Description:
Found while analysing the test failures of the 6.3.17 build:
We have 91 cases of tests aborting with this message:
=====
<<testname>> [ fail ]
mysqltest: In included file "./include/ndb_not_readonly.inc": At line 25: Failed while waiting for mysqld to come out of readonly mode
Stopping All Servers
=====
90 of them occurred on the SuSE9-x86_64-RPM platform,
1 on the RedHat3-IA64-RPM platform.
It happened in all 3 configurations ("cluster-gpl", "cluster-com", and "cluster-com-pro")
and in various runs ("ndb", "ndb+rpl_ndb+ps", "funcs1+ps", "partitions"
There are other failures related to this include file which I will report separately.
-----
Only some of these reports were followed by additional information:
Warnings from just before the error:
Error 1146 Table 'mysql.ndb_schema' doesn't exist
(Tests "ndb_storedproc_03", "ndb_restore_compat", "rpl_ndb_ddl", "rpl_ndb_delete_nowhere", "rpl_ndb_sp003", "rpl_ndb_trig004")
Warnings from just before the error:
Error 1296 Got error 4009 'Cluster Failure' from NDB
Error 1296 Got error 157 'Unknown error code' from NDBCLUSTER
(Test "ndb_binlog_format")
The result from queries just before the failure was:
select * from mysql.ndb_apply_status;
More results from queries before failure can be found in /PATH/mysql-test/var/log/rpl_ndb_apply_status.log
(Test "rpl_ndb_apply_status")
The result from queries just before the failure was:
select * from mysql.ndb_apply_status;
More results from queries before failure can be found in /PATH/mysql-test/var/log/rpl_ndb_apply_status.log
Warnings from just before the error:
Error 1146 Table 'mysql.ndb_schema' doesn't exist
(Also test "rpl_ndb_apply_status", other config/run)
How to repeat:
This happened in a release build and its test runs.
Suggested fix:
Not really sure.
AIUI, this happens when 60 seconds (600 turns of a loop with a 0.1 s sleep) are not sufficient for the server to report successful execution of a command requiring access to NDB.
It would surprise me if this machine, and just this one, were so loaded that it could not set up that connection in time.
Description: Found while analysing the test failures of the 6.3.17 build: We have 91 cases of tests aborting with this message: ===== <<testname>> [ fail ] mysqltest: In included file "./include/ndb_not_readonly.inc": At line 25: Failed while waiting for mysqld to come out of readonly mode Stopping All Servers ===== 90 of them occurred on the SuSE9-x86_64-RPM platform, 1 on the RedHat3-IA64-RPM platform. It happened in all 3 configurations ("cluster-gpl", "cluster-com", and "cluster-com-pro") and in various runs ("ndb", "ndb+rpl_ndb+ps", "funcs1+ps", "partitions" There are other failures related to this include file which I will report separately. ----- Only some of these reports were followed by additional information: Warnings from just before the error: Error 1146 Table 'mysql.ndb_schema' doesn't exist (Tests "ndb_storedproc_03", "ndb_restore_compat", "rpl_ndb_ddl", "rpl_ndb_delete_nowhere", "rpl_ndb_sp003", "rpl_ndb_trig004") Warnings from just before the error: Error 1296 Got error 4009 'Cluster Failure' from NDB Error 1296 Got error 157 'Unknown error code' from NDBCLUSTER (Test "ndb_binlog_format") The result from queries just before the failure was: select * from mysql.ndb_apply_status; More results from queries before failure can be found in /PATH/mysql-test/var/log/rpl_ndb_apply_status.log (Test "rpl_ndb_apply_status") The result from queries just before the failure was: select * from mysql.ndb_apply_status; More results from queries before failure can be found in /PATH/mysql-test/var/log/rpl_ndb_apply_status.log Warnings from just before the error: Error 1146 Table 'mysql.ndb_schema' doesn't exist (Also test "rpl_ndb_apply_status", other config/run) How to repeat: This happened in a release build and its test runs. Suggested fix: Not really sure. AIUI, this happens when 60 seconds (600 turns of a loop with a 0.1 s sleep) are not sufficient for the server to report successful execution of a command requiring access to NDB. It would surprise me if this machine, and just this one, were so loaded that it could not set up that connection in time.