After having to take down both nodes to upgrade to 5.4, I was hoping that the re-sync procedure for a mirrored pair of servers would be much faster. Unfortunately this still doesn't seem to be the case. In our situation we are looking at a re-sync rate of about 1 TB/day. On the bright side, the speed seems reasonably consistant regardless if we re-sync a single empty LUN with no data (other than a VMFS file system) or a 50% full LUN. Still overall, I won't see the two nodes synced for at least 48 hours.
Speeds between servers are amazing, I can easily saturate 1 Gbps for over an hour when running a file transfer between the two servers. So why can't the sync saturate the connection (not to mention there are currently 4 active NIC's on the server)? Anyone else see this issue as well?
The Latest Gartner® Magic Quadrant™Hyperconverged Infrastructure Software
