Dale Ghent wrote:
> Sorry for the double reply, but by the way, what sort of compression
ratio are you seeing on your ZFS filesystems?
{cyrus1:vf5:136} zfs get compressratio cyrus/mail
NAME PROPERTY VALUE SOURCE
cyrus/mail compressratio 1.26x -
I'd be interested to see what happens when gzip compression is available.
I'm not sure whether that will be 10u4 or have to wait for 11.
The T2000 is a 4-way 8-core box so we are not short of CPU, loads running
between 1-3 so it's mostly idle now. I haven't seen the 10u3 CPU-locked
behaviour ZFS-compression to be a constraint yet don't see that much kernel
CPU-time that would indicate it. Most activity is the Cyrus processes.
I should have specified correctly earlier, we have ~70K accounts of which
we've only migrated 27,400 at last count, so we have a ways to go. I can't
see any reason the hardware couldn't handle the whole load on a single
setup.
We are doing a SunCluster failover cluster pair of T2000, but this
effectively
means a single T2000 is handling the job at any given time. Based on
current
loads I can't see any reason for a MURDER as long as you have at least
mid-range
hardware to work with.
We currently have 8 gigs of RAM, I should appeal for more..........
----
Cyrus Home Page: http://cyrusimap.web.cmu.edu/
Cyrus Wiki/FAQ: http://cyrusimap.web.cmu.edu/twiki
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html
