Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-8315

recurrently we got LOST_EVENTS on slave

        Details

        • Type: Bug
        • Status: Closed
        • Priority: Major
        • Resolution: Fixed
        • Affects Version/s: 10.0
        • Fix Version/s: N/A
        • Component/s: Documentation, Replication
        • Labels:
        • Environment:
          Ubuntu 12.04.5 LTS
          Master slave relication, without GTID, parallell replication and row_based replication

          Description

          After we upgraded to 10.0.19-MariaDB-1~precise-log our replication stopped weekly.

          We got on slave 

          The incident LOST_EVENTS occured on the master. Message: error writing to the binary log

          We use paralell replication without GTID

          Relay_Log_File: mysqld-relay-bin.002349
                Relay_Log_Pos: 588779

          #150612 22:47:09 server id 1  end_log_pos 804903        Query   thread_id=292902        exec_time=0     error_code=0
SET TIMESTAMP=1434142029/*!*/;
COMMIT
/*!*/;
# at 805188
#150612 22:19:29 server id 1  end_log_pos 804956
# Incident: LOST_EVENTS
RELOAD DATABASE; # Shall generate syntax error
# at 805241

            Gliffy Diagrams

              Attachments

                Activity

                Ascending order - Click to sort in descending order
                15 older comments
                Hide
                Permalink
                bulepage bulepage added a comment -

                Can I help ?

                Show
                bulepage bulepage added a comment - Can I help ?
                Hide
                Permalink
                elenst Elena Stepanova added a comment -

                bulepage, did it ever happen again after you had increased the cache size?
                If not, I assume the theory was correct, and it was indeed the reason.

                Show
                elenst Elena Stepanova added a comment - bulepage , did it ever happen again after you had increased the cache size? If not, I assume the theory was correct, and it was indeed the reason.
                Hide
                Permalink
                bulepage bulepage added a comment -

                The event did not occured again.
                I think the tehory is correct.

                Show
                bulepage bulepage added a comment - The event did not occured again. I think the tehory is correct.
                Hide
                Permalink
                elenst Elena Stepanova added a comment -

                In this case it is a configuration problem – naturally, if an event exceeds hard limits, it gets lost.
                Also, apparently the max_binlog_stmt_cache_size description in KB needs some adjustments (default value and limitations), I am assigning it to Ian as a documentation issue.

                Show
                elenst Elena Stepanova added a comment - In this case it is a configuration problem – naturally, if an event exceeds hard limits, it gets lost. Also, apparently the max_binlog_stmt_cache_size description in KB needs some adjustments (default value and limitations), I am assigning it to Ian as a documentation issue.
                Hide
                Permalink
                greenman Ian Gilfillan added a comment -

                The documentation has been updated (it was based on an old MySQL version and a Windows bug limiting the size, not relevant to MariaDB)

                Show
                greenman Ian Gilfillan added a comment - The documentation has been updated (it was based on an old MySQL version and a Windows bug limiting the size, not relevant to MariaDB)

                  People

                  • Assignee:
                    greenman Ian Gilfillan
                    Reporter:
                    bulepage bulepage
                  • Votes:
                    0 Vote for this issue
                    Watchers:
                    3 Start watching this issue

                    Dates

                    • Created:
                      Updated:
                      Resolved: