Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-33011

mariabackup --backup: FATAL ERROR: ... Can't open datafile cool_down/t3

    XMLWordPrintable

Details

    Description

      The problem was found on
         origin/10.6-MDEV-32939 028a35b50640fe94e659d58fe8c7241af2696e89
      but Marko means its not caused by the modifications for 10.6-MDEV-32939.
       
      Scenario:
      1. Start the source DB server, generate some unimportant initial data
      2. One connection runs a random mix of
           - CREATE TABLE <table> IF NOT EXISTS ...
           - RENAME TABLE <table>  TO cool_down . <table>
           Please note that the schema "cool_down" was never created.
      3. During 2. is ongoing some "mariabackup --backup" connects to the source DB server
           and wants to backup data to some other location.
      # 2023-12-13T13:53:51 [1677124] | [rr 1712197 19322][00] 2023-12-13 13:53:51 Finished backing up non-InnoDB tables and files
      # 2023-12-13T13:53:51 [1677124] | [rr 1712197 19406][01] 2023-12-13 13:53:51 Copying /data/results/1702475558/8/1/data//aria_log.00000001 to /data/results/1702475558/8/1_clone/data/aria_log.00000001
      # 2023-12-13T13:53:51 [1677124] | [rr 1712197 19426][01] 2023-12-13 13:53:51         ...done
      # 2023-12-13T13:53:51 [1677124] | [rr 1712197 19508][01] 2023-12-13 13:53:51 Copying /data/results/1702475558/8/1/data//aria_log_control to /data/results/1702475558/8/1_clone/data/aria_log_control
      # 2023-12-13T13:53:51 [1677124] | [rr 1712197 19528][01] 2023-12-13 13:53:51         ...done
      # 2023-12-13T13:53:51 [1677124] | [rr 1712197 19855][00] 2023-12-13 13:53:51 Waiting for log copy thread to read lsn 80003
      # 2023-12-13T13:53:51 [1677124] | [rr 1712197 19954][00] FATAL ERROR: 2023-12-13 13:53:51 Can't open datafile cool_down/t3
       
      It wonders me that some
           RENAME <existing table> TO <not existing schema> . <tablename> which should fail during parsing/check prerequisites
           gets memorized by the system at all.
           And if at all than why as something which was executed with success/needs attention when making a backup?
       
      pluto:/data/results/1702475558/MB-1270$ _RR_TRACE_DIR=./1_clone/rr rr replay --mark-stdio # mariabackup --backup --> Can't open datafile cool_down/t3
      pluto:/data/results/1702475558/MB-1270$ _RR_TRACE_DIR=./1/rr rr replay --mark-stdio # source DB server
       
      RQG
      ------
      # git clone https://github.com/mleich1/rqg --branch <pick the right branch> RQG
      #
      # GIT_SHOW: HEAD -> master, origin/master, origin/HEAD 1ed8eb38d6f08d76cb64aebdec71e668389df20f 2023-12-04T18:55:11+01:00
      # rqg.pl  : Version 4.4.0 (2023-08)
      #
      # $RQG_HOME/rqg.pl \
      # --duration=300 \
      # --queries=10000000 \
      # --no_mask \
      # --seed=random \
      # --engine=InnoDB \
      # --gendata_sql=conf/mariadb/table_stress.sql \
      # --gendata=conf/mariadb/table_stress.zz \
      # --max_gd_duration=1200 \
      # --rpl_mode=none \
      # --mysqld=--plugin-load-add=file_key_management.so \
      # --mysqld=--innodb-buffer-pool-size=5M \
      # --mysqld=--lock-wait-timeout=86400 \
      # --mysqld=--loose-table_lock_wait_timeout=50 \
      # --mysqld=--file-key-management-filename=$RQG_HOME/conf/mariadb/encryption_keys.txt \
      # --mysqld=--log-output=none \
      # --mysqld=--loose-debug_assert_on_not_freed_memory=0 \
      # --mysqld=--wait_timeout=28800 \
      # --mysqld=--connect_timeout=60 \
      # --mysqld=--net_write_timeout=60 \
      # --mysqld=--innodb-lock-wait-timeout=50 \
      # --mysqld=--innodb_page_size=4K \
      # --mysqld=--interactive_timeout=28800 \
      # --mysqld=--slave_net_timeout=60 \
      # --mysqld=--log-bin \
      # --mysqld=--sync-binlog=1 \
      # --mysqld=--loose-plugin-load-add=provider_lz4.so \
      # --mysqld=--net_read_timeout=30 \
      # --mysqld=--loose-max-statement-time=30 \
      # --mysqld=--loose-idle_readonly_transaction_timeout=0 \
      # --mysqld=--loose_innodb_lock_schedule_algorithm=fcfs \
      # --mysqld=--loose-innodb_read_only_compressed=OFF \
      # --mysqld=--loose-idle_write_transaction_timeout=0 \
      # --mysqld=--sql_mode=STRICT_TRANS_TABLES,ERROR_FOR_DIVISION_BY_ZERO,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION \
      # --mysqld=--loose-innodb-log-file-size=200M \
      # --mysqld=--loose-idle_transaction_timeout=0 \
      # --mysqld=--log_bin_trust_function_creators=1 \
      # --mysqld=--loose-innodb_fatal_semaphore_wait_threshold=300 \
      # --reporters=Deadlock,Mariabackup_linux,None \
      # --validators=None \
      # --grammar=MB-1270.yy \
      # --threads=1 \
      # <local settings>
       
       
      
      

      Attachments

        1. 10.5.patch
          7 kB
        2. 10.6.patch
          8 kB
        3. MB-1270.cfg
          45 kB
        4. MB-1270.yy
          0.5 kB

        Issue Links

          Activity

            People

              debarun Debarun Banerjee
              mleich Matthias Leich
              Votes:
              0 Vote for this issue
              Watchers:
              3 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.