MariaDB Development
  1. MariaDB Development
  2. MDEV-5029

Crash in MariaDB 5.5.33 with .frm from older MariaDB release

    Details

    • Type: Bug Bug
    • Status: Closed
    • Priority: Critical Critical
    • Resolution: Fixed
    • Affects Version/s: 5.5.33
    • Fix Version/s: 5.5.33a
    • Labels:
      None
    • Global Rank:
      3623

      Description

      An frm file found "in the wild" had 0x2b as its 4th byte. This "works" fine in MariaDB <5.5.33 but causes crashes in MariaDB 5.5.33. Changing this byte in the .frm file to 0x2a allows the table to be used in MariaDB 5.5.33, but the first use gives this warning:

      | Note  | 1194 | Zerofilling moved table ./test/kb3_domains |
      

      It is not clear how this .frm file was originally created, but I am assured that it was generated by MariaDB.

      Stacktrace:

      ??:0(my_print_stacktrace)[0xa8962e]
      ??:0(handle_fatal_signal)[0x6d002b]
      ??:0(??)[0x7fa17becb500]
      ??:0(TABLE_SHARE::destroy())[0x61da55]
      ??:0(my_hash_delete)[0xa6ea34]
      ??:0(_Z15get_table_shareP3THDP10TABLE_LISTPcjjPij.clone.7)[0x541ae7]
      ??:0(open_table(THD*, TABLE_LIST*, st_mem_root*, Open_table_context*))[0x54ac1d]
      ??:0(open_tables(THD*, TABLE_LIST**, unsigned int*, unsigned int, Prelocking_strategy*))[0x54bf01]
      ??:0(open_and_lock_tables(THD*, TABLE_LIST*, bool, unsigned int, Prelocking_strategy*))[0x54cc54]
      ??:0(execute_sqlcom_select(THD*, TABLE_LIST*))[0x585915]
      ??:0(mysql_execute_command(THD*))[0x58c0f6]
      ??:0(mysql_parse(THD*, char*, unsigned int, Parser_state*))[0x58ee19]
      ??:0(dispatch_command(enum_server_command, THD*, char*, unsigned int))[0x590270]
      ??:0(do_handle_one_connection(THD*))[0x649c1f]
      ??:0(handle_one_connection)[0x649cbc]
      ??:0(??)[0x7fa17bec3851]
      ??:0(??)[0x7fa17a83790d]
      

        Issue Links

          Activity

          Hide
          Kolbe Kegel added a comment -

          Sorry, I should specify that this is an Aria table and is supposed to be an Aria table. It was, about a year ago, converted to Aria from MyISAM. I haven't done too much investigation into what version of MariaDB might have been used at the time or anything like that...

          ALTER TABLE ... ENGINE=ARIA; seems to fix this and put the correct engine ID (0x2a) into the .frm file.

          Show
          Kolbe Kegel added a comment - Sorry, I should specify that this is an Aria table and is supposed to be an Aria table. It was, about a year ago, converted to Aria from MyISAM. I haven't done too much investigation into what version of MariaDB might have been used at the time or anything like that... ALTER TABLE ... ENGINE=ARIA; seems to fix this and put the correct engine ID (0x2a) into the .frm file.
          Hide
          Jean Weisbuch added a comment - - edited

          Modifying the 4th byte to 0x2b on a MyISAM table does exactly the same : works as if nothing were wrong on MariaDB 5.5.32 but crashes on 5.5.33.

          Here is what i got on 5.5.33 if i try to do a SELECT on the table or a SHOW TABLE STATUS (i can do both without any error on the logs on 5.5.32) :
          mysqld: Version: '5.5.33-MariaDB-1~squeeze' socket: '/var/run/mysqld/mysqld.sock' port: 3306 mariadb.org binary distribution
          mysqld: 130918 3:10:45 [ERROR] mysqld: Incorrect information in file: './test123/zz_comments.frm'
          mysqld: 130918 3:10:45 [ERROR] mysqld got signal 11 ;
          mysqld: This could be because you hit a bug. It is also possible that this binary
          mysqld: or one of the libraries it was linked against is corrupt, improperly built,
          mysqld: or misconfigured. This error can also be caused by malfunctioning hardware.
          mysqld:
          mysqld: To report this bug, see http://kb.askmonty.org/en/reporting-bugs
          mysqld:
          mysqld: We will try our best to scrape up some info that will hopefully help
          mysqld: diagnose the problem, but since we have already crashed,
          mysqld: something is definitely wrong and this may fail.
          mysqld:
          mysqld: Server version: 5.5.33-MariaDB-1~squeeze
          mysqld: key_buffer_size=2147483648
          mysqld: read_buffer_size=2097152
          mysqld: max_used_connections=1
          mysqld: max_threads=252
          mysqld: thread_count=1
          mysqld: It is possible that mysqld could use up to
          mysqld: key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 3133826 K bytes of memory
          mysqld: Hope that's ok; if not, decrease some variables in the equation.
          mysqld:
          mysqld: Thread pointer: 0x0x7fd4e71ee680
          mysqld: Attempting backtrace. You can use the following information to find out
          mysqld: where mysqld died. If you see no messages after this, something went
          mysqld: terribly wrong...
          mysqld: stack_bottom = 0x7fd1c29cfe58 thread_stack 0x48000
          mysqld: /usr/sbin/mysqld(my_print_stacktrace+0x2e)[0x7fd4dbbe280e]
          mysqld: /usr/sbin/mysqld(handle_fatal_signal+0x4ac)[0x7fd4db7e1d7c]
          mysqld: /lib/libpthread.so.0(+0xeff0)[0x7fd4daedeff0]
          mysqld: /usr/sbin/mysqld(_ZN11TABLE_SHARE7destroyEv+0x9d)[0x7fd4db72368d]
          mysqld: /usr/sbin/mysqld(my_hash_delete+0x314)[0x7fd4dbbc74d4]
          mysqld: /usr/sbin/mysqld(+0x32d6c8)[0x7fd4db6396c8]
          mysqld: /usr/sbin/mysqld(_Z10open_tableP3THDP10TABLE_LISTP11st_mem_rootP18Open_table_context+0x6d7)[0x7fd4db642ee7]
          mysqld: /usr/sbin/mysqld(_Z11open_tablesP3THDPP10TABLE_LISTPjjP19Prelocking_strategy+0x59d)[0x7fd4db6443bd]
          mysqld: /usr/sbin/mysqld(_Z30open_normal_and_derived_tablesP3THDP10TABLE_LISTjj+0x4f)[0x7fd4db64501f]
          mysqld: /usr/sbin/mysqld(_Z18mysqld_list_fieldsP3THDP10TABLE_LISTPKc+0x29)[0x7fd4db6e5119]
          mysqld: /usr/sbin/mysqld(_Z16dispatch_command19enum_server_commandP3THDPcj+0x19b9)[0x7fd4db68eab9]
          mysqld: /usr/sbin/mysqld(_Z24do_handle_one_connectionP3THD+0x419)[0x7fd4db753809]
          mysqld: /usr/sbin/mysqld(handle_one_connection+0x51)[0x7fd4db7538c1]
          mysqld: /lib/libpthread.so.0(+0x68ca)[0x7fd4daed68ca]
          mysqld: /lib/libc.so.6(clone+0x6d)[0x7fd4d9855b6d]
          mysqld:
          mysqld: Trying to get some variables.
          mysqld: Some pointers may be invalid and cause the dump to abort.
          mysqld: Query (0x7fd4e7239a18):
          mysqld:
          mysqld: Connection ID (thread ID): 1
          mysqld: Status: NOT_KILLED
          mysqld:
          mysqld: Optimizer switch: index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,index_merge_sort_intersection=off,engine_condition_pushdown=off,index_condition_pushdown=on,derived_merge=on,derived_with_keys=on,firstmatch=on,loosescan=on,materialization=on,in_to_exists=on,semijoin=on,partial_match_rowid_merge=on,partial_match_table_scan=on,subquery_cache=on,mrr=off,mrr_cost_based=off,mrr_sort_keys=off,outer_join_with_cache=on,semijoin_with_cache=on,join_cache_incremental=on,join_cache_hashed=on,join_cache_bka=on,optimize_join_buffer_size=off,table_elimination=on,extended_keys=off

          ps: i used the same table structure as the original crashing .frm for my test : http://paste.karbowiak.dk/84e86c5921f0eb3ede3d427295e4dd421bfbb0a2/

          Show
          Jean Weisbuch added a comment - - edited Modifying the 4th byte to 0x2b on a MyISAM table does exactly the same : works as if nothing were wrong on MariaDB 5.5.32 but crashes on 5.5.33. Here is what i got on 5.5.33 if i try to do a SELECT on the table or a SHOW TABLE STATUS (i can do both without any error on the logs on 5.5.32) : mysqld: Version: '5.5.33-MariaDB-1~squeeze' socket: '/var/run/mysqld/mysqld.sock' port: 3306 mariadb.org binary distribution mysqld: 130918 3:10:45 [ERROR] mysqld: Incorrect information in file: './test123/zz_comments.frm' mysqld: 130918 3:10:45 [ERROR] mysqld got signal 11 ; mysqld: This could be because you hit a bug. It is also possible that this binary mysqld: or one of the libraries it was linked against is corrupt, improperly built, mysqld: or misconfigured. This error can also be caused by malfunctioning hardware. mysqld: mysqld: To report this bug, see http://kb.askmonty.org/en/reporting-bugs mysqld: mysqld: We will try our best to scrape up some info that will hopefully help mysqld: diagnose the problem, but since we have already crashed, mysqld: something is definitely wrong and this may fail. mysqld: mysqld: Server version: 5.5.33-MariaDB-1~squeeze mysqld: key_buffer_size=2147483648 mysqld: read_buffer_size=2097152 mysqld: max_used_connections=1 mysqld: max_threads=252 mysqld: thread_count=1 mysqld: It is possible that mysqld could use up to mysqld: key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 3133826 K bytes of memory mysqld: Hope that's ok; if not, decrease some variables in the equation. mysqld: mysqld: Thread pointer: 0x0x7fd4e71ee680 mysqld: Attempting backtrace. You can use the following information to find out mysqld: where mysqld died. If you see no messages after this, something went mysqld: terribly wrong... mysqld: stack_bottom = 0x7fd1c29cfe58 thread_stack 0x48000 mysqld: /usr/sbin/mysqld(my_print_stacktrace+0x2e) [0x7fd4dbbe280e] mysqld: /usr/sbin/mysqld(handle_fatal_signal+0x4ac) [0x7fd4db7e1d7c] mysqld: /lib/libpthread.so.0(+0xeff0) [0x7fd4daedeff0] mysqld: /usr/sbin/mysqld(_ZN11TABLE_SHARE7destroyEv+0x9d) [0x7fd4db72368d] mysqld: /usr/sbin/mysqld(my_hash_delete+0x314) [0x7fd4dbbc74d4] mysqld: /usr/sbin/mysqld(+0x32d6c8) [0x7fd4db6396c8] mysqld: /usr/sbin/mysqld(_Z10open_tableP3THDP10TABLE_LISTP11st_mem_rootP18Open_table_context+0x6d7) [0x7fd4db642ee7] mysqld: /usr/sbin/mysqld(_Z11open_tablesP3THDPP10TABLE_LISTPjjP19Prelocking_strategy+0x59d) [0x7fd4db6443bd] mysqld: /usr/sbin/mysqld(_Z30open_normal_and_derived_tablesP3THDP10TABLE_LISTjj+0x4f) [0x7fd4db64501f] mysqld: /usr/sbin/mysqld(_Z18mysqld_list_fieldsP3THDP10TABLE_LISTPKc+0x29) [0x7fd4db6e5119] mysqld: /usr/sbin/mysqld(_Z16dispatch_command19enum_server_commandP3THDPcj+0x19b9) [0x7fd4db68eab9] mysqld: /usr/sbin/mysqld(_Z24do_handle_one_connectionP3THD+0x419) [0x7fd4db753809] mysqld: /usr/sbin/mysqld(handle_one_connection+0x51) [0x7fd4db7538c1] mysqld: /lib/libpthread.so.0(+0x68ca) [0x7fd4daed68ca] mysqld: /lib/libc.so.6(clone+0x6d) [0x7fd4d9855b6d] mysqld: mysqld: Trying to get some variables. mysqld: Some pointers may be invalid and cause the dump to abort. mysqld: Query (0x7fd4e7239a18): mysqld: mysqld: Connection ID (thread ID): 1 mysqld: Status: NOT_KILLED mysqld: mysqld: Optimizer switch: index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,index_merge_sort_intersection=off,engine_condition_pushdown=off,index_condition_pushdown=on,derived_merge=on,derived_with_keys=on,firstmatch=on,loosescan=on,materialization=on,in_to_exists=on,semijoin=on,partial_match_rowid_merge=on,partial_match_table_scan=on,subquery_cache=on,mrr=off,mrr_cost_based=off,mrr_sort_keys=off,outer_join_with_cache=on,semijoin_with_cache=on,join_cache_incremental=on,join_cache_hashed=on,join_cache_bka=on,optimize_join_buffer_size=off,table_elimination=on,extended_keys=off ps: i used the same table structure as the original crashing .frm for my test : http://paste.karbowiak.dk/84e86c5921f0eb3ede3d427295e4dd421bfbb0a2/
          Hide
          Jean Weisbuch added a comment -

          If you want to check that you don't have one of these frm file with a bad value (don't forget to adjust to your datadir if its not /var/lib/mysql/) :
          for frm in /var/lib/mysql//.frm; do [[ "$(od -An -j2 -N2 -x "$frm")" = " 2b0a" ]] && echo "$frm"; done

          It will return a list of frm files that has this DB_TYPE then if the tables are Aria ones, you will hit the bug.

          Show
          Jean Weisbuch added a comment - If you want to check that you don't have one of these frm file with a bad value (don't forget to adjust to your datadir if its not /var/lib/mysql/) : for frm in /var/lib/mysql/ / .frm; do [[ "$(od -An -j2 -N2 -x "$frm")" = " 2b0a" ]] && echo "$frm"; done It will return a list of frm files that has this DB_TYPE then if the tables are Aria ones, you will hit the bug.
          Hide
          Sergei Golubchik added a comment -

          For the record:

          • before Aria engine type code was frozen in MDEV-4466 it was dynamic and when OQGraph of Sphinx is compiled statically (they also have dynamic engine codes), Aria could've had not 0x2a, but 0x2b or even 0x2c
          • after MDEV-4466 Aria always has 0x2a
          • after MDEV-4507 TokuDB has 0x2b (it cannot have dynamic code because of MDEV-4466)
          • now the server refuses to load frms that have engine code 0x2b (TokuDB) and the engine name "Aria"
          Show
          Sergei Golubchik added a comment - For the record: before Aria engine type code was frozen in MDEV-4466 it was dynamic and when OQGraph of Sphinx is compiled statically (they also have dynamic engine codes), Aria could've had not 0x2a, but 0x2b or even 0x2c after MDEV-4466 Aria always has 0x2a after MDEV-4507 TokuDB has 0x2b (it cannot have dynamic code because of MDEV-4466 ) now the server refuses to load frms that have engine code 0x2b (TokuDB) and the engine name "Aria"
          Show
          Daniel Bartholomew added a comment - http://bazaar.launchpad.net/~maria-captains/maria/5.5/revision/4041

            People

            • Assignee:
              Sergei Golubchik
              Reporter:
              Kolbe Kegel
            • Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: