2017-01-31 01:16 PM - edited 2017-01-31 03:19 PM
We have a pair of 3240's in an HA pair, and they're each using two single mode VIF's, one each for ISCSI and CIFS.
Each single mode VIF contains, in turn, two static multimode VIF's, each containing two physical interfaces. Each multimode VIF is connected to a separate Cisco 6500 chassis.
We are now upgrading those Cisco 6500 chassis with a couple of Nexus switches using vPC, which allows a single port channel to be configured across both switches. This allows the use of all 4 interfaces in a single multimode VIF. Yay!
What this means, however, is that I must destroy the existing single mode and static multimode CIFS and ISCSI VIF's to create a new dynamic multimode VIF for each, each containing four interfaces. I think I'm probably OK with the mechanics of the VIF destruction and creation process, either from the CLI or OnCommand, but I imagine there are other things I'll need to look at afterwards.
I'll assign the same IP address to each VIF on each filer and configure the switches, but what do I need to worry about on the filers to ensure that HA is still working properly, that my Hyper-V cluster nodes still see the filer as the same ISCSI portal, and that my Windows clients can still see the CIFS shares at the same sharename?
Can anyone give me a hand on a good plan to approach the reconfig? Thanks for the help!
Solved! SEE THE SOLUTION
2017-01-31 04:15 PM
If you choose the CLI approach, remember to also commit the changes to /etc/rc and /etc/hosts files. If you make the changes in a current version of system manager this should be done for you, but you still need to validate the on-disk config and perform a takeover/giveback in both directions to confirm the configuration works as intended in takeover mode and persists across reboots.
2017-01-31 04:23 PM - edited 2017-01-31 04:25 PM
Thanks for the reminder about etc\rc. I think I'll have to use the CLI for the CIFS VIF at the very least.
Still looking for any tips\hints for a configuration plan which will ensure ISCSI and CIFS services will be available to clients as before the reconfiguration however...
2017-01-31 04:33 PM
You're just reconfiguring phycial interfaces, vifs and maybe vlans. Keep the same IPs in the new config, and protocol access should work again after the cutover. Similar to a headswap, where you fix up the rc & hosts during the cutover, and verify TO/GB and client access before exiting the maintenance window. No need to do any protocol reconfiguration. Plan the new config files, backup the old ones, and have a test plan ready to execute against that includes verifying protocol access from a representative sampling of hosts.
2017-02-01 12:28 PM
"Keep the same IPs in the new config, and protocol access should work again after the cutover."
That's what I wanted to hear. So it sounds like just destroy the old VIF's, create the new ones, assign the same IP's, and things will just work?
Will test of course, just looking at what to expect. Thanks for your input!