How elastic search works?
We are upgrading from Pega 7.1.7 to Pega 7.1.9 and changing the indexing mechanism from lucene to elastic.
We previously used to have customised solution where the searching of indexes used to be directly from each node to the shared mount accessible on all the nodes.
With elastic search we would like to use the distributed way.
Few questions:-
1. What is the minimum number of nodes that needs to serve as indexing nodes?
2. Host nodes for indexing will be used for both generation and retrieval as well?
3. Would like to understand how the search works with elastic indexes from a node that is not the host node. Does it check initially whether the location is directly accessible or always goes to the host nodes?
Accepted Solution
Pegasystems
IN
1. What is the minimum number of nodes that needs to serve as indexing nodes?
One node is sufficient if you don't need failover and high availability. But for production systems, having more indexing nodes does help.
2. Host nodes for indexing will be used for both generation and retrieval as well?
Host nodes host the index files (or replicas). Searches also go through those nodes. Note that you don't need to run the incremental indexing agent only on those nodes. They can run on all nodes.
3. Would like to understand how the search works with elastic indexes from a node that is not the host node. Does it check initially whether the location is directly accessible or always goes to the host nodes?
Elastic Search uses its own protocol to communicate between the nodes over port range 9300~9399. You can find the port number in pr_sys_statusnodes table (pyIndexerAddress column).
Pegasystems Inc.
US
1. Depends on your work load. If you have enabled WO- indexing and lot of work objects are being created / touched distributing the load across a 2 - 3 indexing hosts is ideal. This helps FTSIncrementalIndexer agent from backing out indexing events.
2. Yes. The search query in any node will internally end up in any of the index hosts.
3. Please refer to link - https://www.elastic.co/guide/en/elasticsearch/guide/current/_distributed_nature.html
Accepted Solution
Pegasystems
IN
1. What is the minimum number of nodes that needs to serve as indexing nodes?
One node is sufficient if you don't need failover and high availability. But for production systems, having more indexing nodes does help.
2. Host nodes for indexing will be used for both generation and retrieval as well?
Host nodes host the index files (or replicas). Searches also go through those nodes. Note that you don't need to run the incremental indexing agent only on those nodes. They can run on all nodes.
3. Would like to understand how the search works with elastic indexes from a node that is not the host node. Does it check initially whether the location is directly accessible or always goes to the host nodes?
Elastic Search uses its own protocol to communicate between the nodes over port range 9300~9399. You can find the port number in pr_sys_statusnodes table (pyIndexerAddress column).
Areteans Technology Solutions Australia Pty Ltd.
IN
Hi,
Can full “Rule Content Search” be achieved by Pega OOTB “Search” functionalities ?
For example , we want to search the extract Rule Name and its details where a StepPage (e.g "MystepPage") , a hardcoded condition check (e.g if(X==”abc”)), parameter value passed (e.g param.EmpID) etc (all these given as input) is used throughout application.
If Pega OOTB Search functionalities provide these then please refer us how can we configure this in my application?
Pegasystems
IN
Copying answer from - https://collaborate.pega.com/question/elastic-search-info
In designer studio search, you can use the "All Content" option to search if the content of the rule has some specific text or not. You might want to refer to this PDN document - https://community.pega.com/knowledgebase/release-note/report-definition-query-filters-can-search-embedded-properties. Note that special characters like equals (=) or double quotes (") are not indexed so you cannot search for the presence of those characters in the content.