workaround:

SELECT COUNT FROM spamov_importacao_arquivos AS a
WHERE tipo='c' AND datahora_gmt>='0000-00-00 00:00:00' AND datahora_gmt<='2012-12-03 01:59:59'
AND (tipo_arquivo,arquivo_id) NOT IN (
SELECT tipo_arquivo,arquivo_id FROM spamov_importacao_arquivos_datas WHERE
tipo_arquivo='c' and arquivo_id=a.arquivo_id
)

OBS: this bug occurs in mariadb 5.5.28a too and 5.5.25 too

Hi Roberto,

Please provide the error log contents (from server startup and up to and including the crash), and your cnf file – the real one. I saw the config you provided in MDEV-3905, but it cannot be right – it contains "skip-pbxt", but we don't have PBXT in 10.0.0, so your server wouldn't have even started (would have complained about an unknown option and aborted). So, either your config is different, or you are using a different version of MariaDB or a custom build. If you attach both the config file and the error log output, we will hopefully be able to figure it out.

yes you can use mdev-3905 config and error log, that's the original one (the only change is password, i put ##### )
i'm using mysqld_multi start
and it starts normally with skip-pbxt option inside my.cnf file
here is ps ax | grep where you can see it running normally with skip-pbxt option

[root@SPADIM-DELL ~]# ps ax | grep mysql
9264 ? Sl 0:14 /usr/bin/mysqld --basedir=/usr --datadir=/home/mysql/data/dev/ --plugin-dir=/usr/lib/mysql/plugin --user=mysql --skip-pbxt --ignore-builtin-innodb --default-storage-engine=MyISAM --skip-name-resolve --lc-messages-dir=/usr/share/mysql --lc-messages=en_US --slow-query-log=ON --slow-query-log-file=/home/mysql/log/dev.slow_query.log --long-query-time=3 --max-connections=20 --key-buffer=8M --table-cache=200 --sort-buffer-size=1M --read-buffer-size=1M --read-rnd-buffer-size=1M --myisam-sort-buffer-size=4M --tmp-table-size=1M --max-allowed-packet=1M --net-buffer-length=16K --wait-timeout=30 --sql-mode=ALLOW_INVALID_DATES,PIPES_AS_CONCAT --concurrent-insert=2 --div-precision-increment=5 --query-cache-size=4M --query-cache-limit=1M --query-cache-type=1 --query-cache-min-res-unit=512 --thread-cache-size=20 --thread-concurrency=4 --myisam-recover=BACKUP,FORCE,QUICK --log-error=/home/mysql/log/dev.error.log --pid-file=/home/mysql/pid/dev.pid --socket=/tmp/mysql.dev.sock --port=3306
28451 ? S 0:02 /bin/sh /usr/bin/mysqld_safe --skip-pbxt --ignore_builtin_innodb --default_storage_engine=MyISAM --skip-name-resolve --socket=/tmp/mysql.dev.sock --port=3306 --pid-file=/home/mysql/pid/dev.pid --datadir=/home/mysql/data/dev/ --lc-messages-dir=/usr/share/mysql --lc-messages=en_US --slow_query_log=ON --slow_query_log_file=/home/mysql/log/dev.slow_query.log --log-error=/home/mysql/log/dev.error.log --long_query_time=3 --max_connections=20 --key_buffer=8M --table_cache=200 --sort_buffer_size=1M --read_buffer_size=1M --read_rnd_buffer_size=1M --myisam_sort_buffer_size=4M --tmp_table_size=1M --max_allowed_packet=1M --net_buffer_length=16K --wait_timeout=30 --sql_mode=ALLOW_INVALID_DATES,PIPES_AS_CONCAT --concurrent_insert=2 --div_precision_increment=5 --query_cache_size=4M --query_cache_limit=1M --query_cache_type=1 --query_cache_min_res_unit=512 --thread_cache_size=20 --thread_concurrency=4 --myisam-recover=BACKUP,FORCE,QUICK

The request for contents of the error log file is still in force. Please provide the full log from server start and up to the crash. Thanks.

Closing as incomplete for now, if you have the error log output requested above, please attach it and re-open the bug report.

hi elena, i will try again in near future
i change the query to the work around that i post, and servers that have this problem are production i will try a time gap to test soon