[MDEV-15572] view.test, server crash with --big-tables=1 Created: 2018-03-15  Updated: 2019-07-18  Resolved: 2019-07-18

Status: Closed
Project: MariaDB Server
Component/s: Views
Affects Version/s: 5.5, 10.0, 10.1, 10.2, 10.3, 10.4
Fix Version/s: 10.2.26, 5.5.65, 10.1.41, 10.3.17, 10.4.7

Type: Bug Priority: Critical
Reporter: Alice Sherepa Assignee: Oleksandr Byelkin
Resolution: Fixed Votes: 0
Labels: affects-tests

Issue Links:
Duplicate
is duplicated by MDEV-18148 Server crashes in ha_maria::end_bulk_... Closed
is duplicated by MDEV-19420 Server crashes in ha_maria::extra upo... Closed
Relates
relates to MDEV-18451 Server crashes in maria_create_trn_fo... Closed

 Description   

to reproduce:

./mtr view.test --mysqld=--big-tables=1


main.view                                [ fail ]
 
        Test ended at 2018-03-15 13:54:25
 
 
 
CURRENT_TEST: main.view
 
mysqltest: At line 1380: query 'insert into v3(a,b) select * from t2' failed with wrong errno 2013: 'Lost connection to MySQL server during query', instead of 1393...
 
 
 
The result from queries just before the failure was:
 
< snip >
 
select * from v3;
 
a	b
 
2	1
 
300	1
 
2	3
 
300	3
 
deallocate prepare stmt1;
 
drop view v3,v2;
 
drop tables t1,t2;
 
create table t1 (a int, primary key (a), b int);
 
create table t2 (a int, primary key (a), b int);
 
insert into t2 values (1000, 2000);
 
create view v3 (a,b) as select t1.a as a, t2.a as b from t1, t2;
 
insert into v3 values (1,2);
 
ERROR HY000: Can not insert into join view 'test.v3' without fields list
 
insert into v3 select * from t2;
 
ERROR HY000: Can not insert into join view 'test.v3' without fields list
 
insert into v3(a,b) values (1,2);
 
ERROR HY000: Can not modify more than one base table through a join view 'test.v3'
 
insert into v3(a,b) select * from t2;


Server version: 10.2.14-MariaDB-debug-log
 
key_buffer_size=1048576
 
read_buffer_size=131072
 
max_used_connections=1
 
max_threads=153
 
thread_count=2
 
It is possible that mysqld could use up to 
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 63177 K  bytes of memory
 
Hope that's ok; if not, decrease some variables in the equation.
 
 
 
Thread pointer: 0x7f3c5c000a98
 
Attempting backtrace. You can use the following information to find out
 
where mysqld died. If you see no messages after this, something went
 
terribly wrong...
 
stack_bottom = 0x7f3c782cfe70 thread_stack 0x49000
 
mysys/stacktrace.c:267(my_print_stacktrace)[0x55c9f9d96af8]
 
sql/signal_handler.cc:168(handle_fatal_signal)[0x55c9f9629037]
 
/lib/x86_64-linux-gnu/libpthread.so.0(+0x11390)[0x7f3c7e443390]
 
maria/ma_write.c:1793(maria_end_bulk_insert)[0x55c9f9c9f217]
 
maria/ha_maria.cc:2223(ha_maria::end_bulk_insert())[0x55c9f9c2706d]
 
sql/handler.h:2905(handler::ha_end_bulk_insert())[0x55c9f939df6b]
 
sql/sql_insert.cc:3948(select_insert::abort_result_set())[0x55c9f939bb98]
 
sql/sql_select.cc:383(handle_select(THD*, LEX*, select_result*, unsigned long))[0x55c9f93f68eb]
 
sql/sql_parse.cc:4525(mysql_execute_command(THD*))[0x55c9f93bb692]
 
sql/sql_parse.cc:7902(mysql_parse(THD*, char*, unsigned int, Parser_state*, bool, bool))[0x55c9f93c61e0]
 
sql/sql_parse.cc:1808(dispatch_command(enum_server_command, THD*, char*, unsigned int, bool, bool))[0x55c9f93b3e0f]
 
sql/sql_parse.cc:1360(do_command(THD*))[0x55c9f93b276b]
 
sql/sql_connect.cc:1335(do_handle_one_connection(CONNECT*))[0x55c9f9500ee9]
 
sql/sql_connect.cc:1242(handle_one_connection)[0x55c9f9500c69]
 
perfschema/pfs.cc:1864(pfs_spawn_thread)[0x55c9f985e8ce]
 
/lib/x86_64-linux-gnu/libpthread.so.0(+0x76ba)[0x7f3c7e4396ba]
 
x86_64/clone.S:111(clone)[0x7f3c7d8ce41d]
 
 
 
Trying to get some variables.
 
Some pointers may be invalid and cause the dump to abort.
 
Query (0x7f3c5c1bdc80): insert into v3(a,b) select * from t2
 
Connection ID (thread ID): 4
 
Status: NOT_KILLED
 
 
 
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=off,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=on,exists_to_in=on,orderby_uses_equalities=on,condition_pushdown_for_derived=on
 
 
 
The manual page at http://dev.mysql.com/doc/mysql/en/crashing.html contains
 
information that should help you find out what is causing the crash.
 
Writing a core file at /home/alice/git/10.2/mysql-test/var/mysqld.1/data/
 
----------SERVER LOG END-------------
 



 
 
Thread 1 (Thread 0x7f3c782d0700 (LWP 31990)):
 
#0  __pthread_kill (threadid=<optimized out>, signo=11) at ../sysdeps/unix/sysv/linux/pthread_kill.c:62
 
#1  0x000055c9f9d96bea in my_write_core (sig=11) at /home/alice/git/10.2/mysys/stacktrace.c:477
 
#2  0x000055c9f962945f in handle_fatal_signal (sig=11) at /home/alice/git/10.2/sql/signal_handler.cc:305
 
#3  <signal handler called>
 
#4  0x000055c9f9c9f217 in maria_end_bulk_insert (info=0x0) at /home/alice/git/10.2/storage/maria/ma_write.c:1793
 
#5  0x000055c9f9c2706d in ha_maria::end_bulk_insert (this=0x7f3c5c1d8110) at /home/alice/git/10.2/storage/maria/ha_maria.cc:2222
 
#6  0x000055c9f939df6b in handler::ha_end_bulk_insert (this=0x7f3c5c1d8110) at /home/alice/git/10.2/sql/handler.h:2905
 
#7  0x000055c9f939bb98 in select_insert::abort_result_set (this=0x7f3c5c1d8930) at /home/alice/git/10.2/sql/sql_insert.cc:3932
 
#8  0x000055c9f93f68eb in handle_select (thd=0x7f3c5c000a98, lex=0x7f3c5c004578, result=0x7f3c5c1d8930, setup_tables_done_option=1073741824) at /home/alice/git/10.2/sql/sql_select.cc:382
 
#9  0x000055c9f93bb692 in mysql_execute_command (thd=0x7f3c5c000a98) at /home/alice/git/10.2/sql/sql_parse.cc:4525
 
#10 0x000055c9f93c61e0 in mysql_parse (thd=0x7f3c5c000a98, rawbuf=0x7f3c5c1bdc80 "insert into v3(a,b) select * from t2", length=36, parser_state=0x7f3c782cf1f0, is_com_multi=false, is_next_command=false) at /home/alice/git/10.2/sql/sql_parse.cc:7902
 
#11 0x000055c9f93b3e0f in dispatch_command (command=COM_QUERY, thd=0x7f3c5c000a98, packet=0x7f3c5c082089 "insert into v3(a,b) select * from t2", packet_length=36, is_com_multi=false, is_next_command=false) at /home/alice/git/10.2/sql/sql_parse.cc:1806
 
#12 0x000055c9f93b276b in do_command (thd=0x7f3c5c000a98) at /home/alice/git/10.2/sql/sql_parse.cc:1360
 
#13 0x000055c9f9500ee9 in do_handle_one_connection (connect=0x55c9fbca1ba8) at /home/alice/git/10.2/sql/sql_connect.cc:1335
 
#14 0x000055c9f9500c69 in handle_one_connection (arg=0x55c9fbca1ba8) at /home/alice/git/10.2/sql/sql_connect.cc:1241
 
#15 0x000055c9f985e8ce in pfs_spawn_thread (arg=0x55c9fbc46ee8) at /home/alice/git/10.2/storage/perfschema/pfs.cc:1862
 
#16 0x00007f3c7e4396ba in start_thread (arg=0x7f3c782d0700) at pthread_create.c:333
 
#17 0x00007f3c7d8ce41d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:109



 Comments   
Comment by Elena Stepanova [ 2018-12-26 ]

Still reproducible.

Comment by Alice Sherepa [ 2019-02-13 ]

test case from MDEV-18148:

set big_tables=ON;
 
 
CREATE TABLE t1 ( f1 int , f2 int , f3 int , f4 int);
 
CREATE TABLE t2 ( f1 int , f2 int , f3 int , f4 int);
 
 
CREATE VIEW v1 AS 
	SELECT t2.f1, t1.f2, t2.f3, t2.f4 FROM (t1 JOIN t2);
 
 
REPLACE INTO v1 (f1, f2, f3, f4)
 
	SELECT f1, f2, f3, f4 FROM t1;

Comment by Oleksandr Byelkin [ 2019-05-08 ]

test suite for 10.2 and above

CREATE TABLE t1 (a INT, b BLOB);
 
CREATE TABLE t2 (c INT);
 
CREATE TRIGGER tr BEFORE UPDATE ON t2 FOR EACH ROW SET @a = 1;
 
CREATE VIEW v AS SELECT a, b FROM t1 JOIN t2;
 
 
--error ER_BAD_FIELD_ERROR
 
INSERT INTO v (x) SELECT 1;
 
 
# Cleanup
 
DROP VIEW v;
 
DROP TABLE t1, t2;

Comment by Oleksandr Byelkin [ 2019-05-08 ]

The problem is that we are trying to deal with not opened table of view dummy.

diff --git a/sql/sql_insert.cc b/sql/sql_insert.cc
 
index 4005153cb64..3bcff7bdd56 100644
 
--- a/sql/sql_insert.cc
 
+++ b/sql/sql_insert.cc
 
@@ -3945,7 +3945,7 @@ void select_insert::abort_result_set() {
 
     will be NULL. In that case, we still need to execute the rollback
 
     and the end of the function.
 
    */
 
-  if (table)
 
+  if (table && table->query_id)
 
   {
 
     bool changed, transactional_table;
 
     /*

"official" way of getting known that TABLE is not opened should be found or the check should be used with check of being internal temporary table.

Comment by Oleksandr Byelkin [ 2019-07-11 ]

commit 985822666e5b72a074e53ff5072a585b0a9725c7 (HEAD > bb-5.5MDEV-15572, origin/bb-5.5-MDEV-15572)
Author: Oleksandr Byelkin <sanja@mariadb.com>
Date: Thu Jul 11 14:50:50 2019 +0200

MDEV-15572: view.test, server crash with --big-tables=1

Check that table is really opened before cleanup using handler.

Comment by Sergei Golubchik [ 2019-07-15 ]

ok to push

Generated at Thu Feb 08 08:22:24 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.