I'm working on a NIP which needs to publishes on my TREP infrastructure. Here's my need:
- One NIP
- Two ADH
- I can't use multicast on the server which hosts the NIP
- I need a fault tolerant solution on the TREP side
Here's what I'm doing currently:
And here’s my test scenario:
- I start the ADH pair:
hot standby : ADH1 active / ADH2 standby
cache resiliency : True
cache percistency : True - I start the NIP
- When the NIP starts, it connects to the first server of its list: ADH1
- My NIP publishes a couple of updates on InstrumentA
- I shutdown the NIP
- I kill ADH1
- If I dump the cache of ADH2, there’s no trace of InstrumentA <= this is a problem
Another scenario:
- I start the ADH pair (same configuration)
- I start the NIP
- When the NIP starts, it connects to ADH1
- My NIP publishes a couple of updates on InstrumentA
- I kill the ADH1. (ADH2 is now active)
- The NIP connects to ADH2
- I restart the ADH1. (ADH2 is still active, so ADH1 is standby)
- I kill the NIP
- I start the NIP
The NIP connects to the first server of its list: ADH1 (which is standby) <= this is another problem…
Am I doing something wrong?
Is there something I should do on my NIP in order to post simultaneously (via RSSL, because I can't use multicat) on both ADH?