1. 11 Jan, 2011 4 commits
    • Michael Widenius's avatar
      Allow one to decrease page-buffer-size down to 1M (from 128M) for maria_read_log · 6e12d3f6
      Michael Widenius authored
      Don't allow too low value of pagecache_buffer_size for mysqld
      
      storage/maria/ha_maria.cc:
        Don't allow too low value of pagecache_buffer_size (as it will stop mysqld from starting)
      storage/maria/maria_read_log.c:
        Allow one to decrease page-buffer-size down to 1M (from 128M)
      6e12d3f6
    • Michael Widenius's avatar
      Allow smaller size for page-buffer-size · 20e8d10a
      Michael Widenius authored
      storage/maria/maria_read_log.c:
        Allow smaller size for page-buffer-size (down to 1M)
        Sorted options for --help
      20e8d10a
    • Michael Widenius's avatar
      Fixed wrong test in maria_rsame() that caused ma_test_all to fail. · 4ef446f2
      Michael Widenius authored
      storage/maria/ma_rsame.c:
        Fixed wrong test of index usage
      storage/maria/ma_search.c:
        Fixed test to avoid compiler warnings.
        Safety fix to ensure that my_error is properly set in case of errors.
      4ef446f2
    • Michael Widenius's avatar
      Safety fix for Aria: · fe37e5dc
      Michael Widenius authored
      - Set lastinx= ~0 when last_key.keyinfo is set.
      
      
      storage/maria/ma_check.c:
        Set lastinx= ~0 when last_key.keyinfo is set
      storage/maria/ma_ft_boolean_search.c:
        Set lastinx= ~0 when last_key.keyinfo is set
      storage/maria/ma_rt_index.c:
        Remove setting of info->lastkey.keyinfo, as this should already be set by caller
      storage/maria/ma_search.c:
        Added ASSERT to ensure that info->last_key.keyinfo is properly set
      storage/maria/ma_unique.c:
         Set lastinx= ~0 when last_key.keyinfo is set
      fe37e5dc
  2. 10 Jan, 2011 1 commit
    • Michael Widenius's avatar
      - Fixed that Aria works with HANDLER commands · 15074383
      Michael Widenius authored
      - Added test case for Aria
      - Tested HANDLER with HEAP (changes to HEAP code will be pushed in 5.3)
      - Moved all HANDLER test to suite/handler.
      
      
      mysql-test/Makefile.am:
        Added suite/handler
      mysql-test/mysql-test-run.pl:
        Added suite/handler
      mysql-test/r/lock_multi.result:
        Remove test that is already in handler test suite
      mysql-test/suite/handler/aria.result:
        Test for HANDLER with Aria storage engine
      mysql-test/suite/handler/aria.test:
        Test for HANDLER with Aria storage engine
      mysql-test/suite/handler/handler.inc:
        Extended the general handler test
        Moved interface testing to 'interface.test'
      mysql-test/suite/handler/init.inc:
        Common init for handler tests.
      mysql-test/suite/handler/innodb.result:
        New results
      mysql-test/suite/handler/innodb.test:
        Update to use new include files
      mysql-test/suite/handler/interface.result:
        Test of HANDLER interface (not storage engine dependent parts)
      mysql-test/suite/handler/interface.test:
        Test of HANDLER interface (not storage engine dependent parts)
      mysql-test/suite/handler/myisam.result:
        New results
      mysql-test/suite/handler/myisam.test:
        Update to use new include files
      mysql-test/t/lock_multi.test:
        Remove test that is already in handler test suite
      mysys/tree.c:
        Added missing handling of read previous (showed up in HEAP testing)
      sql/handler.cc:
        Don't marka 'HA_ERR_RECORD_CHANGED' as fatal (can be used with HANDLER READ, especially with MEMORY ENGINE)
      sql/handler.h:
        Added prototype for can_continue_handler_scan()
      sql/sql_handler.cc:
        Re-initialize search if we switch from key to table search.
        Check if handler can continue searching between calls (via can_continue_handler_scan())
        Don't write common not fatal errors to log
      storage/maria/ma_extra.c:
        Don't set index 0 as default. This forces call to ma_check_index() to set up index variables.
      storage/maria/ma_ft_boolean_search.c:
        Ensure that info->last_key.keyinfo is set
      storage/maria/ma_open.c:
        Don't set index 0 as default. This forces call to ma_check_index() to set up index variables.
      storage/maria/ma_rkey.c:
        Trivial optimization
      storage/maria/ma_rnext.c:
        Added missing code from mi_rnext.c to ensure that handler next/prev works.
      storage/maria/ma_rsame.c:
        Simple optimizations
      storage/maria/ma_search.c:
        Initialize info->last_key once and for all when we change keys.
      storage/maria/ma_unique.c:
        Ensure that info->last_key.keyinfo is up to date.
      15074383
  3. 07 Jan, 2011 1 commit
    • Michael Widenius's avatar
      Don't do DBUG_ASSERT for checksum errors when using REPAIR · 91ecf748
      Michael Widenius authored
      mysql_convert_table_format ignored --engine option.
      Fix that zerofill() doesn't write out wrong data to client if run with auto repair.
      Ensure that pagecache is properly flushed, even in case of errors.
      Handle checksum errors in BLOCK_RECORD format.
      
      
      scripts/mysql_convert_table_format.sh:
        Fixed that --engine option works
      storage/maria/ha_maria.cc:
        Fix that zerofill() doesn't write out wrong data to client if run with auto repair.
      storage/maria/ma_check.c:
        Set in_check_table when scanning table to not get DBUG_ASSERT for checksum error.
        Ensure that pagecache is properly flushed, even in case of errors.
        Handle checksum errors in BLOCK_RECORD format.
      storage/maria/ma_sort.c:
        Set in_check_table when scanning table to not get DBUG_ASSERT for checksum error.
      91ecf748
  4. 06 Jan, 2011 1 commit
    • Michael Widenius's avatar
      Bug fixing in Aria: · 896fb79e
      Michael Widenius authored
      - Fixed some bugs in recovery of blobs
      - Don't ASSERT() on checksum errors when running check table
      - Added to maria_read_log option --tables-to-redo=list-of-tables to only recover some tables (good for debugging)
      
      
      storage/maria/ma_blockrec.c:
        Don't ASSERT() on checksum errors when running check table
        Fixed bug in recovery of blog page that was not in dirty pages list
      storage/maria/ma_check.c:
        Don't ASSERT() on checksum errors when running check table
      storage/maria/ma_recovery.c:
        Handling of --tables-to-redo
      storage/maria/ma_recovery.h:
        Handling of --tables-to-redo
      storage/maria/ma_recovery_util.c:
        Give better warning if table was not in dirty pages list
      storage/maria/maria_def.h:
        Added in_check_table
      storage/maria/maria_read_log.c:
        Added --tables-to-redo=list-of-tables to only recover some tables (good for debugging)
        Cleaned up message when wrong arguments
      896fb79e
  5. 05 Jan, 2011 3 commits
  6. 04 Jan, 2011 2 commits
    • Michael Widenius's avatar
      Fixed recovery problem in Aria where bitmap had wrong information after recovery. · 308e11b8
      Michael Widenius authored
      LP#619731: Aria recovery corruption "Page 1: Row: 1 has an extent with wrong information in bitmap
      
      storage/maria/ma_bitmap.c:
        Don't send broadcast if no one is waiting for it
      storage/maria/ma_blockrec.c:
        Don't update bitmap if the page is not in the dirty_page list (or LSN is after checkpoint start)
        Fixes the case where we have in the log redo_free_block followed by another redo entry for the same page which is ignored.
        Also fixed that _ma_apply_redo_insert_row_blobs() doesn't update the bitmap in similar circumstances.
      storage/maria/ma_blockrec.h:
        Updated prototype
      storage/maria/ma_check.c:
        Added printing of bitmap information if used with maria_chk -vvv (for debugging)
      storage/maria/ma_recovery.c:
        Updated call parameters to _ma_apply_redo_free_blocks().
      308e11b8
    • Michael Widenius's avatar
      Updated README for MTR test suite · 7c521183
      Michael Widenius authored
      7c521183
  7. 03 Jan, 2011 1 commit
    • unknown's avatar
      Speed up `mtr --parallel=<lots>` by scheduling some slow tests earlier. · bf14edc2
      unknown authored
      The patch also fixes a race in rpl_stop_slave.test.
      
      On machines with lots of CPU and memory, something like `mtr --parallel=10`
      can speed up the test suite enormously. However, we have a few test cases
      that run for long (several minutes), and if we are unlucky and happen to
      schedule those towards the end of the test suite, we end up with most
      workers idle while waiting for the last slow test to end, significantly
      delaying the finish of the entire suite.
      
      Improve this by marking the offending tests as taking "long", and trying
      to schedule those tests early. This reduces the time towards the end of
      the test suite run where some workers are waiting with nothing to do for
      the remaining workers each to finish their last test.
      
