User: Password:
|
|
Subscribe / Log in / New account

RAID 5/6 code merged into Btrfs

RAID 5/6 code merged into Btrfs

Posted Feb 4, 2013 19:44 UTC (Mon) by dlang (subscriber, #313)
Parent article: RAID 5/6 code merged into Btrfs

> That's important if you don't want to spin every disk for each
metadata read.

when you have a raid setup, especially a raid5/6 setup, you very seldom have drives spinning down.

Also, if you don't read the entire stripe, how do you check the parity information? or do you not do that level of validation ond only look at the parity information if you get a read failure?


(Log in to post comments)

RAID 5/6 code merged into Btrfs

Posted Feb 4, 2013 20:06 UTC (Mon) by masoncl (subscriber, #47138) [Link]

I probably should have written seek every disk. Metadata reads tend to be more seek intensive, and if every drive in the raid set needs to be involved to read every metadata block, you end up seek bound over the whole array pretty quickly.

We're only checking parity if we have to rebuild a block. The rebuilds only happen if the crc check fails, or if you get an IO error.

-chris


Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds