Hi all, I'm trying to implement a Fault Tolerant scenario with one primary and secondary SS, trying the solutions suggested in this post:
Notifier in Fault Tolerance
I'm having trouble starting Alarm Notifier in the secondary one. If the primary is running, I get the following message when I start it:
I believe it's trying to connect first to the MLS, but the SS is rejecting the connection. According to the posts I have read, I assume it should start normally?
If I stop my primary SS and secondary becomes active (it's in hot state), Alarm Notifier tries to connect against primary first, acknowledges there is no answer (Timeout), and then reverts to secondary, and starts working. So the issue appears to be when primary SS actively rejects the connection.
Any ideas/help on this?
Thank you in advance!
I did a quick search to see if anyone else had reported this same issue. I found two similar cases. In one a restart of the SpectroSERVERs resolved the issue.
In another, a reload of the database on the secondary resolved the issue.
Just sanity checks just in case....hostrc file on each of the SSs involved allows the other to access? ( I guess if your online backup sync from pri to sec is working ok that couldn't be the problem). Does SS-Tools/MapUpdate -v report all the machines involved (3 entries for each one)?
Please check the below link , it may be useful for you:
Spectrum Alarm Notifier Fault Tolerance?
Also please refer the below Technical document :
Alarm Notifier SANM Fault Tolerant - CA Knowledge
Thank you Joe. I'll try both and come back to you!
Venkat, thanks for your response! I checked both of the links you suggested before attempting to do this. However, both talk about Alarm Notifier working just the same in a Secondary or a Primary SS, however, my Alarm Notifier Instances just won't start in the secondary SS, showing the error I attached in the Screen.
I believe every Alarm Notifier instance should try to get alarms from all landscapes in a DSS environment, regardless of FT, is this right? If that's so, I think the issue here is that the primary spectroserver is refusing to let the Alarm Notifier instance connect to it while it's running.
Could this be the case?