sh.status() hangs forever if first configdb is unreachable

XMLWordPrintableJSON

    • Type: Bug
    • Resolution: Won't Fix
    • Priority: Major - P3
    • None
    • Affects Version/s: None
    • Component/s: Shell
    • None
    • Environment:
      Debian Linux 6.0 (Squeeze) / 10gen apt repository
    • ALL
    • Hide

      1) Start three configdb servers.
      2) Start mongos with --configdb server1:27019,server2:27019,server3:27019
      3) Connect to mongos with mongo shell.
      4) Run sh.status(), verify success.
      5) Disable networking on server1 to create a network partition.
      6) Run sh.status(), expect infinite hang.

      Show
      1) Start three configdb servers. 2) Start mongos with --configdb server1:27019,server2:27019,server3:27019 3) Connect to mongos with mongo shell. 4) Run sh.status(), verify success. 5) Disable networking on server1 to create a network partition. 6) Run sh.status(), expect infinite hang.
    • None
    • 3
    • None
    • None
    • None
    • None
    • None
    • None

      Running sh.status() from a mongos instance configured with three configdb servers can hang forever (tested hang for about 5 minutes), if the first configured configdb server is unreachable.

      Various monitoring systems use this command to check status and therefore causes them to hang.

      It should be noted that queries to shards continue to work normally despite the failure of sh.status().

            Assignee:
            DO NOT USE - Backlog - Platform Team
            Reporter:
            Christian Hergert (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

              Created:
              Updated:
              Resolved: