Hey all,
I'm having a bit of an issue with storage vMotion in our cluster
We had 3 ESXi boxes (2 ESXi 5 and 1 ESXi4) running independantly. We have started to see an increase in usage, and i finally convineced the bosses we needed some form of HA in it, so I've setup vSphere to move everything over - but i have hit a bit of an issue and was wondering if you could help
I have setup 3 new ESXi 5 boxes and connected them all to our Openfiler SAN that presents an iSCSI mount to all 3 of the new ESXi nodes. Thats all fine and it works. All versions of ESXi5 are all the same as they were installed from the same disk
All the VM's on ESXi 4 have migrated over with no downtime, so everything is working fine. Now we are left with the 2 "old" ESXi5 nodes
I am trying to migrate the disks to the iSCSI mount so i can move them over to the new nodes (they are using local storage at the moment)
I have moved the redhat 5 VM's over and all the windows ones. What i am left with is 3 Debian 6 VM's and 1 CentOS 6 VM - all 64 bit
Now, when i try to move them (Right click > Migrate > change datastore) and select the OpenFiler SAN it starts, gets to 76% then fails with:
A general system error occurred: The migration has exceeded the maximum switchover time of 100
second(s). ESX has preemptively failed the migration to allow the virtual machine to continue running
on the source. To avoid this failure, either increase the maximum allowable switchover time or wait
until the virtual machine is performing a less intensive workload.
I googled around, and found that i can increase the timeout time and it should fix it, but i dont think it will as some people have said they have and it hasn't fixed it. I also need to avoid downtime if i can, we have a 1 week notice period.
I brought up a temporary NFS server to test as well, and same thing when i try to migrate it to the NFS mount.
Network layout (Old nodes)
NIC1, Switch 1 - Untagged 1002 for network
NIC2, Switch 2 - tagged 301 (SAN), tagged 2198 (private network) <- vMotion enabled on this nic
Network layour (New nodes)
NIC1,2 - Switch 1 - Tagged 1002 and 2198 (vMotion here)
NIC3,4 - Switch 2 - Untagged 301 (LACP trunk through)
Network labels line up on all
This is ONLY effecting the Debian and CentOS 6 VM's - Windows and CentOS 5 work fine - so i believe my network setup is fine as is all the storage as everything works fine, just storage vMotion for these 3 friggen VM's
It gets weirder - if i right click on one of the VM's that wont migrate, and clone it to a new VM and then migrate it - it works fine with the network cables disconnected.
I know what you are thinking - reboot the thing and then migrate it - want to avoid downtime, but that will be my next option. Its actually annoying me and playing on my mind why it wont work, so would like to figure it out
All VM's run the latest version of VMWare tools and it is running
If anyone can offer any ideas that would fantastic
Thanks guys
