Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-24796

Distlock acquisition must not retry at the level of individual 'findAndModify' command

    XMLWordPrintableJSON

Details

    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Major - P3 Major - P3
    • 3.3.10
    • None
    • Sharding
    • None
    • Fully Compatible
    • ALL
    • 0

    Description

      Currently, the dist lock catalog retries the 'findAndModify' operations on NotMaster errors. In addition, in the dist lock manager there is a retry logic for the same set of errors.

      The calls inside the dist lock catalog must be made to not retry because if the previous write succeeded, it would return a status indicating that the lock is busy, when in fact it is actually not. This throws off the outer retry logic.

      Attachments

        Activity

          People

            kaloian.manassiev@mongodb.com Kaloian Manassiev
            randolph@mongodb.com Randolph Tan
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: