pre-processing rules on 2 hub environment

[Aduno SA]

Hello, in my nimbus domain I have two hub, A and B. All the auto-operator rules and pre-processing rules are set on the nas on hub A. Hub B is configured with ha probe for failover and auto-operator disabled. If I have a robot connected on hub B, the preprocessing rule on hub A is not working when an alarm is raised. Is this correct? I.E.: I have set a preprocessing rules on nas A to lower the critical to major. I have also set an autoperator rule on nas A to send an email for major alarm. If an alarm is raised from a robot on hub B, i get the email but the preprocessing rules is not working so alarm is not lowered.  I made a wrong configuration or what? thank you

[Anon Anon]

Hi,

 

If I remember correcly, nas replication forwards alarm events, which are "alarm_*" messages, no the "alarm" (or "alarm2") messages, which are handled by preprocessor. Therefore preprocessing would only take place in the first nas to handle the message.

 

-jon

[Aduno SA]

Thanks John,

do you think is there any workaround possible? The only way maybe is enable auto operator on secondary nas so also preprocessing rules becomes available but is this "correct" on a ha environment?

[Garin Walsh]

John, I think that you have it backwards. Nas replicates the results of reading the alarm2 messages.

 

Preprocessing operates on the alarm messages before anything else gets to them.

 

And preprocessing isn't part of the "disable auto operator" checkbox on nas.

 

So if you have preprocessing, you need to have it set up on all the nas probes and expect it to operate on the alarms regardless of the HA state.

 

The flow would be something like:

 

Alarm happens on hub B

alarm enrichment on B reads alarm message and applies preprocessor script to it

alarm2 is published back onto the bus and written to the nas database

replication picks up the new entry to the nas database and forwards to hub A

Hub A stores entry to its database and evaluates AO profiles

relevant AO profiles fire.

 

At least this is what I observe happening in my environment.

 

-Garin

[Anon Anon]

surely preprocessing is applied by nas and not alarm_enrichment? At least in terms of "nas preprocessing", which I was referring to. Otherwise I believe what you said is what I was trying to say as well.

 

-jon

[Legacy User]

Normally in a three tier architecture with nas HA setup, you would have the core primary hub with nas set with AO enabled, core secondary hub with nas set to AO disabled and HA probe.  The second level/proxy hubs would only connect to core primary hub and nas under normal operations but on HA failover would connect to the core secondary hub and nas. 

What you would have configured on the core secondary hub are identical queues for your second level/proxy hubs (like what is configured on core primary) that are not active.  The HA probe on core secondary would be configured to start needed probes, enable queues,and enable the AO.  You would need to configure your nas forwarding and replication on primary nas to replicate scripts and configuration so that the AO pre-processing and profile scripts are copied to the HA nas.  However, the configuration is not merged so you need to periodically manually merge the nas.cfg from your primary with the nas.cfg on your HA. 

Doing this allows you to run the same pre-processing and profiles on your HA nas during failover.

[Aduno SA]

Ok Sane, but this is not a case of failover. The question is, an alarm raised from  a robot under a secondary hub is not lowered by a preprocessing rules on nas A.  Is this correct?

[Halim Liauw]

Hi All,

 

We have same requirement.

 

What is the conclusion or solution to this? Can we turn on AO on secondary hub with HA Probe so that the pre-processing rules will work for those robot reported on secondary hub?