Investigate document count mismatch on deleteMany of sharded timeseries concurrent with upgrade/downgrade

XMLWordPrintableJSON

    • Type: Task
    • Resolution: Unresolved
    • Priority: Major - P3
    • None
    • Affects Version/s: 9.0.0-rc0
    • Component/s: None
    • Catalog and Routing
    • 🟩 Routing and Topology
    • None
    • None
    • None
    • None
    • None
    • None

      Investigate why the timeseries_delete_compressed_buckets.js may fail to delete all documents when running in the fcv_upgrade_downgrade_sharded_collections_jscore_passthrough suite with viewless timeseries upgrade downgrade.

       

      A factor may be SERVER-20361, since we have a deleteMany on a sharded collection we set ShardVersion::IGNORED. If the shards don't re-target this delete to the correct main/buckets NSS if a concurrent upgrade/downgrade happens, this may explain this issue.

            Assignee:
            Unassigned
            Reporter:
            Joan Bruguera Micó
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated: