hello, is anybody else experiencing long restart of rke2 server version 1.35.5 or later? Till 1.34, it was expectably short, but now it seems to take significantly longer to start. also the upgrade 1.34 -> 1.35 took about 5 minutes to restart the rke2 server (I do not think because of images pull).

I have not yet upgraded to 1.35 am curious, if you see anything here:

journalctl -u rke2-server.service

while the service is starting up

What specifically is taking so long?

maybe this one?

Jun 16 22:57:06 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:57:06+02:00" level=info msg="ETCD server is now running"
Jun 16 22:57:06 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:57:06+02:00" level=info msg="Saving cluster bootstrap data to datastore"
...
Jun 16 22:57:20 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:57:20+02:00" level=info msg="Pod for etcd not synced (waiting for termination of old pod sandbox), retrying"
Jun 16 22:57:20 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:57:20+02:00" level=info msg="Pod for etcd not synced (waiting for termination of old pod sandbox), retrying"
Jun 16 22:57:40 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:57:40+02:00" level=info msg="Pod for etcd not synced (waiting for termination of old pod sandbox), retrying"
Jun 16 22:57:40 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:57:40+02:00" level=info msg="Pod for etcd not synced (waiting for termination of old pod sandbox), retrying"
Jun 16 22:57:40 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:57:40+02:00" level=error msg="Failed to check local etcd status for learner management: rpc error: code = Unavailable desc = connection error: desc = \"transport: Error while dialing: dial tcp 127.0.0.1:2379: connect: connection refused\""
Jun 16 22:57:40 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:57:40+02:00" level=error msg="Failed to check local etcd status for learner management: rpc error: code = Unavailable desc = connection error: desc = \"transport: Error while dialing: dial tcp 127.0.0.1:2379: connect: connection refused\""
Jun 16 22:58:00 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:58:00+02:00" level=info msg="Pod for etcd not synced (waiting for termination of old pod sandbox), retrying"
Jun 16 22:58:00 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:58:00+02:00" level=info msg="Pod for etcd not synced (waiting for termination of old pod sandbox), retrying"
Jun 16 22:58:20 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:58:20+02:00" level=info msg="Pod for etcd not synced (waiting for termination of old pod sandbox), retrying"
Jun 16 22:58:20 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:58:20+02:00" level=info msg="Pod for etcd not synced (waiting for termination of old pod sandbox), retrying"
Jun 16 22:58:40 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:58:40+02:00" level=info msg="Pod for etcd not synced (waiting for termination of old pod sandbox), retrying"
Jun 16 22:58:40 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:58:40+02:00" level=info msg="Pod for etcd not synced (waiting for termination of old pod sandbox), retrying"
Jun 16 22:59:00 <http://kub-as1.priv.cerit-sc.cz|kub-as1.priv.cerit-sc.cz> rke2[3390]: time="2026-06-16T22:59:00+02:00" level=info msg="Pod for etcd is synced"

so that the etcd is starting for some time. and maybe sometimes even longer?

It's not that it's not starting, it's that the old one is taking forever to stop. Are you restarting all your nodes at once? What do the etcd pod logs say?

one by one