Re: [tlug] Making better use of SSDs?

[Edward Middleton <edward.middleton@example.com>]

On 05/29/2012 06:02 AM, Attila Kinali wrote:
> On Tue, 29 May 2012 01:26:02 +0800
> Raymond Wan<rwan.kyoto@example.com>  wrote:
>
> > Hi Andrew,
> >
> >
> > On Monday, May 28, 2012 11:57 PM, Andrew Holway wrote:
> > > btrfs is really an alpha file system. Use with extreme caution. ie.
> > > not on production systems.
> >
> >
> > Ok!  Thanks for the warning!
>
> And you should listen to that warning!
> Or just hang out on #btrfs on freenode for a couple of days.
> There is every day someone who lost his file system because
> he hit some strange bug, or upgraded the kernel.
>
> BTRFS is cool and i'd like to use it. But it will probably take
> them another 5 years to get it prime time ready.

After the discussion of btrfs last time I tried a raid1 partition with 
btrfs.  It seemed that the major issues people were having occurred when 
they had power failures[1], kernel crashes or hardware failures.  If you 
are using it as a scratch space then this may be less of an issue. i.e. 
if your data can be easily regenerated and you are just using it for 
improved io.  Is anyone having issues with slow corruption of btrfs 
data?  I would have thought with block checksums[2] it would be more 
reliable then other filesystem in this regard.

Edward

1. thats what a ups is for.
2. http://en.wikipedia.org/wiki/Btrfs#Checksum_tree