In our multi node environment we are considering not having a search node (<node-type> =< search>). if we don't have a specific dedicated search node(s) (the standard Pega behaviour is that the first node that starts become the search node), and to avoided the first node that start becoming the search index node, is it ok to configure <-Dindex.directory="server/searchindexdir/nsomedexdir> to point to a specific node in the cluster?
So in layman terms, if I don't specify a search node, but specify which node-type the search index will be on, will that cause any issues that you know of?
***Edited by Moderator Marissa to update platform capability tags****
What version of Pega is in use? That may impact the answer to this.
If you specify the -Dindex.directory argument with a valid directory, that node is going to startup as an indexing host node and start indexing whether or not it is a Search node (I believe, but could be version specific). The Search node type I don't believe dictates whether or not a node can host indexes, but rather where other processing occurs, like the FTSIncrementalIndexer.
If you don't want any of your nodes to host indexes, you should set -Dindex.directory but provide no value.