Uploaded image for project: 'MariaDB MaxScale'
  1. MariaDB MaxScale
  2. MXS-5526

Maxscale breaks connection complaining the 'retry_failed_reads' is disabled while it is enabled

    XMLWordPrintable

Details

    • Bug
    • Status: Closed (View Workflow)
    • Major
    • Resolution: Incomplete
    • 24.02.2
    • N/A
    • readwritesplit
    • None
    • MariaDB 10.5.19 on Ubuntu 20.04.6 x64

    Description

      Galera4 cluster on 5 nodes, one is going offline to make backup every night
      Service with readwritesplit router
      ===========

      [PROrw]
      type=service                                                                                         
      router=readwritesplit                                                                                
      servers=pro1,pro2,pro3,pro4,pro5
      slave_selection_criteria=ADAPTIVE_ROUTING
      master_reconnection=yes                                                                              
      master_failure_mode=fail_on_write                                                                    
      lazy_connect=no
      reuse_prepared_statements=yes                                                                        
      transaction_replay=yes                                                                               
      transaction_replay_max_size=2M
      transaction_replay_retry_on_deadlock=no                                                              
      transaction_replay_checksum=full                                                                     
      transaction_replay_timeout=30s                                                                                                                                                                            master_accept_reads=yes
      enable_root_user=yes                                                                                 
      connection_timeout=9999s                                                                             
      optimistic_trx=true
      retry_failed_reads=true
      

      =====

      complains that

      2025-02-22 01:24:39   error  : (42550243) (PROrw); Cannot retry read because 'retry_failed_reads' is disabled and a query was active on 'pro5' when it failed: Lost connection to backend server: Server is no longer usable (pro5, session=42550243, conn_id=44811166: Generated event) Error from 'pro5': Lost connection to backend server: Server is no longer usable (pro5, session=42550243, conn_id=44811166: Generated event)
       
      2025-02-24 01:24:48   error  : (49331853) (PROrw); Cannot retry read because 'retry_failed_reads' is disabled and a query was active on 'pro5' when it failed: Lost connection to backend server: Server is no longer usable (pro5, session=49331853, conn_id=53964285: Generated event) Error from 'pro5': Lost connection to backend server: Server is no longer usable (pro5, session=49331853, conn_id=53964285: Generated event)
      

      $ maxctrl show service PROrw|grep retry_failed_reads
                                "retry_failed_reads": true,   
      

      Attachments

        Activity

          People

            markus makela markus makela
            CTAC Stanislav Sukholet
            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.