Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-44506

Retry jstestfuzz concurrency preamble commands

    XMLWordPrintable

    Details

    • Type: Task
    • Status: Needs Scheduling
    • Priority: Major - P3
    • Resolution: Unresolved
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: Replication
    • Labels:
      None
    • Linked BF Score:
      15

      Description

      The jstestfuzz concurrency stepdown suite consists of 10 jstest clients that run commands against the same replica set. These jstest clients set up a connection to a replica set, run some preamble code, and then start running test statements against the replica set.

      Currently, it is possible that a client lags behind and enters the preamble a little behind the others. At this point, the other clients have started running test statements and it is reasonable that a node enters rollback after a stepdown command. If the lagging client starts to run the preamble code and issues commands on the rolled back node, we will fail with a network error. We should find a way to be able to retry these commands in the preamble portion of the generated jstest.

      This could be as simple as adding the network_error_and_txn_override but it may not make sense to wrap all these commands in transactions so some refactoring might be required if this is the chosen approach.

        Attachments

          Activity

            People

            • Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

              Dates

              • Created:
                Updated: