Hi Shishir,
Great to see this being progressed!
Would it make sense to compare the checksums for a file across replica's too? ie. in a replicated volume, one the local checksum compare is complete, compare the checksum against the file's replica on another brick node. This would account for 'funnies' in the data path during the write phase introducing corruption.
Cheers,
Paul C
From: "shishir gowda" <address@hidden>
To: address@hidden
Sent: Tuesday, 24 December, 2013 8:35:08 PM
Subject: [Gluster-devel] bit rot support for glusterfs
Hi All,
I will be starting to work on bit rot detection for glusterfs.
This work will be primarily based on Doug Willam's prototype
(https://github.com/williamsdoug/glusterfs_bitrot_proto).
The existing design can be found @
http://gluster.org/community/documentation/index.php/Arch/BitRot_Detection#Current_Prototype
Few design updates:
1. Depend on change-log to recompute checksum. This eliminates
periodic crawl of brick/volume to update the checksum.
2. Policy to determine when checksum to recomputed. If a file is under
going active I/O, then compute checksum only after a delay
3. Ability to turn off/on bit rot detection in volumes.
4. If bit rot is turned on for a volume, a crawl would be necessary in
this case to compute checksum.
This just a initial draft, and will update the design doc, and the thread soon.
Please feel free to comment.
With regards,
Shishir
_______________________________________________
Gluster-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/gluster-devel