I am getting the “Unable to contact search service” on my Spark 2.6.3 client while trying to search a remote 3.7.1 server with the search 1.5.1 plugin. We are always trying to conserve bits across our satellite links so the server sessions are configured to tear themselves down after five minutes. These are high latency links so the sessions do take some time to become bi-directional again. If the server session is not already active, the Spark client search returns the error. If I keep refreshing the server sessions on the local server, I can see the session become bi-directional shortly after the error on the client. At this point I can run the search without a problem. Is there any way to bump up the timeout value for searches on the client?
Thanks,
Scott