n>3 nodes are in a pool at LTM(F5). No clustering at backend. Do we need to reserve multiple ports in the range 5701-5800 for cluster communication?
Hello Rajiv Nistala and @Aditya .... Can you please clarify my below queries on the same, we recently upgraded to 7.1.9:
1. We have n(n>3) nodes are in a pool at the LTM(F5) and there is no clustering on the backend even at the weblogic end for pega application so do we still need to reserve multiple ports in the range 5701-5800 for cluster communication?
2. As per pega recommendation - To enable communication between Elasticsearch host nodes in the cluster, open a TCP port in the range 9300-9399 on each node. - So what if the desired port in not available in Production, I mean what could be consequences?
Message was edited by: Lochan to branch it from https://mesh.pega.com/thread/55503
**Moderation Team has archived post**
This post has been archived for educational purposes. Contents and links will no longer be updated. If you have the same/similar question, please write a new post.
Accepted Solution
Pegasystems Inc.
US
Upon reviewing the associated SR, I see that SA-23077 has been created to resolve this. Please view "Service requestor could not be authenticated" if you have the same question.
Updated: 3 Mar 2016 6:16 EST
Pegasystems
IN
The ports 5701 onwards for Hazelcast and 9300 onwards for ElasticSearch is for the inter node communication between all the nodes. These ports need not be exposed outside of the local network on which these nodes reside.
There are prconfig settings available which can override the ports to be used if the default ports are reserved for some other purpose in your machine / network.
Elastic Search - indexing/distributed/transport/port
Hazelcast - cluster/hazelcast/ports
Updated: 3 Mar 2016 6:16 EST
JPMC
US
Thank you Rajiv Nistala for your response!
Could you please answer my query - If we have non-cluster setup with multiple nodes then do we need to do these changes regarding cluster communication and elastic search?
What if we do not create these DSS and upgrade system in Production without worrying about this setup and port availability?
Updated: 3 Mar 2016 6:16 EST
Pegasystems
IN
By cluster setup you mean application server cluster or cluster as defined by Pega (which is all nodes using the same PegaRULES schema)?
Updated: 3 Mar 2016 6:16 EST
JPMC
US
Yes Rajiv, I was referring to application server cluster. Currently I have opened a SR and following up with Ian Fitzpatrick on the same.
Updated: 3 Mar 2016 6:16 EST
Pegasystems
IN
The opening of these ports is irrespective of whether you have application server cluster or not.
Hi Dhirendra,
We have separated your discussion to a new post. Please let us know the SR number so that we may track it here.
Regards,
Lochan | Online Community Moderator | Pegasystems Inc.
Accepted Solution
Pegasystems Inc.
US
Upon reviewing the associated SR, I see that SA-23077 has been created to resolve this. Please view "Service requestor could not be authenticated" if you have the same question.
Cognizant
IN
When we update our applications to 7.3.X from a lower version of Pega7; we may be expected to run the following SQLs and handover the results to Pegasystems. They verify...
- If the ports allocated for Hazelcast ( default 5701 ) & Elastic search ( default 9300 ) are not already in use; for a different purpose
- If Hazelcast & Elastic search are using any other ports; that are not allocated for them.
Verification Query for Hazelcast nodes ( 5701 – 5800 )
select pyClusterAddress from PegaDATA.pr_sys_statusnodes;
Verification Query for Elastic search nodes ( 9300 – 9399 )
select pyIndexerAddress from PegaDATA.pr_sys_statusnodes;