[MDEV-9152] PPC64 build fails - binary crashes on all tests Created: 2015-11-18  Updated: 2016-10-20  Resolved: 2016-10-20

Status: Closed
Project: MariaDB Server
Component/s: Platform Debian
Affects Version/s: 10.0
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Otto Kekäläinen Assignee: Sergey Vojtovich
Resolution: Cannot Reproduce Votes: 1
Labels: foundation

Sprint: 10.0.28

 Description   

From log https://buildd.debian.org/status/fetch.php?pkg=mariadb-10.0&arch=ppc64&ver=10.0.22-1&stamp=1446192750 :

federated.federated_partition 'X,innodb_plugin' [ fail ]
 
        Test ended at 2015-10-30 07:32:20
 
 
 
CURRENT_TEST: federated.federated_partition
 
mysqltest: Could not open connection 'default': 2013 Lost connection to MySQL server at 'waiting for initial communication packet', system error: 110 "Connection timed out"
 
 
 
 - saving '/«PKGBUILDDIR»/builddir/mysql-test/var/log/federated.federated_partition-X,innodb_plugin/' to '/«PKGBUILDDIR»/builddir/mysql-test/var/log/federated.federated_partition-X,innodb_plugin/'
 
***Warnings generated in error logs during shutdown after running tests: federated.federated_partition
 
 
 
151030  7:30:19 [ERROR] mysqld got signal 11 ;
 
Attempting backtrace. You can use the following information to find out
 
151030  7:30:19 [ERROR] mysqld got signal 11 ;
 
Attempting backtrace. You can use the following information to find out
 
 
 
federated.federated_transactions 'X,innodb_plugin' [ fail ]
 
        Test ended at 2015-10-30 07:34:21
 
 
 
CURRENT_TEST: federated.federated_transactions
 
mysqltest: Could not open connection 'default': 2013 Lost connection to MySQL server at 'waiting for initial communication packet', system error: 110 "Connection timed out"
 
 
 
 - saving '/«PKGBUILDDIR»/builddir/mysql-test/var/log/federated.federated_transactions-X,innodb_plugin/' to '/«PKGBUILDDIR»/builddir/mysql-test/var/log/federated.federated_transactions-X,innodb_plugin/'
 
***Warnings generated in error logs during shutdown after running tests: federated.federated_transactions
 
 
 
151030  7:32:21 [ERROR] mysqld got signal 11 ;
 
Attempting backtrace. You can use the following information to find out
 
151030  7:32:21 [ERROR] mysqld got signal 11 ;
 
Attempting backtrace. You can use the following information to find out
 
 
 
federated.federated_innodb 'X,innodb_plugin' [ fail ]
 
        Test ended at 2015-10-30 07:36:24
 
 
 
CURRENT_TEST: federated.federated_innodb
 
mysqltest: Could not open connection 'default': 2013 Lost connection to MySQL server at 'waiting for initial communication packet', system error: 110 "Connection timed out"
 
 
 
 - saving '/«PKGBUILDDIR»/builddir/mysql-test/var/log/federated.federated_innodb-X,innodb_plugin/' to '/«PKGBUILDDIR»/builddir/mysql-test/var/log/federated.federated_innodb-X,innodb_plugin/'
 
***Warnings generated in error logs during shutdown after running tests: federated.federated_innodb
 
 
 
151030  7:34:23 [ERROR] mysqld got signal 11 ;
 
Attempting backtrace. You can use the following information to find out
 
151030  7:34:23 [ERROR] mysqld got signal 11 ;
 
Attempting backtrace. You can use the following information to find out
 
 
 
federated.federated_partition 'X,xtradb' [ fail ]
 
        Test ended at 2015-10-30 07:36:26
 
 
 
CURRENT_TEST: federated.federated_partition
 
 
 
 
 
Server [mysqld.2 - pid: 14820, winpid: 14820, exit: 256] failed during test run
 
Server log from this test:
 
----------SERVER LOG START-----------
 
151030  7:36:24 [Note] /«PKGBUILDDIR»/builddir/sql/mysqld (mysqld 10.0.22-MariaDB-1-log) starting as process 14821 ...
 
151030  7:36:24 [Note] Plugin 'BLACKHOLE' is disabled.
 
151030  7:36:24 [Note] InnoDB: Using mutexes to ref count buffer pool pages
 
151030  7:36:24 [Note] InnoDB: The InnoDB memory heap is disabled
 
151030  7:36:24 [Note] InnoDB: Mutexes and rw_locks use GCC atomic builtins
 
151030  7:36:24 [Note] InnoDB: GCC builtin __atomic_thread_fence() is used for memory barrier
 
151030  7:36:24 [Note] InnoDB: Compressed tables use zlib 1.2.8
 
151030  7:36:24 [Note] InnoDB: Using Linux native AIO
 
151030  7:36:24 [Note] InnoDB: Not using CPU crc32 instructions
 
151030  7:36:24 [Note] InnoDB: Initializing buffer pool, size = 8.0M
 
151030  7:36:24 [Note] InnoDB: Completed initialization of buffer pool
 
151030  7:36:24 [Note] InnoDB: Highest supported file format is Barracuda.
 
