Bug #81007 rpl_show_slave_hosts fails with valgrind in 7.5
Submitted: 8 Apr 2016 9:14 Modified: 28 Apr 2016 12:35
Reporter: Magnus Blåudd Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S2 (Serious)
Version:5.7.2 OS:Any
Assigned to: CPU Architecture:Any

[8 Apr 2016 9:14] Magnus Blåudd
Description:
The rpl_show_slave_hosts test fails in MySQL Cluster 7.5 when running test with valgrind. The test hangs and it seems like the third mysqld(mysqld.3) is the problem since it fails to connect to the master.

The mysqld.3 log file indicates it's trying to connect and you my feeling is that the interesting log lines are below:

2016-04-07T23:09:24.978322Z 2 [Note] 'CHANGE MASTER TO FOR CHANNEL '' executed'. Previous state master_host='', master_port= 3306, master_log_file='', master_log_pos= 4, master_bind=''. New state master_host='127.0.0.1', master_port= 13030, master_log_file='', master_log_pos= 4, master_bind=''. 2016-04-07T23:09:25.118752Z 3 [Warning] Storing MySQL user name or password information in the master info repository is not secure and is therefore not recommended. Please consider using the USER and PASSWORD connection options for START SLAVE; see the 'START SLAVE Syntax' in the MySQL Manual for more information.
2016-04-07T23:09:30.496304Z 3 [ERROR] Slave I/O for channel '': error connecting to master 'root@127.0.0.1:13030' - retry-time: 60  retries: 1, Error_code: 2026
2016-04-07T23:10:35.536533Z 3 [ERROR] Slave I/O for channel '': error connecting to master 'root@127.0.0.1:13030' - retry-time: 60  retries: 2, Error_code: 2026
2016-04-07T23:11:40.555211Z 3 [ERROR] Slave I/O for channel '': error connecting to master 'root@127.0.0.1:13030' - retry-time: 60  retries: 3, Error_code: 2026

How to repeat:
Run rpl_show_slave_hosts with valgrind. Reproduced both in PB and locally.

./mtr rpl_show_slave_hosts --valgrind

Suggested fix:
My suspicion is that this is caused by the retry count backport which we have in 7.5 only. Nothing else differs compared to 5.7. Checking now by compiling 5.7 and shoudl try to revert/disable that patch to confirm.

The suspicion comes from the fact that rpl_show_slave_hosts.cnf has these line configuring mysqld.3:

[mysqld.3]
server_id=3
report-host=slave2
slave-net-timeout=5
^^^^^^^^^^^^^^^^^^^

Don't understand if the mysqld is supposed to fail to connect to master but anyway it apparently hangs somehow.
[28 Apr 2016 12:35] Jon Stephens
Fixed in NDB 7.5.2. Affects testing only; no changelog entry needed.

Closed.