Migration from the KQ Implementation

Gordan Bobic gordan.bobic at gmail.com
Wed May 4 20:58:32 EDT 2011


On 05/05/2011 00:07, Jason J. W. Williams wrote:
>> The 3 guiding principles of data retention are paranoia, paranoia and
>> paranoia.
> Agreed. Which is why we've got snapshots, backups, and zdb. :)

Snapshots won't do you any good if the pool is damaged in an unfortunate 
way that stops it from reassembling.

Backups are of paramount importance, sure, but fsck is important, too, 
especially when a fsck pass might take 10 minutes vs. 2 days for 
restoring all the data from a backup (and the inevitable loss of 
anything since the last backup.

I actually use lsyncd to trigger file syncing to a backup volume on 
every write-close, and the backup volume is running CopyFS to ensure 
that every version of the file that was ever saved (well, within reason, 
space allowing) can be retrieved. But as I said, restoring from a backup 
is much slower than fsck would be.

Gordan



More information about the zfs-discuss mailing list