Long story but the current Dell Storage Switches have a fault, jumbo frames has failed due to ASIC issues. (confirmed by Dell).
We've new Dell Storage switches to replace the two in the datacentre, BUT the entire infrastructure runs through these storage switches.
We are trying to decide if we can do this without downtime although risky.
The new Dell Storage switches have been tested and jumbo frames is indeed working.
Currently Filer is running 8.2.3.7-mode (yes we know old! Flame retardant under-crackers on!), both heads are configured LACP into both switches on the stack
Head A
e0c - switch 1 e0d - switch 2 e0e - switch 1 e0f - switch 1
Head B e0c - switch 2 e0d - switch 1 e0e - switch 2 e0f - switch 1
Both Heads are in use, with volumes and NFS exports.
Failover has been configured properly e.g. we can lose switch 1 or 2 (not at the same time) and service is still available. (just writing this I've had a thought!).
We need to replace switch 1 and switch 2 above with switch 1a and switch 2a.....
I did wonder if we removed all cables from switch 1, and put them in switch 1a, would this work.... but the LAG would disappear from the Stack.... and when we remove all cables from Switch 2 to Switch 2a, would break....
Is the simplest option just to shutdown the DC and rip and remove.
Does failing over the heads help me here, e.g. take over Head A, so all is running on Head B.....
Or maybe it cannot be avoided but to shutdown all the VMware vSphere.
The more I think about this, I still think the safest is to arrange downtime.
Useful Comments ?
Thanks
Andrew