Hi all, long time lurker, first time poster.
As SSD per-GB prices continue to drop at a healthy rate I've been thinking more
and more about the benefit of creating pools out of SSDs rather than HDDs. My
virtualization hosts seem to have at least a 2x dedup ratio, but I've been
cautious on dedup due to the reports here of much more random I/O.
I'm wondering what experiences or input some might have already had in running
SSD pools and if doing so might be a "magic pill" to make some zfs features
such as dedup and/or wider vdev stripes mostly pain free?
My assumption is that the SSD pool could handle the higher random I/O
requirement without really needing to worry so much about ARC or L2ARC
requirements for dedup. Would this be true?
If I wanted a pool that spans multiple chassis, should I be concerned about
elevated IOPS rate if I were using a handful of JBOD/SAS expanders each
connected with an external 6gbit link?
Anything else I might be overlooking?
Any input would be appreciated!
Thanks,
Tyler
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