151030  7:36:25 [Note] InnoDB: 128 rollback segment(s) are active.
 
151030  7:36:25 [Note] InnoDB: Waiting for purge to start
 
151030  7:36:25 [Note] InnoDB:  Percona XtraDB (http://www.percona.com) 5.6.26-74.0 started; log sequence number 1629444
 
151030  7:36:25 [Note] Plugin 'XTRADB_READ_VIEW' is disabled.
 
151030  7:36:25 [Note] Plugin 'XTRADB_INTERNAL_HASH_TABLES' is disabled.
 
151030  7:36:25 [Note] Plugin 'XTRADB_RSEG' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_LOCK_WAITS' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_CMP_RESET' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_CMPMEM_RESET' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_CMP_PER_INDEX_RESET' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_FT_DEFAULT_STOPWORD' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_FT_DELETED' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_FT_BEING_DELETED' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_FT_CONFIG' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_FT_INDEX_CACHE' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_FT_INDEX_TABLE' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_SYS_TABLESTATS' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_SYS_INDEXES' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_SYS_COLUMNS' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_SYS_FIELDS' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_SYS_TABLESPACES' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_SYS_DATAFILES' is disabled.
 
151030  7:36:25 [Note] Plugin 'INNODB_CHANGED_PAGES' is disabled.
 
151030  7:36:25 [Note] Plugin 'ARCHIVE' is disabled.
 
151030  7:36:25 [Note] Plugin 'FEEDBACK' is disabled.
 
151030  7:36:25 [Warning] /«PKGBUILDDIR»/builddir/sql/mysqld: unknown variable 'loose-debug-sync-timeout=300'
 
151030  7:36:25 [Note] Server socket created on IP: '127.0.0.1'.
 
151030  7:36:25 [Note] /«PKGBUILDDIR»/builddir/sql/mysqld: ready for connections.
 
Version: '10.0.22-MariaDB-1-log'  socket: '/«PKGBUILDDIR»/builddir/mysql-test/var/tmp/mysqld.2.sock'  port: 16001  Debian unstable
 
151030  7:36:25 [ERROR] mysqld got signal 11 ;
 
This could be because you hit a bug. It is also possible that this binary
 
or one of the libraries it was linked against is corrupt, improperly built,
 
or misconfigured. This error can also be caused by malfunctioning hardware.
 
 
 
To report this bug, see http://kb.askmonty.org/en/reporting-bugs
 
 
 
We will try our best to scrape up some info that will hopefully help
 
diagnose the problem, but since we have already crashed, 
something is definitely wrong and this may fail.
 
 
 
Server version: 10.0.22-MariaDB-1-log
 
key_buffer_size=1048576
 
read_buffer_size=131072
 
max_used_connections=3
 
max_threads=153
 
thread_count=1
 
It is possible that mysqld could use up to 
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 62779 K  bytes of memory
 
Hope that's ok; if not, decrease some variables in the equation.
 
 
 
Thread pointer: 0x0x3fff957fb008
 
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 = 0x3fff9a9865c0 thread_stack 0x48000
 
mysys/stacktrace.c:246(my_print_stacktrace)[0x10a17e70]
 
sql/signal_handler.cc:153(handle_fatal_signal)[0x103e3034]
 
linux-vdso64.so.1(__kernel_sigtramp_rt64+0x0)[0x3fff9b1b0478]
 
myisam/mi_open.c:65(test_if_reopen)[0x1092c398]
 
mysys/my_malloc.c:67(update_malloc_size)[0x10a11a14]
 
myisam/mi_create.c:638(mi_create)[0x109199c4]
 
myisam/ha_myisam.cc:2069(ha_myisam::create(char const*, TABLE*, HA_CREATE_INFO*))[0x10905bb8]
 
sql/handler.cc:4328(handler::ha_create(char const*, TABLE*, HA_CREATE_INFO*))[0x103ef334]
 
sql/handler.cc:4699(ha_create_table(THD*, char const*, char const*, char const*, HA_CREATE_INFO*, st_mysql_const_unsigned_lex_string*))[0x103f00b8]
 
sql/unireg.cc:378(rea_create_table(THD*, st_mysql_const_unsigned_lex_string*, char const*, char const*, char const*, HA_CREATE_INFO*, handler*, bool))[0x102e397c]
 
sql/sql_table.cc:4826(create_table_impl(THD*, char const*, char const*, char const*, char const*, char const*, HA_CREATE_INFO*, Alter_info*, int, bool*, st_key**, unsigned int*, st_mysql_const_unsigned_lex_string*) [clone .constprop.108])[0x102a817c]
 
