ACE statefull connections sync

[cajalat]

In an FT setup how can I tell if a flow on the ACTIVE ACE is sync'd to the HOT ACE? Is there a way to see the various connections that are replicated to the HOT ACE?

I've looked at the "sh conn" command but it doesn't seem to yield what I'm looking for.

[agiaccone]

What's the output of the 'sh ft group detail' in the context ?

On the standby ACE you should see the My State set to FSM_FT_STATE_STANDBY_HOT. In this case the sh conn on the standby ACE will let you see all of the stateful connections.

Hope this helps,

alberto

[cajalat]

I do show FSM_FT_STATE_STANDBY_HOT on the standby and FSM_FT_STATE_ACTIVE on the active (which is correct). The problem is that I'm not seeing the same connections which I would expect to see. I'm running A1(6.3b). I first noticed this when I upgraded from A1(6.3) to A1(6.3b). As a test I have a controlled connection. Starting with a replicated set of connections on both active and standby I issued my "sh conn" command to see that both active/hot displayed the same connections which they did. After I reloaded the standby and waited for it to be in a HOT state. The connections disappeared. So I restarted the app and the connections appeared on the HOT ACE. So then I failed over the ACTIVE and the HOT went to active and the connections were there. When the partner came back up and was finally in HOT mode I checked the connections but they didn't show up. So I restarted the application and still I don't see the connections. It is as if the connections aren't replicating. Doing a random check on various connections I would expect to be active and alive I noticed more inconsistencies. So that got me to think maybe I wasn't using the right command.

Casey

[agiaccone]

as far as I know, you're using the right command. It's the same command I used on my lab to test ft before deploying the ACEs.

Unfortunately I have no more lab so I checked the commands on my production couple where situation is too dynamic to have a precise idea of replication of connections.

I observed, however, that on my 1(6.3) couple the sh conn count command issued on the same second on the two ACEs gave me the same value with a 5% error (which could be caused by normal fluctuation of flows on applications).

HTH,

alberto