@sdodson, I found the 'etcdctl2'&'etcdctl3' still exists in 3.10, but 'etcdctl' doesn't work in 3.10, but in 3.11, it work as expected that 'etcdctl2&etcdctl3' deprecated, and etcdctl works well. my 3.10 version: openshift v3.10.57 kubernetes v1.10.0+b81c8f8
(In reply to ge liu from comment #1) > @sdodson, I found the 'etcdctl2'&'etcdctl3' still exists in 3.10, but > 'etcdctl' doesn't work in 3.10, but in 3.11, it work as expected that > 'etcdctl2&etcdctl3' deprecated, and etcdctl works well. > > my 3.10 version: > openshift v3.10.57 > kubernetes v1.10.0+b81c8f8 The two bash functions are etcdctl2 and etcdctl3 which specify flags for etcd v2 and v3 which are different. If those two commands work then that seems like what I'd expect. We've never deployed an alias for `etcdctl`.
@sdodson, It still ambiguous to me, and what i know is 'etcdctl' works in 3.11 at master server, but doesn't work in 310's master server, and this will bring some problems like below: Etcd backup/restore: we may use etcdctl3 to do etcd bakcup, then we remove etcd.yaml and rm -rf '/var/lib/etcd', then we try to do etcd restore, but etcdctl3 command is not work, so we can only reinstall etcdctl command tools to do etcd restore. and compare to 3.11, we do etcd backup with etcdctl command, then we remove etcd.yaml and rm -rf '/var/lib/etcd', and etcdctl command still work for etcd restore. so the difference between 310 and 311 will bring problems to us.
https://github.com/openshift/openshift-ansible/pull/9757 3.10 backport Fixed in openshift-ansible-3.10.36-1
Fail to verify on: oc v3.10.79 kubernetes v1.10.0+b81c8f8 features: Basic-Auth GSSAPI Kerberos SPNEGO # etcdctl bash: etcdctl: command not found
Only etcdctl2 and etcdctl3 bash functions are expected to work. These will execute insider the pod.
@sdodson, regarding to Comment 5, if etcdctl only work inside pod, the etcd restore process could not be executed because etcd pod be destroyed. Is it work as design? [root@qe-geliu310master-etcd-1 master]# etcdctl3 Component etcd is stopped or not running [root@qe-geliu310master-etcd-1 master]# etcdctl bash: etcdctl: command not found [root@qe-geliu310master-etcd-1 master]#
I've not kept up with the latest on the restore process. These commands will definitely fail if the etcd pod is not running or otherwise crash looping. I'm not sure what we can do about that.
As my personal feeling, 3.11's design is make sense, could it possible do same with 3.11?
According to comment 6, close this bug.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2019:0026