Ltm upgrade 10 to 11

Once you upgrade one unit, both units will be unaware of the other.

When i have done this in the past i make sure to disable the link between my production interface and my core switch to ensure there is no chance of going 'active/active'.

We would need more details in order to give you an informed answer.

If you are simply upgrading an LTM that has Virtual Servers on different LTM Devices as pool members then it should not interfere with its operation because it treats pool members like a simple Node (IP and Port Combination).

HA Pairs, GTM, and custom interactions are where you would need to be careful when doing upgrades, but the more information you provide the better we can say for certain.

Hope this helps.

To clarify: There is 2 DCs in each DC there are 2 x LTMs running as active/standby. I can only upgrade 1 dc each month. My query is that in each DC there is an exchange cas virtual server that contains 2 pools one from its own site and one from the other site - so will the ltms still all talk together when one pair is on v10 and the other on v11

So, you have two distinct pairs of LTMs, one pair per site.

I don't see any reason why you couldn't upgrade each pair separately. Each pair is entirely distinct from the other - they do not 'talk together' or rely upon each other in any way it seems?

I disagree there must be some connection between them as how would the virtual server in DC1 know that the pool in DC2 is still available - I believe ltms talk via ssh but not sure on this point

The HA pair in DC1 has a virtual server with a pool (and thereby nodes) from DC2. The question is therefore will the virtual server still have the pool as up as there must be a communication between the ltms and will this be affected when one pair is on v10 and the other on v11

Mikey, why are you so sure there is communication between the HA pairs? A node or Pool Member can be anywhere on the network as long as its reachable. The DC2 pair doesn't have to be involved.

It could be if the Pool Member is actually a VS on the DC2 pair but this doesn't seem to be the case. Perhaps you could check?

If you were on v11 you could perhaps have a DSC Device Group with four members but that can't be the case here.

Mikey,

I agree with the others. If you have two Datacenters, each with distinct HA Pairs then they will not communicate with one another unless they are in a Sync Group (which they are not because you are running v10.x.x and Sync Groups were not introduced (for LTMs until v11, but have always existed for GTMs).

Your question "My query is that in each DC there is an exchange cas virtual server that contains 2 pools one from its own site and one from the other site": EACH LTM Pair should have Health Monitors checking the status of its pool members and their availability, so the LTM pairs appear to be independent of one another from then information you have given and you should not have any issues when you are upgrading.

Hope this helps.