What is the bug?
When using a JOIN timeout hint with a PIT-enabled cluster, the PIT will get deleted before the timeout expires.

Caused by: java.lang.IllegalArgumentException: Point In Time id: [y_T7MwDnSG-rblpw9Xa6mQ] doesn't exist

This is due to a 1-minute keepalive

Used by PIT at

How can one reproduce the bug?
Steps to reproduce the behavior:

1. Create indices that require more than 60s to join on (>5 mil records each)
2. Configure the cluster to use PIT joins per

   sql/docs/dev/opensearch-pagination.md

   Line 497 in 35518dc

   "plugins.sql.pagination.api" : "true"

3. Run a join query with a longer join timeout hint /*! JOIN_TIME_OUT(90) */
4. See in the logs that the join failed because the PIT no longer exists (deleted by the 1m keepalive)

What is the expected behavior?
The JOIN_TIME_OUT should also update the keepalive

What is your host/environment?

• OS: Latest

Do you have any screenshots?
N/A

Do you have any additional context?
Add any other context about the problem.