Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-19257

Configure ReplicaSetMonitorWatcher recv timeout

    XMLWordPrintableJSON

Details

    • Icon: Improvement Improvement
    • Resolution: Unresolved
    • Icon: Major - P3 Major - P3
    • None
    • 3.0.4
    • Sharding
    • Cluster Scalability

    Description

      RE: https://github.com/mongodb/mongo/blob/r3.1.5/src/mongo/client/replica_set_monitor.cpp#L75

      The ReplicaSetMonitorWatcher thread has a hard-coded socket timeout of 5 seconds. It would be convenient in several ways if this could be configured as it would permit some coarse control over when mongos triggers re-routing decisions - particularly where tag-sets are involved. For a co-located mongos (co-located with all shards) might be appropriate to set a lower timeout, while a remotely located mongos might be better to set a little higher.

      Attachments

        Activity

          People

            backlog-server-cluster-scalability Backlog - Cluster Scalability
            andrew.ryder@mongodb.com Andrew Ryder (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            7 Start watching this issue

            Dates

              Created:
              Updated: