I am need help regarding below issue i am facing while configuring Fresh installation.
Number of Servers = 8
Number of Access Zone = 2
Number of Cluster = 3
1. Cluster Node = Central Node
2. Cluster Node = Reload Engine
3. Cluster Node = User Engine / Proxy
Note : We have install Database ( PostgreSQL) on stand alone server * both central node getting data from this database.
Note : All the installation are on AWS EC2 .
Step 1 :
We create / Install cluster on AZ1 ( Access Zone 1 ) and joined cluster AZ2 ( Access Zone 2 )
Step 2 :
We add AZ2 node in AZ1 QMC.( I dont know about this , should i do this or not, i did this just for testing purpose )
Step 3:
We install reload engine on AZ1. (Cluster) ( is it possible )
Step 4:
We join reload engine AZ1 as cluster in AZ2. ( is it possible )
Step 5 :
Installed 3 rim nodes and added in AZ1 QMC.
Step 6 :
Installed 3 rim nodes and added in AZ2 QMC.
Question # 1 : if AZ1 is down, do the AZ2 work in this case ?
Question # 2 : All available nodes need to be added on QMC ?
Question # 3 : How can we distinguish which server is central node and which is reload engine. * At the time of cluster installation it is not asking anything.
Can you please share your past experience , and please share best practice and recommended document.
Hi Osama - I think i am understanding. A couple of things:
- Two active central nodes against one database is not supported. There should only be one central node in a single cluster active against the database.
- You can failover to a second central node *within the same cluster* automatically if the first central node fails or manually via (Failover ‒ Qlik Sense) by marking a node as a failover candidate in the QMC.
- You can have components active against in two AZ's within the same cluster provided latency is <4 ms between AZs. Persistence ‒ Qlik Sense
- I am not sure what you are trying to accomplish here - full DR or more of an HA configuration. If you are looking for DR with a complete second cluster, then be sure that only one cluster is reading/writing to the fileshare and database at a time.
Question # 1 : if AZ1 is down, do the AZ2 work in this case ? See above.
Question # 2 : All available nodes need to be added on QMC ? You register each node in a cluster, but do not register nodes from one cluster into a QMC of another cluster.
Question # 3 : How can we distinguish which server is central node and which is reload engine. * At the time of cluster installation it is not asking anything. This is found under QMC > Nodes where you specify the workload of the nodes.
