Uploaded image for project: 'Couchbase Server'
  1. Couchbase Server
  2. MB-47129

[BP 6.6.3] - Extreme interval between checkInactiveUprStreams

    XMLWordPrintable

Details

    • Untriaged
    • Unknown

    Description

      The inactive streams are reattempted every 20minutes, it doesn't look optimal. 

      Attachments

        Issue Links

          No reviews matched the request. Check your Options in the drop-down menu of this sections header.

          Activity

            Build couchbase-server-6.6.3-9777 contains goxdcr commit 215a80a with commit message:
            MB-47129 - shorten DCP inactive timeout wait time to 5 minutes instead of 20

            build-team Couchbase Build Team added a comment - Build couchbase-server-6.6.3-9777 contains goxdcr commit 215a80a with commit message: MB-47129 - shorten DCP inactive timeout wait time to 5 minutes instead of 20

            I see "MaxCountDCPStreamsInactive":10 on 6.6.3-9777. Neil Huang, does this translate to 5 mins after some calculation?

            pavithra.mahamani Pavithra Mahamani (Inactive) added a comment - I see "MaxCountDCPStreamsInactive":10 on 6.6.3-9777. Neil Huang , does this translate to 5 mins after some calculation?
            neil.huang Neil Huang added a comment -

            It is 30 seconds (dcp_inactive_stream_check_interval) * MaxCountStreamsInactive (10) == 300 seconds -> 5 minutes.

            neil.huang Neil Huang added a comment - It is 30 seconds (dcp_inactive_stream_check_interval) * MaxCountStreamsInactive (10) == 300 seconds -> 5 minutes.

            Description for release notes:
            Summary: Due to a longer interval between checks for inactive DCP streams, changes from some vBuckets were getting stuck and not being propagated. Also, the XDCR mutations left statistic was seen to steadily increase on all buckets. This has been fixed by reducing the DCP inactive timeout wait time to 5 minutes instead of 20 minutes.

            amarantha.kulkarni Amarantha Kulkarni (Inactive) added a comment - Description for release notes: Summary : Due to a longer interval between checks for inactive DCP streams, changes from some vBuckets were getting stuck and not being propagated. Also, the XDCR mutations left statistic was seen to steadily increase on all buckets. This has been fixed by reducing the DCP inactive timeout wait time to 5 minutes instead of 20 minutes.

            People

              pavithra.mahamani Pavithra Mahamani (Inactive)
              neil.huang Neil Huang
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved:

                Gerrit Reviews

                  There are no open Gerrit changes

                  PagerDuty