I was facing the same issue with version 1.11.0 on dual stack kubernetes 1.24.
I found the script /usr/bin/backup.sh in image percona-xtradb-cluster-operator:1.11.0-pxc8.0-backup using $(hostname -i) to retrieve the pod's ip, which would be something like
My workaround is using $(hostname -i | sed 's/.* //') instead and making my own operator backup image.
The fixed
is attached. The only difference with the origin file is the $(hostname -i | sed 's/.* //') part.
Then make a workaround operator backup image with dockerfile like below:
Change the image of the backup section to your own one and the backup would be running smoothly again
When I try to run a (manual) backup my backup job seems to fail to perform the SST.
I can't seem to find out why, but I have one hunch: I have IPv6 enabled in my cluster (dual stack). Maybe something is going wrong there.
I have attached:
My cr.yaml
My backup.yaml
The logs of the backup job
The (hopefully) relevant logs of the PXC pod that was responding to the SST
This is the error from that last file that pointed me towards IPv6:
If any more info is needed, please let me know!
I am using PXC image
8.0.26-16.1
and backup image1.10.0-pxc8.0.25-backup
.