recovery scripted failed with following stderr, even though etcd container is already running. 

$ upgrade-recovery.sh --resume

"##### Tue Jan 31 22:40:55 UTC 2023: Waiting for etcd container to restart",
##### Tue Jan 31 23:01:03 UTC 2023: etcd container is not Running. Please investigate"



The error is from this line, where it assumes specific column from crictl cmd. This mechanism is unreliable in general. Suggest to compose the cmd with explicit/specific columns or finding matching column by comparing the header line.
https://github.com/openshift-kni/cluster-group-upgrades-operator/blob/release-4.12/recovery/bindata/upgrade-recovery.sh#L114

4.11.25, 4.12.z

100%

1. Enable ocp upgrade with backup via TALM
2. Before ocp upgrade completes, start recovery via: upgrade-recovery.sh
3. Start second phase of recovery via: upgrade-recovery.sh --resume 

Second phase of recovery fails at finding Running etcd container 

- due to script issue at https://github.com/openshift-kni/cluster-group-upgrades-operator/blob/release-4.12/recovery/bindata/upgrade-recovery.sh#L114

Second phase of recovery passes.

Manually tried the following, where first cmd was from the product code, and second is the working cmd.
[root@cnfde17 core]# name=etcd
[root@cnfde17 core]# crictl ps 2>/dev/null | awk -v name="${name}" '{if ($(NF-2) == name) {print $(NF-3); exit 0}}'
[root@cnfde17 core]# crictl ps 2>/dev/null | awk -v name="${name}" '{if ($(NF-3) == name) {print $(NF-4); exit 0}}'
Running