We sure have been having problems with our server. Long story short we have been having IOPS issues and it resulted in a very very unclean shutdown of the server. I was unable to recover one dataset in ZFS. Unfortunately it was the largest one being the fedora buffet. I've been pulling the data back in as fast as I can with the current server. It is a fine line between pulling data in, serving data out, and having the server light on fire. I do know that ZFS with 17T of active data on a box with 64G of ram is simply not a good mix. I'm working on going back to LVM with XFS. I still don't have replacement hardware that I trust so my ability to move forward is limited. There is talk about having mirrors.rit.edu become an officially supported production service instead of the current model of "This is what I have, let's see what works". Of course the talks for this are moving at the speed of bureaucracy.

-paul -----Original Message----- From: David Caro [mailto:dcaroest@redhat.com] Sent: Monday, September 15, 2014 12:23 PM To: Cristóbal Palmer; Paul Mezzanini Cc: infra(a)ovirt.org; Brian Proffitt Subject: Mirrors unsynched Hi! We have detected that your mirrors seem not to be updated since (at much) August 21rst, are you having any problems synching them? Cheers! -- David Caro Red Hat S.L. Continuous Integration Engineer - EMEA ENG Virtualization R&D Tel.: +420 532 294 605 Email: dcaro(a)redhat.com Web: www.redhat.com RHT Global #: 82-62605