during a crash with beta9, I had a disk redball.  Instead of rebuilding parity from scratch, I want to force the array up with valid parity and let the parity check correct wrong blocks.  What would be the procedure to bring the array up as valid?

Figured it out...

 

Just do a new config, put the disks back into the same slot, and then click the "valid" checkbox for parity before starting the array.  It will startup with parity valid and run a parity check and correct any inconsistencies...

New config forces the system to accept the information on the data drives and write the result to the parity drive.

 

I want to force the array up with valid parity and let the parity check correct wrong blocks.

 

Which drive do you think is correct? Parity or the red balled data drive? A red ball means a write failed, so any information written to the data drive after it red balled didn't actually get to the physical drive, it was written to the emulated drive, which is created by the parity disk and all the other data drives.

 

 

My server went nuts and caused IO errors on most of the disks as well as the parity drive.  All the IO at the time was an rsync which was still in the process of copying the hidden .file so i am pretty confident in the data.  I think there are a few parity bits that are not correct based off of previous crashes.  I would rather it correct a thousand or so parity in consistencies than have to rebuild all 4TB of parity.  Doesn't really save any time, but the data is more protected incase of another drive failure.  I would rather have a few corrupt files than loose a whole drive if a drive happened to fail before the rebuild..

 

Ok. Sounds like you have it under control. Your original post was a little ambiguous on what was going on, and what you expected to accomplish.

Sorry, I should have elaborated more on my first post.  I get a little freaked out when I get a red ball on a disk...

 

Thanks for your help...  It is always good to get a second opinion when destroying your parity...