JWK - Jackall's IT Wiki

Sécurité, Système, Réseau

Outils pour utilisateurs

Outils du site


unsollicited ClusterXL Active member failover


When configuring a cluster of checkpoint Gateway, especially when applying configuration, unsollicited active member failover occurs. Change is lost, and cluster is often in collision state.

In Smart Tracker if you filter on 'clusterXL' string:

Member X ...detected a problem (fwd)
Member X...detected a problem (cphad)

This is a timeout problem on the two process (fwd and cphad).

How to solve the problem

On each member:

[Expert@HostName]# cphaprob -d cphad -p unregister
[Expert@HostName]# cphaprob -d fwd -p unregister 


[Expert@HostName]# cphaprob -d cphad -t 10 -s ok -p register
[Expert@HostName]# cphaprob -d fwd -t 10 -s ok -p register 

This will change the original timeout for the cphad and fwd process (2 sec by default) to 10 seconds.


unsollicited_clusterxl_failover.txt · Dernière modification: 2018/02/18 16:01 (modification externe)