sql/sql_table.cc:4941(mysql_create_table_no_lock(THD*, char const*, char const*, HA_CREATE_INFO*, Alter_info*, bool*, int))[0x102a87f0]
 
sql/sql_table.cc:5001(mysql_create_table(THD*, TABLE_LIST*, HA_CREATE_INFO*, Alter_info*))[0x102a8af4]
 
sql/sql_parse.cc:3035(mysql_execute_command(THD*))[0x102046e4]
 
sql/sql_parse.cc:6546(mysql_parse(THD*, char*, unsigned int, Parser_state*))[0x10204a58]
 
sql/sql_parse.cc:1309(dispatch_command(enum_server_command, THD*, char*, unsigned int))[0x10206954]
 
sql/sql_connect.cc:1377(do_handle_one_connection(THD*))[0x10305ff8]
 
sql/sql_connect.cc:1294(handle_one_connection)[0x10306124]
 
perfschema/pfs.cc:1860(pfs_spawn_thread)[0x10608f44]
 
/lib/powerpc64-linux-gnu/libpthread.so.0(+0xc688)[0x3fff9b16c688]
 
/lib/powerpc64-linux-gnu/libc.so.6(clone-0xa6cec)[0x3fff9ac05cd4]
 
 
 
Trying to get some variables.
 
Some pointers may be invalid and cause the dump to abort.
 
Query (0x3fff891a7020): create temporary table error_log ( row int auto_increment primary key, suspicious int default 1, file_name varchar(255), line varchar(1024) default null ) engine=myisam
 
Connection ID (thread ID): 7
 
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=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=on,exists_to_in=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

This problem has been open for long, but at least in 10.0.14 some tests passed successfully so the binary has not always been completely broken: https://buildd.debian.org/status/fetch.php?pkg=mariadb-10.0&arch=ppc64&ver=10.0.14-1&stamp=1412211365

 Comments   
Comment by Elena Stepanova [ 2015-11-19 ]

Assigned to Monty because he said he would look into this.

Comment by Elena Stepanova [ 2016-08-21 ]

svoj,

Please take ownership of this one and decide what we do about it.
I don't see how it can be solved from the test perspective if everything fails – can't disable them all.

Comment by Otto Kekäläinen [ 2016-08-22 ]

Current Debian unstable build status of package mariadb-10.0 is visible at https://buildd.debian.org/status/package.php?p=mariadb-10.0
The ppc64el arch build is currently OK. It has one override defined in https://anonscm.debian.org/cgit/pkg-mysql/mariadb-10.0.git/tree/debian/unstable-tests.ppc64el but it is not this MDEV.

I don't recall who fixed this and when. I ran on the 10.0 branch git log | grep -F MDEV-9152 without results.

It would be good to document what fixed this and then close the issue. If nobody recalls committing a fix, then just close this.

Comment by Daniel Black [ 2016-09-01 ]

otto, your original bug here was ppc64 (which is big endian). https://buildd.debian.org/status/logs.php?pkg=mariadb-10.0&arch=ppc64 which did fail its last build however it seemed to be a timeout of some sort. It did seem to pass federated.federated_innodb test which failed here. I agree that it seems fixed by reasons unknown.

Comment by Sergey Vojtovich [ 2016-10-20 ]

There're about 600 revisions between 10.0.22 (last failing) and 10.0.26 (first not failing). FWICS versions between those haven't been tested by this builder.

Traces don't look meaningful either: it usually fails in MyISAM, but there's suspicious frame that must not be there, e.g.:

myisam/mi_open.c:65(test_if_reopen)[0x10930714]
 
psi/mysql_thread.h:681(inline_mysql_mutex_lock)[0x109306e4]
 
myisam/ha_myisam.cc:755(ha_myisam::open(char const*, int, unsigned int))[0x109066d4]

inline_mysql_mutex_lock must not be there.

Sometimes mysqltest crashes.

Said the above I doubt it's feasible to find revision that fixed this. It might happen so that the problem was in some third party library that we were linking against. I guess this host went through a few upgrades meanwhile.

Closing as can't repeat. Feel free to reopen if the problem is still there.

Comment by Otto Kekäläinen [ 2016-10-20 ]

Note that currently ppc64 and ppc64el builds find and test suite passes in Debian with the patches and skip-tests applied.

Latest build status: https://buildd.debian.org/status/package.php?p=mariadb-10.0

Patches: https://anonscm.debian.org/cgit/pkg-mysql/mariadb-10.0.git/tree/debian/patches

Skiplist for ppc64el: https://anonscm.debian.org/cgit/pkg-mysql/mariadb-10.0.git/tree/debian/unstable-tests.ppc64el

cvicentiu Can you please upstream the patches (as you choose) in Debian so that they will eventually benefit everybody? Thanks

Generated at Thu Feb 08 07:32:31 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.