On 6/14/2011 2:47 PM, Adam Levin wrote:
>
> S dedup is not yet mature.  But aside from that...
> I have heard some anecdotal evidence which I have yet to believe about
> data corruption, since I have seen no proof for it.  However, Doug even
> pointed out some occasional gotchas.  So, it's going places, and I think
> it will probably be huge someday, but right now if I have
> business-critical data, I would not choose to put it on ZFS if I could
> avoid it.
>
The glitches that we experience are around the occassional kernel hang 
that requires a machine reset. We have one machine that gets a hard-hang 
about once a month. it is also the most heavily used. We, so far, have 
not been able to get a good kernel dump. We also have a few IDRs in 
place (pre-patch patch) to fix some signficant bugs that happen when a 
zpool gets too full.

That said, zfs is solid, solid, solid for integrity of the data, but 
it's nowhere near as solid for high availability environments as you'd 
get out of a NetApp, BlueArc, or the like.

I will continue to use it in production on ~30 servers.

_______________________________________________
Tech mailing list
[email protected]
https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech
This list provided by the League of Professional System Administrators
 http://lopsa.org/

Reply via email to