issues with drives falling out of sync after crash (VIRTUALIZED FREENAS)


Feb 21, 2019
First of all, all hardware specs are in my signature.

Long story short, I had a server crash, and after a reboot, two drives were out of sync. I possibly could have synced them again (not sure if this is possible), but I decided to just resilver them one at a time.

While resilvering the first one, it slowed down to a halt (original resilver time was 2-3 days, but after slow down, it kept increasing to 40-50 days and icreasing), indicating an issue. I tried to do a smart check on the drive I was resilvering, and it wasnt responding. I decided to reboot freenas, which worked fine, but it seems like it started an automatick full scrub on restart. This seems to be trucking on without issues, but I am unable to access freenas in the browser, and I am unable to type anything in the command line interface (direct access, see screenshot).

I assume I should just let it finish. It seems to be progressing nicely, and the stability issues were most likely due to insufficient cooling. We were going through a period of extreme heat when the server crashed, and I have no Air Conditioning.

The question. Does it seem normal (SEE SCREENSHOT), and can I estimate a completeion time based on the reported speed, and size of pool?

Pool: 10x10TB WD Red White Label drives, single vdev RAIDz2. Total raw space 90TiB, total usable space, 67TiB.

According to the screenshot, the reported speed seems to be around 2 800 000 000 to 3 700 000 000 (bits/bytes?). I am uncertain of the unit that is reported, but I believe it is bytes.

Assuming avg_seg_size is reported in bytes, and assuming an average scrub speed of 3 000 000 000 bytes per 5 seconds:
This gives me 600 000 000 Bytes/second, or 572MB/s scrub speed. Does that seem realistic for my setup?

Also, assuming 572MB/s, and a total raw space of about 90TB, should I expect the scrubbing process to take about 165 000 seconds, or 45 hours.

Are my assumptions correct?