Seen during another run (Laurynas expects it is the same issue):
==16835== HEAP SUMMARY: ==16835== in use at exit: 152 bytes in 1 blocks ==16835== total heap usage: 55,843,989 allocs, 55,843,988 frees, 420,201,789,971 bytes allocated ==16835== ==16835== 152 bytes in 1 blocks are definitely lost in loss record 1 of 1 ==16835== at 0x4C278FE: malloc (vg_replace_malloc.c:270) ==16835== by 0x84B0D2: ut_malloc_low (ut0mem.c:102) ==16835== by 0x883C58: buf_LRU_free_block (buf0buf.ic:871) ==16835== by 0x885016: buf_LRU_search_and_free_block (buf0lru.c:940) ==16835== by 0x8856EE: buf_LRU_get_free_block (buf0lru.c:1285) ==16835== by 0x87471A: buf_page_get_gen (buf0buf.c:2681) ==16835== by 0x851534: btr_free_but_not_root (btr0btr.ic:59) ==16835== by 0x88B12E: dict_drop_index_tree (dict0crea.c:795) ==16835== by 0x81EA2E: row_upd_step (row0upd.c:2211) ==16835== by 0x913E9C: que_run_threads (que0que.c:1238) ==16835== by 0x91448F: que_eval_sql (que0que.c:1445) ==16835== by 0x80C47C: row_drop_table_for_mysql (row0mysql.c:3443) ==16835== by 0x7EE140: ha_innobase::delete_table(char const*) (ha_innodb.cc:8382) ==16835== by 0x69C400: ha_delete_table(THD*, handlerton*, char const*, char const*, char const*, bool) (handler.cc:2211) ==16835== by 0x5E1ABC: quick_rm_table(handlerton*, char const*, char const*, unsigned int) (sql_table.cc:2410) ==16835== by 0x5E42AF: mysql_alter_table(THD*, char*, char*, st_ha_create_information*, TABLE_LIST*, Alter_info*, unsigned int, st_order*, bool) (sql_table.cc:6998) ==16835== ==16835== LEAK SUMMARY: ==16835== definitely lost: 152 bytes in 1 blocks ==16835== indirectly lost: 0 bytes in 0 blocks ==16835== possibly lost: 0 bytes in 0 blocks ==16835== still reachable: 0 bytes in 0 blocks ==16835== suppressed: 0 bytes in 0 blocks ==16835== ==16835== For counts of detected and suppressed errors, rerun with: -v ==16835== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 4 from 4)
Run details:
[roel@fusion1 109]$ cat cmd230 ps -ef | grep 'cmdrun_230' | grep -v grep | awk '{print $2}' | xargs sudo kill -9 rm -Rf /sdd/qa/109/cmdrun_230 mkdir /sdd/qa/109/cmdrun_230 cd /sdd/qa/109/randgen bash -c "set -o pipefail; perl runall.pl --queries=100000000 --seed=8747 --duration=300 --querytimeout=60 --short_column_names --sqltrace --reporter=Shutdown,Backtrace,QueryTimeout,ErrorLog,ErrorLogAlarm --mysqld=--log-output=none --mysqld=--sql_mode=ONLY_FULL_GROUP_BY --grammar=conf/percona_qa/percona_qa.yy --gendata=conf/percona_qa/percona_qa.zz --basedir=/sdd/qa/109/Percona-Server-5.5.28-rel29.2-392.Linux.x86_64 --valgrind --reporter=ValgrindErrors --validator=MarkErrorLog --threads=25 --views --validator=Transformer --mysqld=--slow_query_log --mysqld=--innodb_lazy_drop_table=1 --mysqld=--innodb_file_per_table=1 --mysqld=--innodb_file_format=barracuda --mtr-build-thread=794 --mask=48380 --vardir1=/sdd/qa/109/cmdrun_230 > /sdd/qa/109/cmdrun230.log 2>&1"
Seen during another run (Laurynas expects it is the same issue):
==16835== HEAP SUMMARY: malloc. c:270) search_ and_free_ block (buf0lru.c:940) get_free_ block (buf0lru.c:1285) but_not_ root (btr0btr.ic:59) index_tree (dict0crea.c:795) table_for_ mysql (row0mysql.c:3443) :delete_ table(char const*) (ha_innodb.cc:8382) table(THD* , handlerton*, char const*, char const*, char const*, bool) (handler.cc:2211) table(handlerto n*, char const*, char const*, unsigned int) (sql_table.cc:2410) table(THD* , char*, char*, st_ha_create_ information* , TABLE_LIST*, Alter_info*, unsigned int, st_order*, bool) (sql_table.cc:6998)
==16835== in use at exit: 152 bytes in 1 blocks
==16835== total heap usage: 55,843,989 allocs, 55,843,988 frees, 420,201,789,971 bytes allocated
==16835==
==16835== 152 bytes in 1 blocks are definitely lost in loss record 1 of 1
==16835== at 0x4C278FE: malloc (vg_replace_
==16835== by 0x84B0D2: ut_malloc_low (ut0mem.c:102)
==16835== by 0x883C58: buf_LRU_free_block (buf0buf.ic:871)
==16835== by 0x885016: buf_LRU_
==16835== by 0x8856EE: buf_LRU_
==16835== by 0x87471A: buf_page_get_gen (buf0buf.c:2681)
==16835== by 0x851534: btr_free_
==16835== by 0x88B12E: dict_drop_
==16835== by 0x81EA2E: row_upd_step (row0upd.c:2211)
==16835== by 0x913E9C: que_run_threads (que0que.c:1238)
==16835== by 0x91448F: que_eval_sql (que0que.c:1445)
==16835== by 0x80C47C: row_drop_
==16835== by 0x7EE140: ha_innobase:
==16835== by 0x69C400: ha_delete_
==16835== by 0x5E1ABC: quick_rm_
==16835== by 0x5E42AF: mysql_alter_
==16835==
==16835== LEAK SUMMARY:
==16835== definitely lost: 152 bytes in 1 blocks
==16835== indirectly lost: 0 bytes in 0 blocks
==16835== possibly lost: 0 bytes in 0 blocks
==16835== still reachable: 0 bytes in 0 blocks
==16835== suppressed: 0 bytes in 0 blocks
==16835==
==16835== For counts of detected and suppressed errors, rerun with: -v
==16835== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 4 from 4)
Run details:
[roel@fusion1 109]$ cat cmd230 109/cmdrun_ 230 109/cmdrun_ 230 column_ names --sqltrace --reporter= Shutdown, Backtrace, QueryTimeout, ErrorLog, ErrorLogAlarm --mysqld= --log-output= none --mysqld= --sql_mode= ONLY_FULL_ GROUP_BY --grammar= conf/percona_ qa/percona_ qa.yy --gendata= conf/percona_ qa/percona_ qa.zz --basedir= /sdd/qa/ 109/Percona- Server- 5.5.28- rel29.2- 392.Linux. x86_64 --valgrind --reporter= ValgrindErrors --validator= MarkErrorLog --threads=25 --views --validator= Transformer --mysqld= --slow_ query_log --mysqld= --innodb_ lazy_drop_ table=1 --mysqld= --innodb_ file_per_ table=1 --mysqld= --innodb_ file_format= barracuda --mtr-build- thread= 794 --mask=48380 --vardir1= /sdd/qa/ 109/cmdrun_ 230 > /sdd/qa/ 109/cmdrun230. log 2>&1"
ps -ef | grep 'cmdrun_230' | grep -v grep | awk '{print $2}' | xargs sudo kill -9
rm -Rf /sdd/qa/
mkdir /sdd/qa/
cd /sdd/qa/109/randgen
bash -c "set -o pipefail; perl runall.pl --queries=100000000 --seed=8747 --duration=300 --querytimeout=60 --short_