      Also, the rpl_stop_slave test had a race which could cause it to take
      a 300 seconds debug_sync timeout; this is fixed.
      
      Testing on a 4-core 8GB machine, this patch speeds up the test suite with
      around 30% for --parallel=10 (debug build), allowing to run the entire
      suite in 5 minutes.
      bf14edc2
  8. 24 Dec, 2010 1 commit
    • Igor Babaev's avatar
      Fixed LP bug #639935 (bug #58727). · b6242087
      Igor Babaev authored
      When the optimizer creates items out of other items it does
      not have to call the fix_fields method. Usually in these
      cases it calls quick_fix_field() that just marks the 
      created item as fixed. If the created item is an Item_func
      object then calling quick_fix_field() works fine if the
      arguments of the created functional item are already fixed.
      Otherwise some unfixed nodes remain in the item tree and
      it triggers an assertion failure whenever the item is 
      evaluated.
      
      Fixed the problem by making the method quick_fix_field
      virtual and providing an implementation for the class
      Item_func objects that recursively calls the method
      for unfixed arguments of any functional item.
      b6242087
  9. 21 Dec, 2010 2 commits
  10. 13 Dec, 2010 1 commit
  11. 10 Dec, 2010 2 commits
    • Michael Widenius's avatar
      Fix for Lbug:#686010 maria.optimize corrupts stack around alloca() call · 4bb42c4e
      Michael Widenius authored
      
      storage/maria/ma_check.c:
        Allocate also memory for nodflag and transid's
      storage/maria/ma_write.c:
        Allow nodflag and transid as part of key.
        (This has nothing to do with the bug report, but it's a safer check).
      storage/maria/maria_def.h:
        Define MARIA_MAX_POINTER_LENGTH (length of pointer to node)
        Added node pointer length to MARIA_INDEX_OVERHEAD_SIZE, as this is part of the key.
        (Safety fix)
      4bb42c4e
    • Michael Widenius's avatar
      Better warning message if lock test fails · b0312896
      Michael Widenius authored
      Made archive.test a bit more safe
      
      
      
      mysql-test/r/archive.result:
        Added removal of files to make rerun of failed test work
      mysql-test/t/archive.test:
        Added removal of files to make rerun of failed test work
      mysys/thr_lock.c:
        Better warning message if lock test fails
      b0312896
  12. 08 Dec, 2010 1 commit
    • unknown's avatar
      MBug#687320: Fix sporadic test failures in innodb_mysql.test and... · 52920994
      unknown authored
      MBug#687320: Fix sporadic test failures in innodb_mysql.test and partition_innodb_semi_consistent.test
      
      Problem is that these tests run with --innodb-lock-wait-timeout=2 in .opt
      (and this is necessary as built-in innodb does not allow to change this
      dynamically). This cases another part of the test to occasionally time
      out an UPDATE, which subsequently caused the test case to timeout due to
      waiting for a condition (successful UPDATE) that never occurs.
      
      Fixed by re-trying the update in case of timeout.
      
      Tested by inserting a sleep() in the connection that the UPDATE is waiting
      for, and checking that the retry loops a couple of times until the other
      connection is done and COMMITs.
      52920994
  13. 07 Dec, 2010 1 commit
  14. 19 Nov, 2010 1 commit
    • unknown's avatar
      Mbug#677407 / MySQL Bug#48883: Stale data from INNODB_LOCKS table. · 40ce82b9
      unknown authored
      The logic for how to check when to update the table cache for
      INNODB_LOCKS with real data was flawed. This could result in both
      not updating the cache often enough (when the table is queried
      repeatedly with less than 100 milliseconds in-between) resulting
      in stale data; as well as updating too often (when multiple
      queries against the table start at around the same time).
      
      This caused occasional test failures in innodb_information_schema.
      
      Fix by updating the "last updated" timestamp in the right place,
      when the cache is updated, not when it is read.
      40ce82b9
  15. 06 Dec, 2010 6 commits
    • Vladislav Vaintroub's avatar
      Fix myisam_crash_before_flush_keys on Windows · a1abfb58
      Vladislav Vaintroub authored
      Problem : mtr reports test failure because it sees
      [ERROR] mysqld got exception 0x80000003 ;
      in the .err file
      
      The exception comes from DBUG_EXECUTE_IF (.. abort())
      
      Fix:  use DBUG_ABORT instead of abort() - it does not throw 
      any exceptions.
      a1abfb58
    • Vladislav Vaintroub's avatar
      Fix LP Bug#686184 - merge_debug test fails. · cd72b793
      Vladislav Vaintroub authored
      The reason for failure is that DBUG_EXECUTE_IF in mi_open()
      only worked for Unix-formatted file names, due to strstr(name, "/crashed")
      
      The fix change strstr() above to strstr(name, "crashed"), to it can work with 
      Windows file names as well.
      cd72b793
    • Vladislav Vaintroub's avatar
      merge · 1208fba6
      Vladislav Vaintroub authored
      1208fba6
    • Vladislav Vaintroub's avatar
      Bug#473914: mysql_client_test fail with in debug compilaton on windows x64 · 7a80fa96
      Vladislav Vaintroub authored
       
      Reason: inconsistent compilation, federatedx is compiled without SAFEMALLOC
      flag, while anything else is compiled with SAFEMALLOC.
      
      As a consequence, my_hash_init used inside federatedx initialization does not
      provide correct caller info parameters (file, line) , so they are initialized with 
      whatever is on stack. When info about allocated memory is output in
      COM_DEBUG command, the server crashes trying to output string starting at
      0xcccccccccccccccc.
      
      The fix is to remove SAFEMALLOC  preprocessor flags 
      from every CMakeLists.txt, except the top-level one.
      
      Also, SAFEMALLOC is not defined by default now, instead
      there is WITH_DEBUG_FULL CMake option which adds 
      -DSAFEMALLOC to C and C++ flags in debug compilation. 
      This option is off by  default, because 
      1) Debug C runtime already has heap debugging builtin with 
       overwrite and leak detection
      2)safemalloc considerably slows down the tests.
      
      
      Note also  that 
      - SAFEMALLOC is gone  in MySQL5.5
      - On Windows, heap related overflows can also be found using free pageheap utility
      (that is also part of application verifier). This is even more efficient if there are no other layers 
      on top of Windows heap allocator, e.g  it is most efficient with release version.
      7a80fa96
    • Michael Widenius's avatar
    • Michael Widenius's avatar
      Merge with 5.1-release · 7720092b
      Michael Widenius authored
      7720092b
  16. 05 Dec, 2010 1 commit
    • Michael Widenius's avatar
      Fixes to not trigger end-consistency testing of pagecache. · 420b6312
      Michael Widenius authored
      - Moved end_pagecache() to after maria_close()
      - Flush page cache before closing files (in maria_pack)
      Fixed test suite failure for PBXT
      
      mysql-test/suite/pbxt/r/select.result:
        We mask out the 'rows' column because it may differ from run to run
      mysql-test/suite/pbxt/t/select.test:
        We mask out the 'rows' column because it may differ from run to run
      storage/maria/maria_chk.c:
        Moved end_pagecache() to after maria_close() to not trigger end-consistency testing of pagecache.
      storage/maria/maria_pack.c:
        Moved end_pagecache() to after maria_close() to not trigger end-consistency testing of pagecache.
        flush page cache before closing files.
      420b6312
  17. 04 Dec, 2010 7 commits
  18. 03 Dec, 2010 1 commit
    • Michael Widenius's avatar
      Fixed compiler warnings. · 60f42a46
      Michael Widenius authored
      Fixed timing test failures.
      Fixed a failure in the Aria engines page cache and log handler (found with maria.maria-big test)
      - This could cause a core dump when deleting big blobs.
      - Added test to end_pagecache() to verify that page cache was correctly used.
        - inc_counter_for_resize_op and dec_counter_for_resize_op are called same number of times.
        - All page cache blocks was properly deallocated (empty)
      
      
      mysql-test/suite/innodb/t/innodb_bug38231.test:
        Fixed timing issue (code comment says it all)
      mysql-test/suite/innodb_plugin/t/innodb_bug38231.test:
        Fixed timing issue (code comment says it all)
      sql/debug_sync.cc:
        Fixed compiler warning
      storage/maria/ma_loghandler.c:
        Fixed bug found by maria.maria-big test:
        - Fixed race condition between update thread logging a very big blog and checkpoint thread.
      storage/maria/ma_pagecache.c:
        Added assert to ensure mutex was properly locked.
        Added test to end_pagecache() to verify that page cache was correctly used.
        - inc_counter_for_resize_op and dec_counter_for_resize_op are called same number of times.
        - All page cache blocks was properly deallocated (empty)
        In pagecache_delete_internal(), properly reset counters and pins if functions aborts.
        Added missing inc_counter_for_resize_op() to pagecache_wait_lock().
        Added missing dec_counter_for_resize_op() to pagecache_delete()
      60f42a46
  19. 02 Dec, 2010 3 commits