Details
Major
Description
Reported by AL13N on IRC.
I am also getting the failure (sporadically, but often enough) on a release build, but not on a debug build.
worker[1] Using MTR_BUILD_THREAD 300, with reserved ports 16000..16019 rpl.show_status_stop_slave_race-7126 'mix' [ pass ] 2121 MTR's internal check of the test case 'rpl.show_status_stop_slave_race-7126' failed. This means that the test case does not preserve the state that existed before the test case was executed. Most likely the test case did not do a proper clean-up. It could also be caused by the previous test run by this thread, if the server wasn't restarted. This is the diff of the states of the servers before and after the test case was executed: mysqltest: Logging to '/data/releases/mariadb-10.0.18-linux-x86_64/mysql-test/var/tmp/check-mysqld_2.log'. mysqltest: Results saved in '/data/releases/mariadb-10.0.18-linux-x86_64/mysql-test/var/tmp/check-mysqld_2.result'. mysqltest: Connecting to server localhost:16001 (socket /data/releases/mariadb-10.0.18-linux-x86_64/mysql-test/var/tmp/mysqld.2.sock) as 'root', connection 'default', attempt 0 ... mysqltest: ... Connected. mysqltest: Start processing test commands from './include/check-testcase.test' ... mysqltest: ... Done processing test commands. --- /data/releases/mariadb-10.0.18-linux-x86_64/mysql-test/var/tmp/check-mysqld_2.result 2015-05-13 02:02:04.000000000 +0300 +++ /data/releases/mariadb-10.0.18-linux-x86_64/mysql-test/var/tmp/check-mysqld_2.reject 2015-05-13 02:02:06.000000000 +0300 @@ -32,8 +32,8 @@ Master_SSL_Key Seconds_Behind_Master NULL Master_SSL_Verify_Server_Cert No -Last_IO_Errno 0 -Last_IO_Error +Last_IO_Errno 2003 +Last_IO_Error error connecting to master 'root@127.0.0.1:16000' - retry-time: 1 retries: 10 message: Can't connect to MySQL server on '127.0.0.1' (4 "Interrupted system call") Last_SQL_Errno 0 Last_SQL_Error Replicate_Ignore_Server_Ids mysqltest: Result length mismatch
Here is what I think is happening:
The test runs "START SLAVE; STOP SLAVE" multiple times concurrently in two connections; so sometimes STOP from one connection interrupts starting from another connection, and the slave IO thread stops with the 'interrupted system call' error. If it happens to be the last state at the end of the concurrent flow, the test ends with this status because it doesn't have replication cleanup at the end.
The change below seems to fix the problem, but I need a confirmation that the "Interrupted system call" status is really allowed here.
diff --git a/mysql-test/suite/rpl/r/show_status_stop_slave_race-7126.result b/mysql-test/suite/rpl/r/show_status_stop_slave_race-7126.result index e71bb2e..c2a0498 100644 --- a/mysql-test/suite/rpl/r/show_status_stop_slave_race-7126.result +++ b/mysql-test/suite/rpl/r/show_status_stop_slave_race-7126.result @@ -1,2 +1,4 @@ include/master-slave.inc [connection master] +start slave; +include/rpl_end.inc diff --git a/mysql-test/suite/rpl/t/show_status_stop_slave_race-7126.test b/mysql-test/suite/rpl/t/show_status_stop_slave_race-7126.test index a79a188..38759c9 100644 --- a/mysql-test/suite/rpl/t/show_status_stop_slave_race-7126.test +++ b/mysql-test/suite/rpl/t/show_status_stop_slave_race-7126.test @@ -10,3 +10,10 @@ --exec $MYSQL_SLAP --silent --socket=$SLAVE_MYSOCK -q "START SLAVE; STOP SLAVE; SHOW GLOBAL STATUS" -c 2 --number-of-queries=100 --create-s # All done. + +--connection slave +start slave; + +--connection master +--source include/rpl_end.inc +
Gliffy Diagrams
Attachments
Activity
- All
- Comments
- Work Log
- History
- Activity
- Transitions
It should be ok. "Interrupted system call" is how STOP SLAVE is able to
cause the IO thread to get out of waiting on the network connection to the
master.
Also applied to 5.5
Kristian Nielsen, could you please take a look to confirm (or not confirm) that it's a test problem only?