MySQL Bugs: #90126: Sig 11 in dict_index_set_merge_threshold

Bug #90126	Sig 11 in dict_index_set_merge_threshold \| dict/dict0dict.cc:5923
Submitted:	19 Mar 2018 14:39	Modified:	20 Mar 2018 16:37
Reporter:	Ramesh Sivaraman	Email Updates:
Status:	Duplicate	Impact on me:	None
Category:	MySQL Server	Severity:	S3 (Non-critical)
Version:	5.7.21	OS:	CentOS
Assigned to:		CPU Architecture:	Any

Description:
mysqld is crashing with sig 11 assertion when we run sysbench with 2048 threads

GDB info

#0  0x00007fb639c0297c in pthread_kill () from /lib64/libpthread.so.0
#1  0x00000000007ce254 in handle_fatal_signal (sig=11) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/sql/signal_handler.cc:220
#2  <signal handler called>
#3  0x00000000011cea15 in dict_index_set_merge_threshold (index=0x7fb1a00adab8, merge_threshold=50) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/storage/innobase/dict/dict0dict.cc:5923
#4  0x0000000000feedb2 in innobase_parse_hint_from_comment (thd=<value optimized out>, table=<value optimized out>, table_share=0x7fb364331a50) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/storage/innobase/handler/ha_innodb.cc:11617
#5  0x0000000000ff2f20 in create_table_info_t::create_table_update_dict (this=0x7fb364331010) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/storage/innobase/handler/ha_innodb.cc:12095
#6  0x0000000000ff4aec in ha_innobase::create (this=<value optimized out>, name=<value optimized out>, form=<value optimized out>, create_info=<value optimized out>) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/storage/innobase/handler/ha_innodb.cc:12176
#7  0x0000000000818c33 in ha_create_table (thd=0x7fb1a0000ae0, path=0x7fb300000000 " ", db=0x7fb1a0006ff8 "test", table_name=0x7fb1a0006a30 "sbtest132", create_info=0x7fb364333c40, update_create_info=false, is_temp_table=false) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/sql/handler.cc:5123
#8  0x0000000000ddc19a in rea_create_table (thd=0x7fb1a0000ae0, path=0x7fb364333320 "./test/sbtest132", db=0x7fb1a0006ff8 "test", table_name=0x7fb1a0006a30 "sbtest132", create_info=0x7fb364333c40, create_fields=<value optimized out>, keys=1, key_info=0x7fb1a000eb60, file=0x7fb1a000e870, no_ha_table=false) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/sql/unireg.cc:555
#9  0x0000000000d7d74a in create_table_impl (thd=0x7fb1a0000ae0, db=0x7fb1a0006ff8 "test", table_name=0x7fb1a0006a30 "sbtest132", error_table_name=0x7fb1a0006a30 "sbtest132", path=0x7fb364333320 "./test/sbtest132", create_info=0x7fb364333c40, alter_info=0x7fb364333640, internal_tmp_table=false, select_field_count=0, no_ha_table=false, is_trans=0x7fb36433359f, key_info=0x7fb364333530, key_count=0x7fb364333538) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/sql/sql_table.cc:5287
#10 0x0000000000d7e01f in mysql_create_table_no_lock (thd=0x7fb1a0000ae0, db=0x7fb1a0006ff8 "test", table_name=0x7fb1a0006a30 "sbtest132", create_info=0x7fb364333c40, alter_info=0x7fb364333640, select_field_count=0, is_trans=0x7fb36433359f) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/sql/sql_table.cc:5417
#11 0x0000000000d7e7bf in mysql_create_table (thd=0x7fb1a0000ae0, create_table=0x7fb1a0006a78, create_info=0x7fb364333c40, alter_info=0x7fb364333640) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/sql/sql_table.cc:5463
#12 0x0000000000d1aa55 in mysql_execute_command (thd=0x7fb1a0000ae0, first_level=true) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/sql/sql_parse.cc:3249
#13 0x0000000000d1af7d in mysql_parse (thd=0x7fb1a0000ae0, parser_state=<value optimized out>) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/sql/sql_parse.cc:5582
#14 0x0000000000d1c19a in dispatch_command (thd=0x7fb1a0000ae0, com_data=0x7fb364334d60, command=COM_QUERY) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/sql/sql_parse.cc:1458
#15 0x0000000000d1d044 in do_command (thd=0x7fb1a0000ae0) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/sql/sql_parse.cc:999
#16 0x0000000000ded7ac in handle_connection (arg=<value optimized out>) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/sql/conn_handler/connection_handler_per_thread.cc:300
#17 0x0000000000f707b4 in pfs_spawn_thread (arg=0x14adb3e0) at /export/home/pb2/build/sb_0-26514852-1514433850.9/mysql-5.7.21/storage/perfschema/pfs.cc:2190
#18 0x00007fb639bfdaa1 in start_thread () from /lib64/libpthread.so.0
#19 0x00007fb63875bbcd in clone () from /lib64/libc.so.6

How to repeat:
Startup;

/data/perf/mysql-5.7.21-linux-glibc2.12-x86_64/bin/mysqld --defaults-file=/home/ramesh/toku-qa/bin/my_inno.cnf --user=ramesh --basedir=/data/perf/mysql-5.7.21-linux-glibc2.12-x86_64 --datadir=/data/perf/mysql-5.7.21-linux-glibc2.12-x86_64/data --innodb_buffer_pool_size=10G --port= --pid-file=/data/perf/mysql-5.7.21-linux-glibc2.12-x86_64/data/pid.pid --core-file --socket=/tmp/ps-perf-mysql.sock --log-error=/data/perf/mysql-5.7.21-linux-glibc2.12-x86_64/data/error.log.out

sysbench run;

sysbench /usr/share/sysbench/oltp_insert.lua --mysql-storage-engine=innodb --rand-type=uniform --threads=2048 --tables=2048 --table-size=100000 --mysql-db=test --mysql-user=root --db-driver=mysql --mysql-socket=/tmp/ps-perf-mysql.sock prepare

my.cnf;

[mysqld]
innodb_flush_method = O_DIRECT
innodb_log_file_size = 2000M
innodb_log_files_in_group = 2
innodb_flush_log_at_trx_commit = 1
innodb_log_buffer_size=128M
innodb_max_dirty_pages_pct=80
innodb_read_io_threads = 8
innodb_write_io_threads = 8
innodb_io_capacity = 5000

#transaction-isolation=READ-COMMITTED
#log-bin=mysql-bin
#server-id=100
max-prepared-stmt-count=100000000
sync_binlog=0
max_connections=5000
table_open_cache=5000
table-definition-cache=1000
query_cache_size=0
query_cache_type=0
performance_schema=0

Hi!

This bug is duplicate of the bug:

https://bugs.mysql.com/bug.php?id=89126

If you disagree with this conclusion, please send us a full CREATE command that produced this assert.

Please, also inform us whether you have used our binary to produce this assert or not ......

Hi!

We would still very much like to have that exact CREATE command that causes the assert.

Error log

Attachment: error.log (application/octet-stream, text), 8.76 KiB.

CREATE TABLE command; 

CREATE TABLE sbtest1907(   id INTEGER NOT NULL AUTO_INCREMENT,   k INTEGER DEFAULT '0' NOT NULL,   c CHAR(120) DEFAULT '' NOT NULL,   pad CHAR(60) DEFAULT '' NOT NULL,   PRIMARY KEY (id) ) /*! ENGINE = innodb */

Also attached error log.

Thank you, very much !!!!!

@Sinisa, hi! What makes you say that this bug is a duplicate of 89126? The stack is different?

Hi!

Several reasons of which I will mention here only one. This bug contains the exact command that causes the assert and this command will be used to reproduce the bug.