After upgrading from 9.0.0. to 9.0.1c the 40G stacking ports stayed down on two-member ICX7450 stack and after upgrading. These were 40G modules in the rear of the unit, in slots 3 and 4. We had unit 1 module 3 wired to unit 2 module 4.
Despite the configuration not changing, the 40G port would not come up. Swapped stacking cables, both modules, and rebooted countless times to take incremental steps before rebuilding the stack from scratch -- still no go. So we built the stack with 10G ports ... and then 15 minutes later the 40G ports came up, causing a traffic loop!
Any ideas why the 40G ports would not link up after the upgrade and many, many reboots?
Ok -- didn't know if I was missing something like have to force the 40G ports to be copper or 40G-full or something else.
I am working on the support contract side to get a case open -- I'm told my our consultants that it should have been covered under the project purchase , but when we contacted TAC they said it was not.
Please try checking "Show stack " try checking if the units in stack are coming up correctly , example unit 1 should display as unit 1 in " show stack " output and unit 2 as unit 2 in " show stack " output . If you see any changes with respect to unit being displayed , try defaulting that member and have the stacking cables connected . The switch will automatically join the active member
Thanks for the suggestion. The stack wasn't re-forming after the software upgrade, or during any subsequent repair attempt, because the 40G ports weren't linking up.
Initially we defaulted just the member (unit 2), but that didn't make a difference, the 40G ports weren't linking up.
And then also defaulting the primary/master didn't make the 40G ports come up.
Only thing that allowed us to restore service to customers was building the stack on a 10G link.