Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-25992

Galera 3 Server crash with signal 6 after RBR event apply failed

    XMLWordPrintable

Details

    • Bug
    • Status: Closed (View Workflow)
    • Major
    • Resolution: Fixed
    • 10.3.29
    • 10.3.32

    Description

      We have a 3 node cluster in our UAT environment, with all traffic going to node 3, an incoming delete conflict causes nodes 1 & 2 to crash. This causes node 3 to go non-primary (as expected).

      The crash is always on the non-write nodes (either one of them or both crash) that are applying the deletes concurrently.

      the delete is always on the same table named "blobs" with a self-referencing foreign key:

      Of note, the logged SQL for the conflict appears to have some garbled data on the end of it (that I can't quite capture in this form):

      "SQL: DELETE FROM blobs WHERE id = '7432858'???`^S^F"

      The table has been rebuilt with an alter table engine = innodb, yet the issue still occours.

      The crashes started 5 days after we upgraded from:

      10.3.24-MariaDB-1:10.3.24+maria~bionic-log
      patch: wsrep_25.24
      prov: 25.3.29(r3902)

      ... and on a related note but I must emphasise different cluster entirely; our Production cluster which has yet to be upgraded (version as above) is uttering "[ERROR] InnoDB: Record field 15 len 18446744073709551615" which I've traced back to the same collection of tables in the same schema. I've been unable to identify any corruption in the tables themselves (by selecting out data and forcing index usage). The UAT cluster for which this report relates hasn't uttered these messaged, but I've a sneeking suspicion there is some, even if loose relationship between the issues.

      Detail of the crash in the UAT env

      *************************** 1. row ***************************
             Table: blobs
      Create Table: CREATE TABLE `blobs` (
        `id` int(11) NOT NULL AUTO_INCREMENT,
        `original_blob_id` int(11) DEFAULT NULL,
        `sys_name` varchar(100) DEFAULT NULL,
        `storage_loc` varchar(50) DEFAULT NULL,
        `storage_loc_pref` varchar(50) DEFAULT NULL,
        `storage_loc_specific` varchar(50) DEFAULT NULL,
        `save_path` varchar(255) DEFAULT NULL,
        `file_url` varchar(255) DEFAULT NULL,
        `filename` varchar(255) NOT NULL,
        `filesize` int(11) NOT NULL,
        `content_type` varchar(50) NOT NULL,
        `authcode` varchar(50) NOT NULL,
        `blob_hash` varchar(40) NOT NULL,
        `is_media_upload` tinyint(1) NOT NULL,
        `title` varchar(255) NOT NULL,
        `dim_w` int(11) NOT NULL,
        `dim_h` int(11) NOT NULL,
        `date_created` datetime NOT NULL,
        `is_temp` tinyint(1) NOT NULL,
        PRIMARY KEY (`id`),
        KEY `IDX_896C3E356BBE2052` (`original_blob_id`),
        KEY `authcode_idx` (`authcode`),
        KEY `storage_loc_idx` (`storage_loc`,`storage_loc_pref`),
        KEY `sys_name_idx` (`sys_name`),
        KEY `date_created_idx` (`date_created`,`is_temp`),
        KEY `storage_loc_pref_idx` (`storage_loc_pref`),
        CONSTRAINT `FK_896C3E356BBE2052` FOREIGN KEY (`original_blob_id`) REFERENCES `blobs` (`id`) ON DELETE SET NULL
      ) ENGINE=InnoDB AUTO_INCREMENT=7700386 DEFAULT CHARSET=utf
      

      node 1 crash:

       
      2021-06-22 12:35:02 11 [Note] WSREP: cluster conflict due to high priority abort for threads:
      2021-06-22 12:35:02 11 [Note] WSREP: Winning thread:
         THD: 11, mode: applier, state: executing, conflict: no conflict, seqno: 89327026
         SQL: DELETE FROM blobs WHERE id = '7432858'???`^S^F
      2021-06-22 12:35:02 11 [Note] WSREP: Victim thread:
         THD: 15, mode: applier, state: idle, conflict: no conflict, seqno: 89327028
         SQL: NULL
      2021-06-22 12:35:02 0 [ERROR] WSREP: Trx 89327026 tries to abort slave trx 89327028. This could be caused by:
              1) unsupported configuration options combination, please check documentation.
              2) a bug in the code.
              3) a database corruption.
       Node consistency compromized, need to abort. Restart the node to resync with cluster.
      210622 12:35:02 [ERROR] mysqld got signal 6 ;
      This could be because you hit a bug. It is also possible that this binary
      or one of the libraries it was linked against is corrupt, improperly built,
      or misconfigured. This error can also be caused by malfunctioning hardware.
       
      To report this bug, see https://mariadb.com/kb/en/reporting-bugs
       
      We will try our best to scrape up some info that will hopefully help
      diagnose the problem, but since we have already crashed,
      something is definitely wrong and this may fail.
       
      Server version: 10.3.29-MariaDB-1:10.3.29+maria~bionic-log
      key_buffer_size=134217728
      read_buffer_size=134217728
      max_used_connections=26
      max_threads=1002
      thread_count=24
      It is possible that mysqld could use up to
      key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 262821945 K  bytes of memory
      Hope that's ok; if not, decrease some variables in the equation.
       
      Thread pointer: 0x0
      Attempting backtrace. You can use the following information to find out
      where mysqld died. If you see no messages after this, something went
      terribly wrong...
      stack_bottom = 0x0 thread_stack 0x12c00000
      /usr/sbin/mysqld(my_print_stacktrace+0x2e)[0x560ee4349b8e]
      /usr/sbin/mysqld(handle_fatal_signal+0x515)[0x560ee3dde8c5]
      /lib/x86_64-linux-gnu/libpthread.so.0(+0x12980)[0x7f9ca6c54980]
      /lib/x86_64-linux-gnu/libc.so.6(gsignal+0xc7)[0x7f9ca688ffb7]
      /lib/x86_64-linux-gnu/libc.so.6(abort+0x141)[0x7f9ca6891921]
      /usr/sbin/mysqld(+0x90a914)[0x560ee3f68914]
      /usr/sbin/mysqld(+0x90f83c)[0x560ee3f6d83c]
      /usr/sbin/mysqld(handle_manager+0x1f3)[0x560ee3bf0a83]
      /lib/x86_64-linux-gnu/libpthread.so.0(+0x76db)[0x7f9ca6c496db]
      /lib/x86_64-linux-gnu/libc.so.6(clone+0x3f)[0x7f9ca697271f]
      The manual page at https://mariadb.com/kb/en/how-to-produce-a-full-stack-trace-for-mysqld/ contains
      information that should help you find out what is causing the crash.
      Writing a core file...
      Working directory at /srv/galera-uat/mysql
      Resource Limits:
      Limit                     Soft Limit           Hard Limit           Units
      Max cpu time              unlimited            unlimited            seconds
      Max file size             unlimited            unlimited            bytes
      Max data size             unlimited            unlimited            bytes
      Max stack size            8388608              unlimited            bytes
      Max core file size        0                    unlimited            bytes
      Max resident set          unlimited            unlimited            bytes
      Max processes             15430                15430                processes
      Max open files            4096                 4096                 files
      Max locked memory         67108864             67108864             bytes
      Max address space         unlimited            unlimited            bytes
      Max file locks            unlimited            unlimited            locks
      Max pending signals       15430                15430                signals
      Max msgqueue size         819200               819200               bytes
      Max nice priority         0                    0
      Max realtime priority     0                    0
      Max realtime timeout      unlimited            unlimited            us
      Core pattern: |/usr/share/apport/apport %p %s %c %d %P %E
      

      node 2 crash:

      2021-06-22 12:35:02 15 [Note] WSREP: cluster conflict due to high priority abort for threads:
      2021-06-22 12:35:02 15 [Note] WSREP: Winning thread:
         THD: 15, mode: applier, state: executing, conflict: no conflict, seqno: 89327026
         SQL: DELETE FROM blobs WHERE id = '7432858'???`^S^F
      2021-06-22 12:35:02 15 [Note] WSREP: Victim thread:
         THD: 9, mode: applier, state: executing, conflict: no conflict, seqno: 89327028
         SQL: DELETE FROM blobs WHERE id = '7432867'???`^S^F
      2021-06-22 12:35:02 0 [ERROR] WSREP: Trx 89327026 tries to abort slave trx 89327028. This could be caused by:
              1) unsupported configuration options combination, please check documentation.
              2) a bug in the code.
              3) a database corruption.
       Node consistency compromized, need to abort. Restart the node to resync with cluster.
      210622 12:35:02 [ERROR] mysqld got signal 6 ;
      This could be because you hit a bug. It is also possible that this binary
      or one of the libraries it was linked against is corrupt, improperly built,
      or misconfigured. This error can also be caused by malfunctioning hardware.
       
      To report this bug, see https://mariadb.com/kb/en/reporting-bugs
       
      We will try our best to scrape up some info that will hopefully help
      diagnose the problem, but since we have already crashed,
      something is definitely wrong and this may fail.
       
      Server version: 10.3.29-MariaDB-1:10.3.29+maria~bionic-log
      key_buffer_size=134217728
      read_buffer_size=134217728
      max_used_connections=8
      max_threads=1002
      thread_count=23
      It is possible that mysqld could use up to
      key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 262821945 K  bytes of memory
      Hope that's ok; if not, decrease some variables in the equation.
       
      Thread pointer: 0x0
      Attempting backtrace. You can use the following information to find out
      where mysqld died. If you see no messages after this, something went
      terribly wrong...
      2021-06-22 12:35:02 9 [Warning] WSREP: conflict state after RBR event applying: 1, 89327028
      2021-06-22 12:35:02 9 [Warning] WSREP: RBR event apply failed, rolling back: 89327028
      stack_bottom = 0x0 thread_stack 0x12c00000
      /usr/sbin/mysqld(my_print_stacktrace+0x2e)[0x5591922ceb8e]
      /usr/sbin/mysqld(handle_fatal_signal+0x515)[0x559191d638c5]
      /lib/x86_64-linux-gnu/libpthread.so.0(+0x12980)[0x7ffb20e45980]
      /lib/x86_64-linux-gnu/libc.so.6(gsignal+0xc7)[0x7ffb20a80fb7]
      /lib/x86_64-linux-gnu/libc.so.6(abort+0x141)[0x7ffb20a82921]
      /usr/sbin/mysqld(+0x90a914)[0x559191eed914]
      /usr/sbin/mysqld(+0x90f83c)[0x559191ef283c]
      /usr/sbin/mysqld(handle_manager+0x1f3)[0x559191b75a83]
      /lib/x86_64-linux-gnu/libpthread.so.0(+0x76db)[0x7ffb20e3a6db]
      /lib/x86_64-linux-gnu/libc.so.6(clone+0x3f)[0x7ffb20b6371f]
      The manual page at https://mariadb.com/kb/en/how-to-produce-a-full-stack-trace-for-mysqld/ contains
      information that should help you find out what is causing the crash.
      Writing a core file...
      Working directory at /srv/galera-uat/mysql
      Resource Limits:
      Limit                     Soft Limit           Hard Limit           Units
      Max cpu time              unlimited            unlimited            seconds
      Max file size             unlimited            unlimited            bytes
      Max data size             unlimited            unlimited            bytes
      Max stack size            8388608              unlimited            bytes
      Max core file size        0                    unlimited            bytes
      Max resident set          unlimited            unlimited            bytes
      Max processes             15430                15430                processes
      Max open files            4096                 4096                 files
      Max locked memory         67108864             67108864             bytes
      Max address space         unlimited            unlimited            bytes
      Max file locks            unlimited            unlimited            locks
      Max pending signals       15430                15430                signals
      Max msgqueue size         819200               819200               bytes
      Max nice priority         0                    0
      Max realtime priority     0                    0
      Max realtime timeout      unlimited            unlimited            us
      Core pattern: |/usr/share/apport/apport %p %s %c %d %P %E
      

      Attachments

        Issue Links

          Activity

            People

              jplindst Jan Lindström (Inactive)
              glynastill Glyn Astill
              Votes:
              3 Vote for this issue
              Watchers:
              7 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.