Uploaded image for project: 'MariaDB MaxScale'
  1. MariaDB MaxScale
  2. MXS-444

Binlog server: Useful information to provide via maxinfo

    XMLWordPrintable

Details

    • New Feature
    • Status: Closed (View Workflow)
    • Minor
    • Resolution: Won't Do
    • 1.3.0
    • N/A
    • binlogrouter, maxinfo
    • None

    Description

      The binlog router provides several metrics via the output of maxadmin show services.
      These need to be parsed and processed and are not in a very useful format.

      It would be convenient to be able to access this status information using the new maxinfo infrastructure, probably in several "tables". The following values below imply having a single master whereas in reality MaxScale could have more than one configured binlog router. For now I'm only assuming the case of having a single active connection.

      The type of thing that I would like to see are:

      • number of connected clients via the SQL interface
      • number of connected slaves via the SQL interface
      • connected_to_master (values: 0 / 1) to indicate if I'm connected to the master
      • number of packets received from the master
      • number of connections to the master
      • number of disconnections from the master
      • number of events received from the master
      • number of transactions received from the master (transaction safety functionality)
      • number of heartbeats events received from the master
      • connection time to the master (in seconds)

      For the slaves we basically want the same values. MaxScale can not at the moment do aggregations so you probably want to create a table with the current connected slaves which looks like this:

      server-id
      hostname
      slave-uuid
      slave_host_post
      username
      state
      binlog_file
      binlog_pos
      num_requests
      num_events_sent
      num_bytes_sent
      num_bursts_sent
      num_transitions_to_follow_mode
      num_flow_control
      num_up_to_date
      num_drained_cbs
      num_failed_reads
      last_binlog_event_timestamp
      seconds_behind_master
      slave_mode

      Given I can't do SELECT sum(...), Sum(...) from ... GROUP BY ... then having a sum_all_slaves table with the sum of the above metrics for all active slaves, would probably be good. Don't forget to add an additional count_slaves column so that we know for how many slaves the values are being shown.

      Providing this sort of output via some maxinfo structures would be good.

      Attachments

        Activity

          People

            Unassigned Unassigned
            simon.mudd Simon J Mudd
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Git Integration

                Error rendering 'com.xiplink.jira.git.jira_git_plugin:git-issue-webpanel'. Please contact your Jira administrators.