Uploaded image for project: 'Blazegraph (by SYSTAP)'
  1. Blazegraph (by SYSTAP)
  2. BLZG-757

HA3 LOAD non-responsive with node failure.

    XMLWordPrintable

    Details

    • Type: Bug
    • Status: Closed - Won't Fix
    • Resolution: Cannot Reproduce
    • Affects Version/s: BIGDATA_RELEASE_1_2_2
    • Fix Version/s: None
    • Component/s: HAJournalServer
    • Labels:
      None

      Description

      A problem was observed where a LOAD in an HA3 configuration became non-responsive. This situation was eventually cured, and the load completed normally on 2 nodes. After a restart, the 3rd node resynchronized automatically. However, I am not sure why the LOAD was blocked for as long as it was without making progress. Also, GET /status was blocking on the AbstractQuorum.lock(). The thread handling the pipelineRemove() event was also blocked waiting for that lock. I have not identified the thread that was actually running with the lock held yet. A stack trace is attached for the leader. This is against r7059.

        Attachments

          Activity

            People

            Assignee:
            bryanthompson bryanthompson
            Reporter:
            bryanthompson bryanthompson
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Dates

              Created:
              Updated:
              Resolved: