Uploaded image for project: 'Percona Server'
  1. Percona Server
  2. PS-5574

innodb.log_corruption / innodb.log_file_size_checkpoint regression on trunk

    Details

    • Type: Bug
    • Status: In Progress
    • Priority: Medium
    • Resolution: Unresolved
    • Affects Version/s: 8.0
    • Fix Version/s: None
    • Component/s: None
    • Labels:

      Description

      On 8.0 trunk, introduced between April 11th and 15th:

      190417  2:38:57 [ 55%] innodb.log_corruption                    w2  [ fail ]
              Test ended at 2019-04-17 02:38:57
      
      CURRENT_TEST: innodb.log_corruption
      mysqltest: At line 93: Command "exec" failed with wrong error: 3, my_errno=2. should have failed with error '1'.
      
      The result from queries just before the failure was:
      # redo log from before we started versioning in MySQL 5.7.9
      Pattern "Unsupported redo log format \(0\). The redo log was created before MySQL 5\.7\.9.*" found
      # redo log from before MySQL 8.0.3, with corrupted log block
      Pattern "Upgrade after a crash is not supported.*" found
      # redo log from MySQL 8.0.3, but with invalid header checksum
      Pattern "Invalid redo log header checksum" found
      # distant future redo log format, with valid header checksum
      Pattern "Unknown redo log format \(4294967295\).*" found
      # redo log from MySQL 8.0.3, with corrupted log checkpoint
      Pattern "No valid checkpoint found.*corrupted redo log" found
      # valid 5.7.9 header, valid checkpoint 1, all-zero (invalid) checkpoint 2,
      # invalid block checksum,
      # thus we cannot determine if the redo log is logically clean.
      Pattern "Database upgrade cannot be accomplished with innodb_force_recovery > 0" found
      # --innodb-force-recovery=6 (skip the entire redo log)
      Pattern "Database upgrade cannot be accomplished in read-only mode" found
      # valid 5.7.9 header, valid checkpoint 1, all-zero (invalid) checkpoint 2, invalid block number
      # Start with innodb-force-recovery=5
      Pattern "Database upgrade cannot be accomplished with innodb_force_recovery > 0" found
      # Start with --innodb-force-recovery=6 (skip the entire redo log)
      Pattern "Database upgrade cannot be accomplished in read-only mode" found
      # valid 8.0.3 header, valid checkpoint 1, all-zero (invalid) checkpoint 2, invalid block checksum
      safe_process[27829]: Child process: 27830, exit: 1
      
      
      
      Server [mysqld.1 - pid: 26814, winpid: 26814] log: 
      Server log from this test:
      ----------SERVER LOG START-----------
      ----------SERVER LOG END-------------
      

      and

      90417  3:24:27 [ 98%] innodb.log_file_size_checkpoint          w5  [ fail ]
              Test ended at 2019-04-17 03:24:27
      
      CURRENT_TEST: innodb.log_file_size_checkpoint
      safe_process[3741]: Child process: 3742, killed by signal: 9
      
      
      Server [mysqld.1 - pid: 3747, winpid: 3747, exit: 256] failed during test run
      Server log from this test:
      ----------SERVER LOG START-----------
      ... truncated ...
      

      Robert Golebiowski, please check whether these are related to your pushes at that date range

        Smart Checklist

          Attachments

            Activity

              People

              • Assignee:
                robert.golebiowski Robert Golebiowski
                Reporter:
                laurynas.biveinis Laurynas Biveinis
              • Votes:
                0 Vote for this issue
                Watchers:
                1 Start watching this issue

                Dates

                • Created:
                  Updated:

                  Time Tracking

                  Estimated:
                  Original Estimate - Not Specified
                  Not Specified
                  Remaining:
                  Remaining Estimate - Not Specified
                  Not Specified
                  Logged:
                  Time Spent - 1 day
                  1d