Getting error from Slave SQL thread when it is started before the PXC 8.0 node is ready to accept connections while joining the PXC cluster
General
Escalation
General
Escalation
Description
Environment
None
AFFECTED CS IDs
CS0028197
Attachments
1
Smart Checklist
Activity
Show:

puneet.kaushik September 6, 2022 at 8:44 AM
MTR test passed successfully !
Done
Details
Details
Assignee

Reporter

Labels
Needs Review
Yes
Fix versions
Affects versions
Priority
Smart Checklist
Open Smart Checklist
Smart Checklist

Open Smart Checklist
Created July 12, 2022 at 6:18 AM
Updated March 6, 2024 at 8:54 PM
Resolved August 10, 2022 at 9:21 AM
If a PXC node is also an asynchronous slave, the SQL thread may fail at startup if the node is not yet ready to join the cluster. To set this up, create a PXC cluster and make one node an asynchronous slave to a master. Then, on another PXC node and master node, let it receive writes from different databases:
Eg.
Master:
Another PXC node:
Then, try to restart the PXC 8.0 node that is also asynchronous slave repeatedly until you get this error:
As stated above that you need to test repeatedly because I cannot consistently get the Slave SQL thread to start before the the node is ready.
On the other hand, in PXC 5.7 it seems that Slave SQL thread is started after the node is ready to receive connections: