1. 05 Jun, 2013 1 commit
    • unknown's avatar
      MDEV-26: Global transaction ID. · 5cb486d1
      unknown authored
      Fix problems related to reconnect. When we need to reconnect (ie. explict
      stop/start of just the IO thread by user, or automatic reconnect due to
      loosing network connection with the master), it is a bit complex to correctly
      resume at the right point without causing duplicate or missing events in the
      relay log. The previous code had multiple problems in this regard.
      
      With this patch, the problem is solved as follows. The IO thread keeps track
      (in memory) of which GTID was last queued to the relay log. If it needs to
      reconnect, it resumes at that GTID position. It also counts number of events
      received within the last, possibly partial, event group, and skips the same
      number of events after a reconnect, so that events already enqueued before the
      reconnect are not duplicated.
      
      (There is no need to keep any persistent state; whenever we restart slave
      threads after both of them being stopped (such as after server restart), we
      erase the relay logs and start over from the last GTID applied by SQL thread.
      But while the SQL thread is running, this patch is needed to get correct relay
      log).
      5cb486d1
  2. 03 Jun, 2013 1 commit
    • unknown's avatar
      MDEV-4605: Failing to load GTID slave position from rpl.gtid_slave_pos · 7ad47ab0
      unknown authored
      There were several cases where the slave GTID position was not loaded
      correctly before being used. This caused various failures such as
      corrupting the position at slave start and empty values of
      @@gtid_slave_pos and @@gtid_current_pos.
      
      Fixed by adding more checks for loaded position, and by always loading
      the position at server startup.
      7ad47ab0
  3. 29 May, 2013 2 commits
    • unknown's avatar
      MDEV-4485: Incorrect error handling in record_gtid(). · 6feadb10
      unknown authored
      Fix the error handling when access to the table mysql.gtid_slave_pos
      fails for whatever reason. Add some test cases.
      6feadb10
    • unknown's avatar
      MDEV-4485: Master did not allow slave to connect from the very start (empty... · 385780f5
      unknown authored
      MDEV-4485: Master did not allow slave to connect from the very start (empty GTID pos) if GTIDs from other multi_source master was present
      
      The idea in the code was to protect the user that tries to connect a slave
      to a master with completely different domains than what was intended. If
      none of the domains in the start position are present at all in the master
      binlog, we gave an error.
      
      However, this is a stupid idea. Because when a slave connects to a master
      to start replication from the very start of binlogs - such as when setting
      up new master->slave servers from scratch - there will be just this
      situation, the requested slave position is empty for all the domains in the
      master's binlog.
      
      So the code that gives this error is wrong, and the solution is simply to
      remove it.
      385780f5
  4. 28 May, 2013 2 commits
    • unknown's avatar
      3061ca2b
    • unknown's avatar
      MDEV-4478: Implement GTID "strict mode" · ee2b7db3
      unknown authored
      When @@GLOBAL.gtid_strict_mode=1, then certain operations result
      in error that would otherwise result in out-of-order binlog files
      between servers.
      
      GTID sequence numbers are now allocated independently per domain;
      this results in less/no holes in GTID sequences, increasing the
      likelyhood that diverging binlogs will be caught by the slave when
      GTID strict mode is enabled.
      ee2b7db3
  5. 25 May, 2013 1 commit
  6. 24 May, 2013 1 commit
    • unknown's avatar
      MDEV-4475: Replication from MariaDB 10.0 to 5.5 does not work · 416aed25
      unknown authored
      The problem was the Gtid_list event which is logged to the binlog in
      10.0 and is not understood by the 5.5 server.
      
      This event is supposed to be replaced with a dummy event for 5.5
      servers. But the very first event logged in the very first binlog
      has an empty list of GTID, which makes the event too short to be
      replacable with an empty event.
      
      The fix is to pad the empty Gtid_list event to be big enough to
      be replacable by a dummy event.
      416aed25
  7. 22 May, 2013 1 commit
    • unknown's avatar
      MDEV-26: Global transaction ID. · 1cd6eb5f
      unknown authored
      Change of user interface to be more logical and more in line with expectations
      to work similar to old-style replication.
      
      User can now explicitly choose in CHANGE MASTER whether binlog position is
      taken into account (master_gtid_pos=current_pos) or not (master_gtid_pos=
      slave_pos) when slave connects to master.
      
      @@gtid_pos is replaced by three separate variables @@gtid_slave_pos (can
      be set by user, replicated GTIDs only), @@gtid_binlog_pos (read only), and
      @@gtid_current_pos (a combination of the two, most recent GTID within each
      domain). mysql.rpl_slave_state is renamed to mysql.gtid_slave_pos to match.
      
      This fixes MDEV-4474.
      1cd6eb5f
  8. 16 May, 2013 1 commit
    • unknown's avatar
      Fix race condition in binlog dump thread during server shutdown. · d795bc9f
      unknown authored
      There was missing a check for THD::killed after THD::enter_cond(). This could
      cause the binlog dump thread to miss the kill signal during server shutdown
      and hang until it was force-closed.
      
      Also fix a race in a test case that occasionally fails in Buildbot.
      d795bc9f
  9. 15 May, 2013 1 commit
    • unknown's avatar
      MDEV-26: Global transaction ID. · 9fae9930
      unknown authored
      Implement START SLAVE UNTIL master_gtid_pos = "<GTID position>".
      
      Add test cases, including a test showing how to use this to promote
      a new master among a set of slaves.
      9fae9930
  10. 10 May, 2013 2 commits
    • Michael Widenius's avatar
      Merge of patch lp:~ahiguti100/maria/handlersocket-fix-78 by Akira Higuchi · 7202c21b
      Michael Widenius authored
      A bugfix of HandlerSocket is not applied to mariadb yet
      7202c21b
    • Michael Widenius's avatar
      Added some fixes that should make MyISAM & Aria REPAIR work with more than 4G records · 53d44ad1
      Michael Widenius authored
      - If one specifies --force twice to myisamchk and aria_chk, then we will try to finnish the repair even if sort_buffer would be too small.
        This was done by dynamically allocate buffer handler objects as long as memory lasts.
      - New option for myisamchk and aria_chk: create-missing-keys
      - Changed default size of myisam_sort_buffer_size from 8M to 128M.
      - Changed default size of sort_buffer_size in aria_chk from 128M to 256M.
      - Increased information in error message about 'sort_buffer_size' beeing to small.
      - Print also to 'show warnings' if repair was retried.
      - Increased size of internal sort-buffer-readers from 16K to 128K
      - Changed printing of 'number of records' to use %ll instead of casting to long
      - Changed buffer sizes for myisam and aria to use MY_ALIGN_DOWN() to get same number of bytes allocated on different machines.
      
      
      
      
      
      include/my_global.h:
        Added MY_ALIGN_DOWN() to get previous alignment (for big memory areas)
      include/myisam.h:
        Increased size of types to be able to handle more records
      include/myisamchk.h:
        Increased size of types to be able to handle more records
        Added T_FORCE_SORT_MEMORY to force repair to work even if sort_buffer would not be big enough
      mysql-test/r/myisam.result:
        Updated result
      mysql-test/r/mysqld--help.result:
        Updated result
      mysql-test/r/repair.result:
        Updated result
      mysql-test/suite/maria/maria.result:
        Added test cases for some fixed bugs in MyISAM to verify that Aria doesn't have them too.
      mysql-test/suite/maria/maria.test:
        Added test cases for some fixed bugs in MyISAM to verify that Aria doesn't have them too.
      mysql-test/suite/maria/maria3.result:
        Updated result after sort buffer size increase
      mysql-test/suite/sys_vars/r/aria_sort_buffer_size_basic.result:
        Updated result after sort buffer size increase
      mysql-test/suite/sys_vars/r/myisam_sort_buffer_size_basic.result:
        Updated result after sort buffer size increase
      mysql-test/suite/sys_vars/t/myisam_sort_buffer_size_basic.test:
        Updated result after sort buffer size increase
      mysql-test/t/myisam.test:
        Fixed error messages to not print system specific data
      mysql-test/t/repair.test:
        Fixed error messages to not print system specific data
      storage/maria/ha_maria.cc:
        Print also to 'show warnings' if repair was retried
        Changed default size of sort_buffer_size from 128M to 256M (same as in mysqld)
      storage/maria/ma_check.c:
        Renamed USE_BUFFER_INIT -> PAGE_BUFFER_INIT
      storage/maria/ma_sort.c:
        Increased size of internal sort-buffer-readers from 16K to 128K
        Increased size of types to be able to handle more records
        Added support for T_FORCE_SORT_MEMORY
        Don't allocate too many extra BUFFPEK at a time (they are probably not needed)
        Improved error message for "sort_buffer_size is too small"
        Changed printing of 'number of records' to use %ll instead of casting to long
        Fixed bug where maria_update_key_parts() was called too early.
        Fixed bug in detecting result from read_to_buffer().
        Added 'out of memory' checking when calling 'alloc_dynamic()'.
      storage/maria/maria_chk.c:
        Added --create-missing-keys
        If one specifies --force twice then we will try to finnish the repair even if sort_buffer would be too small.
        check_param.sort_buffer_length varialble was used with wrong type.
      storage/maria/maria_def.h:
        Increased size of types to be able to handle more records
        Use MY_ALIGN_DOWN() to get same number of bytes allocated on different machines
        Renamed USE_BUFFER_INIT -> PAGE_BUFFER_INIT
      storage/maria/maria_ftdump.c:
        Renamed USE_BUFFER_INIT -> PAGE_BUFFER_INIT
      storage/maria/maria_read_log.c:
        Use PAGE_BUFFER_INIT for page cache
      storage/myisam/ha_myisam.cc:
        Changed default size of myisam_sort_buffer_size from 8M to 128M
      storage/myisam/mi_check.c:
        Renamed USE_BUFFER_INIT -> KEY_BUFFER_INIT
      storage/myisam/myisam_ftdump.c:
        Renamed USE_BUFFER_INIT -> KEY_BUFFER_INIT
      storage/myisam/myisamchk.c:
        Added --create-missing-keys
        If one specifies --force twice then we will try to finnish the repair even if sort_buffer would be too small.
        check_param.sort_buffer_length varialble was used with wrong type.
        Renamed USE_BUFFER_INIT -> KEY_BUFFER_INIT
      storage/myisam/myisamdef.h:
        Increased SORT_BUFFER_INIT to 64M
        (speeds up repair a lot and most machines have nowadays a lot of memory)
        Use MY_ALIGN_DOWN() to get same number of bytes allocated on different machines
        Renamed USE_BUFFER_INIT -> KEY_BUFFER_INIT
      storage/myisam/sort.c:
        Increased size of internal sort-buffer-readers from 16K to 128K
        Increased size of types to be able to handle more records
        Added support for T_FORCE_SORT_MEMORY
        Don't allocate too many extra BUFFPEK at a time (they are probably not needed)
        Improved error message for "sort_buffer_size is too small"
        Changed printing of 'number of records' to use %ll instead of casting to long
        Fixed bug in detecting result from read_to_buffer().
        Added 'out of memory' checking when calling 'alloc_dynamic()'.
      53d44ad1
  11. 06 May, 2013 1 commit
  12. 05 May, 2013 1 commit
    • Michael Widenius's avatar
      Fixed errors and compiler warnings found by buildbot · 5333dafa
      Michael Widenius authored
      Solaris fixes:
      - Fixed that wait_timeout_func and wait_timeout tests works on solaris
      - We have to compile without NO_ALARM on Solaris as Solaris doesn't support timeouts on sockets with setsockopt(.. SO_RCVTIMEO).
      - Fixed that compile-solaris-amd64-debug works (before that we got a wrong ELF class: ELFCLASS64 on linkage)
      - Added missing sync_with_master
      Other bug fixes:
      - Free memory for rpl_global_gtid_binlog_state before exit() to avoid 'accessing uninitalized mutex' error.
      
      
      
      BUILD/FINISH.sh:
        Fixed issues on Solaris with ksh
      BUILD/compile-solaris-amd64-debug:
        Added missing -m64 flag
      configure.cmake:
        We have to compile without NO_ALARM on Solaris as Solaris doesn't support timeouts on sockets with setsockopt(.. SO_RCVTIMEO)
      mysql-test/suite/rpl/t/rpl_gtid_mdev4473.test:
        - Added missing sync_with_master (fix by knielsen)
      sql-common/client.c:
        Added () to get rid of compiler warning
      sql/item_strfunc.cc:
        Fixed compiler warning
      sql/log.cc:
        Free memory for static variable rpl_global_gtid_binlog_state before exit()
        - If we are compiling with safemalloc, we would try to call sf_free() for some members after sf_terminate() was called, which would result of trying to access the uninitalized mutex 'sf_mutex'
      sql/multi_range_read.cc:
        Fixed compiler warnings of converting double to ulong.
      sql/opt_range.cc:
        Fixed compiler warnings of converting double to ulong or uint
        - Better to have all variables that can be number of rows as 'ha_rows'
      sql/rpl_gtid.cc:
        Added rpl_binlog_state::free() to be able to free memory for static objects before exit()
      sql/rpl_gtid.h:
        Added rpl_binlog_state::free() to be able to free memory for static objects before exit()
      sql/set_var.cc:
        Fixed compiler warning
      sql/sql_join_cache.cc:
        Fixed compiler warnings of converting double to uint
      sql/sql_show.cc:
        Added cast to get rid of compiler warning
      sql/sql_statistics.cc:
        Remove code that didn't do anything.
        (store_record() with record[0] is a no-op)
      storage/xtradb/os/os0file.c:
        Added  __attribute__ ((unused))
      support-files/compiler_warnings.supp:
        Ignore warnings from atomic_add_64_nv
        (was not able to fix this with a cast as the macro is a bit different between systems)
      vio/viosocket.c:
        Added more DBUG_PRINT
      5333dafa
  13. 03 May, 2013 1 commit
    • unknown's avatar
      MDEV-4473: mysql_binlog_send() starts sending events from wrong GTID position... · 5aa0d185
      unknown authored
      MDEV-4473: mysql_binlog_send() starts sending events from wrong GTID position in some master failover scenarios
      
      Suppose binlog file X has in its Gtid_list_event: 0-1-3,0-2-5, and suppose the
      slave requests to start replicating after 0-1-3.
      
      In this case the bug was that master would start sending events from the start
      of X. This is wrong, because 0-2-4 and 0-2-5 are contained in X-1, and are
      needed by the slave. So these events were lost.
      
      On the other hand, if the slave requested 0-2-5, then it _is_ correct to start
      sending from the beginning of binlog file X, because 0-2-5 is the last GTID
      logged in earlier binlogs. The difference is that 0-2-5 is the last of the
      GTIDs in the Gtid_list_event. The problem was that the code did not check that
      the matched GTID was the last one in the list.
      
      Fixed by checking if the gtid requested by slave that matches a gtid in the
      Gtid_list_event is the last event for that domain in the list. If not, go back
      to a prior binlog to ensure all needed events are sent to slave.
      
      mysql-test/include/show_events.inc:
        Backport --let $binlog_file=LAST, used by MDEV-4473 test case.
      5aa0d185
  14. 02 May, 2013 2 commits
    • Michael Widenius's avatar
      Instead of writing "Errcode" to the log for Slave errors, use "Internal MariaDB error code" · eb75edfb
      Michael Widenius authored
      This makes it clear that the error code has nothing to do with errno.
      
      
      mysql-test/include/mtr_warnings.sql:
        Fixed suppression for new slave error messages
      mysql-test/lib/My/Test.pm:
        Use 'send' instead of 'print' to avoid errors about "wrong class ... back attempt"
      mysql-test/lib/v1/mtr_report.pl:
        Fixed suppression for new slave error messages
      mysql-test/mysql-test-run.pl:
        Fixed suppression for new slave error messages
        Removed warning from perl 5.16.2 about arrays
      mysql-test/r/flush_read_lock.result:
        Fixed suppression for new slave error messages
      sql/rpl_reporting.cc:
        Instead of writing "Errcode" to the log for Slave errors, use "Internal MariaDB error code"
      eb75edfb
    • Michael Widenius's avatar
      Fixed: MDEV-4352; LOAD DATA was not multi-source safe · 8cdb118a
      Michael Widenius authored
      - Calls to cleanup_load_tmpdir() could delete temporary files for another master connection
      - Concurrent LOAD DATA commands from two master connections could use the same file name
      
      Other bug fixes:
      - Enlarge buffer for connection names with 'special characters' one can't store in filenames
      
      Optimization:
      - Don't do 'lower case' of connection names. We can use cmp_connection_name, where we already have the connection name in lower case.
      
      
      mysql-test/suite/multi_source/load_data.result:
        Test case for MDEV-4352
      mysql-test/suite/multi_source/load_data.test:
        Test case for MDEV-4352
      sql/log_event.cc:
        Fixed: MDEV-4352
        - Calls to cleanup_load_tmpdir() could delete temporary files for another master connection
        - Concurrent LOAD DATA commands from two master connections could use the same file name
        
        The fix was to add the connection name (if one exists) to all slave temporary files used by LOAD DATA
      sql/rpl_mi.cc:
        Enlarge buffer for connection names with 'special characters' one can't store in filenames
        Use mi->cmp_connection_name for connection file names.
      sql/rpl_rli.cc:
        Use mi->cmp_connection_name for connection file names.
      sql/slave.cc:
        Removed not needed empty line
      sql/sql_const.h:
        Added MAX_FILENAME_MBWIDTH to be able to calculate buffer length for connection_names stored in file names
      sql/sql_repl.cc:
        Use mi->cmp_connection_name for connection file names.
      8cdb118a
  15. 29 Apr, 2013 1 commit
  16. 25 Apr, 2013 1 commit
    • unknown's avatar
      Fix more failures in buildbot. · ced65e21
      unknown authored
      Server shutdown timeout of 10 seconds in test cases is too little for heavily
      loaded test servers.
      
      Fix innodb_bug12902967 to not fail with wrong error log output if we have
      warnings about too few AIO handles for InnoDB.
      
      Fix typo which could lead to unnecessarily replacing GTID event with dummy
      event.
      ced65e21
  17. 24 Apr, 2013 1 commit
    • unknown's avatar
      Add missing check for thd->killed in mysql_binlog_send(). · 6b97512b
      unknown authored
      The slave dump thread running on the master only checked thd->killed whenever
      it reached the end of a binlog file, not between events. This could
      unnecessarily delay server shutdown.
      
      This was found by code inspection while tracking down some occasional "forcing
      close of thread..." errors in Buildbot. Hopefully this will fix the failures,
      but the fix is correct in any case.
      
      Also increase the wait during server shutdown, 2 seconds is a bit tight in
      case of heavy I/O stall, and it seems better to delay shutdown a bit than
      force-kill threads unnecessarily.
      
      Also fix some races in test cases that restart the mysqld server. The .expect
      file should be changed with --append_file, --remove_file + --write_file
      creates a short window where mysqld can error out due to .expect file missing.
      6b97512b
  18. 22 Apr, 2013 4 commits
  19. 21 Apr, 2013 3 commits
  20. 20 Apr, 2013 3 commits
  21. 19 Apr, 2013 4 commits
  22. 18 Apr, 2013 4 commits
  23. 17 Apr, 2013 1 commit