Bug #86956 Joining group timeout on permissions error
Submitted: 5 Jul 2017 12:04 Modified: 6 Jul 2017 9:57
Reporter: Jervin R Email Updates:
Status: Verified Impact on me:
None 
Category:MySQL Server: Group Replication Severity:S3 (Non-critical)
Version:5.7.18 OS:Any
Assigned to: CPU Architecture:Any

[5 Jul 2017 12:04] Jervin R
Description:
Some errors when joining the group does not need the 60 second timeout, for example - if the port cannot be used, it should fail immediately.

2017-07-05T11:56:13.468301Z 0 [Note] Plugin group_replication reported: 'Unable to bind to 0.0.0.0:24901 (socket=65, errno=13)!'
2017-07-05T11:56:13.468342Z 0 [ERROR] Plugin group_replication reported: 'Unable to announce tcp port 24901. Port already in use?'
2017-07-05T11:56:13.468483Z 0 [ERROR] Plugin group_replication reported: '[GCS] Error joining the group while waiting for the network layer to become ready.'
2017-07-05T11:56:13.468509Z 0 [Note] Plugin group_replication reported: 'state 4257 action xa_exit'
2017-07-05T11:56:13.468943Z 0 [Note] Plugin group_replication reported: 'Exiting xcom thread'
2017-07-05T11:56:13.468970Z 0 [ERROR] Plugin group_replication reported: '[GCS] The member was unable to join the group. Local port: 24901'
2017-07-05T11:57:13.467621Z 14 [ERROR] Plugin group_replication reported: 'Timeout on wait for view after joining group'

How to repeat:
- Enable SELinux
- Configure GR to start on standard port 24901

Suggested fix:
Unable to bind error should fail immediately instead of waiting 60 seconds.
[6 Jul 2017 9:57] Umesh Shastry
Hello Jervin R,

Thank you for the report and feedback.

Thanks,
Umesh
[25 Aug 2017 10:48] Tiago Jorge
Posted by developer:
 
From Group Replication perspective, currently when it joins a group, there are only two possible outcomes:
- Success on view delivery
- Timeout due to some eventual failure

One needs a mechanism to report fail-fast errors between GCS and GR.