Comment 2 for bug 1522385
Revision history for this message
| Krunal Bauskar (krunal-bauskar) wrote | #2 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
73416ba
(Get the code!)
With all that analysis in place
a. GTID holes can happen only if user explicitly configures server to suppress error. With normal flow this is not an issue.
b. Action of inserting a dummy trx is slave thread specific action which is outside the galera purview. (Also, it is not an user generated action for Galera to take care of it). May be it is not correct to expect that such actions should be replicated across galera cluster.
------
There are 2 different eco-systems (and both of them operate independently):
1. Master-Slave eco-system
2. Galera eco-system (Multi-Master Eco-system)
Expecting event of one eco-system to trigger series of event in other system doesn't look feasible and creates un-needed dependency path
(especially when these are error events triggered by configuration setting done by user).
I don't foresee we should try to even explore this dependency path.