There is a problem with pxc pod livenes during data replication when starting from existing data.
I've got ready data dir for one instance with 40gb of data. I deploy cluster when first instance is ready and ok, it starts another, but during first instance is being clone, another one is starting, becouse readiness lets to think that the second one is read. It makes second one to crash , and starts with not fully cloned data.
I think that problem is liveness:
Probably this makes pod ok,becouse this file exists during data cloning.
Environment
Kubernetes self hosted v1.18.4
Smart Checklist
Activity
Slava Sarzhan September 7, 2020 at 6:27 PM
thank you for your report. The issue was fixed under and will be available in 1.6.0.
There is a problem with pxc pod livenes during data replication when starting from existing data.
I've got ready data dir for one instance with 40gb of data. I deploy cluster when first instance is ready and ok, it starts another, but during first instance is being clone, another one is starting, becouse readiness lets to think that the second one is read.
It makes second one to crash , and starts with not fully cloned data.
I think that problem is liveness:
Probably this makes pod ok,becouse this file exists during data cloning.