Uploaded image for project: 'WiredTiger'
  1. WiredTiger
  2. WT-6955

Test restart performance with large history store due to a pinned stable timestamp

    XMLWordPrintable

    Details

    • Type: Task
    • Status: Open
    • Priority: Major - P3
    • Resolution: Unresolved
    • Affects Version/s: None
    • Fix Version/s: Backlog
    • Component/s: None
    • Labels:
      None

      Description

      I recently looked at MongoDB performance when it can't advance the majority commit point and it accumulates a lot of updates in the WT history store and oplog. (See WT-6776 and WT-6786.)  

      We should test a related corner case.  In a PSA replica set where the secondary has failed and the history store grows a lot because MongoDB isn't advancing the stable timestamp, what happens if the primary then fails?  In particular, what is the effect on the recovery time for the primary.

      As I understand, WT will run rollback-to-stable and need to get rid of most of that accumulated state in the history store.  So this is essentially asking how long RTS takes with a large (multi GB) HS file.

        Attachments

          Activity

            People

            Assignee:
            backlog-server-storage-engines Backlog - Storage Engines Team
            Reporter:
            keith.smith Keith Smith
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

              Dates

              Created:
              Updated: