Is there any documentation, or whitepaper on the considerations in running active/active datamovers in a two-datamover Celerra? We have a unique (to me) use case where dual-switch network infrastructure runs active/passive. i.e., only one of the switches is active at a time. I am thinking that *maybe* we can use active/active datamovers to provide higher levels of availability in case of primary switch failure (and subsequent switch failover).
Has anyone else had a similar use case, and what did you do about it?
Thank you, Rainer. Curious about possibly using VDM's for this purpose, so that we can split the workload across two active DM's. Seems easier to failover a VDM onto a logical network construction than doing it at the physical level. But I could be wrong.
A little embarassed to ask this follow-up: Cross-stack LACP? Can you add a little color to that?
As a follow-up, we're using 10GbE connections, and each DM only has two 10GbE ports on it. If we run in a standard active/passive config, then we have only one active port to the network. That's why I'm asking the question... Trying to provide higher levels of availability on the network in case of the active (and only) link failure.
Rainer_EMC
6 Operator
•
8645 Posts
128
0
Posted March 4th, 2011 09:00
No – data movers are alway primary / standby - Or primary / primary with no failover
The options I quickly see:
- Do a FSN to both switches or cross-stack LCAP
- Write a scripts that monitors the switches and initates a data mover failover in case of the primary switch isn’t working
Rainer