Microk8s: snap refresh microk8s fails (from v1.18.2)

[tomtor]

Ubuntu 18.04, currently on:

installed: v1.18.2 (1378) 201MB classic

snap refresh microk8s fails with:

+ for dir in "${SNAP_DATA}/credentials/ ${SNAP_DATA}/certs/ ${SNAP_DATA}/args/ ${SNAP_DATA}/[0/1836]
"
+ chmod -R ug+rwX /var/snap/microk8s/1443/credentials/ /var/snap/microk8s/1443/certs/ /var/snap/micr
ok8s/1443/args/ /var/snap/microk8s/1443/var/lock
+ chmod -R o-rwX /var/snap/microk8s/1443/credentials/ /var/snap/microk8s/1443/certs/ /var/snap/micro
k8s/1443/args/ /var/snap/microk8s/1443/var/lock
+ getent group microk8s
+ getent group microk8s
+ chgrp microk8s -R /var/snap/microk8s/1443/credentials/ /var/snap/microk8s/1443/certs/ /var/snap/mi
crok8s/1443/args/ /var/snap/microk8s/1443/var/lock/
+ '[' -f /var/snap/microk8s/1443/args/flanneld ']'
+ grep -e etcd.socket:2379 /var/snap/microk8s/1443/args/etcd
+ grep '\-\-enable\-v2' /var/snap/microk8s/1443/args/etcd
--enable-v2=true
+ false
+ '[' -L /var/snap/microk8s/1443/bin/cilium ']'
+ '[' -e /var/snap/microk8s/1443/args/containerd-template.toml ']'
+ grep -e 'plugins.cri.registry.mirrors."local.insecure-registry.io"' /var/snap/microk8s/1443/args/c
ontainerd-template.toml
+ '[' -e /var/snap/microk8s/1443/var/lock/clustered.lock ']'
+ '[' -e /var/snap/microk8s/1443/args/flannel-template.conflist ']'
+ grep -e cniVersion /var/snap/microk8s/1443/args/flannel-template.conflist
    "cniVersion": "0.3.1",
+ grep -e '\-\-timeout' /var/snap/microk8s/1443/args/cluster-agent
grep: /var/snap/microk8s/1443/args/cluster-agent: No such file or directory
+ refresh_opt_in_config timeout 240 cluster-agent
+ local opt=--timeout
+ local value=240
+ local config_file=/var/snap/microk8s/1443/args/cluster-agent
+ local replace_line=--timeout=240
++ grep -qE '^--timeout=' /var/snap/microk8s/1443/args/cluster-agent
grep: /var/snap/microk8s/1443/args/cluster-agent: No such file or directory
+ run_with_sudo /snap/microk8s/1443/bin/sed -i '$ a --timeout=240' /var/snap/microk8s/1443/args/clus
ter-agent
+ '[' -n '' ']'
+ '[' /snap/microk8s/1443/bin/sed == preserve_env ']'
+ sudo /snap/microk8s/1443/bin/sed -i '$ a --timeout=240' /var/snap/microk8s/1443/args/cluster-agent
/snap/microk8s/1443/bin/sed: can't read /var/snap/microk8s/1443/args/cluster-agent: No such file or
directory
-----)

[tomtor]

EDIT:
Is the problem that 1378 (currently installed snap) does not have a cluster-agent file in the args directory?

tom@Tom-k8s:~$ ls /var/snap/microk8s/1378/args
cni-network               ctr                         kube-apiserver           kubectl-env
containerd                etcd                        kube-controller-manager  kubelet
containerd-env            flannel-network-mgr-config  kube-proxy
containerd-template.toml  flannel-template.conflist   kube-scheduler
containerd.toml           flanneld                    kubectl

[balchua]

I think there an issue with the configure hook. Is the cluster originally installed with 1.18 or its been being upgraded since then.

[tomtor]

@balchua No, it's an old cluster, at least 12 months old. All previous automatic updates went smoothly.

[balchua]

That's what i suspect. Looks like an easy fix.
But if you are in a hurry, you can copy the /snap/microk8s/current/default-args/cluster-agent to /var/snap/microk8s/current/args/.
Then try to do refresh.

[tomtor]

@balchua Thanks!

cp /snap/microk8s/current/default-args/cluster-agent /var/snap/microk8s/current/args/
snap refresh microk8s

worked!

I assume many users have this issue. I only noticed it because I have a watchdog monitoring the cluster and it goes down a few minutes every few hours (due to a failed update) since the last 3 days:

2020-06-11 13:35:02.004615+00:00 0:02:00.390599
2020-06-11 09:49:01.960818+00:00 0:03:00.285431
2020-06-11 09:23:01.199726+00:00 0:02:59.145173
2020-06-11 09:05:02.067254+00:00 0:02:00.187082
2020-06-11 08:58:01.129220+00:00 0:10:59.307220
2020-06-11 07:20:01.438388+00:00 0:01:59.825661
2020-06-11 00:39:01.975900+00:00 0:02:00.457308
2020-06-10 22:40:01.123141+00:00 0:01:59.433738
2020-06-10 17:00:02.066083+00:00 0:03:00.988085
2020-06-10 10:56:01.992499+00:00 0:03:00.589398
2020-06-10 02:30:01.643774+00:00 0:03:00.056305
2020-06-09 19:01:01.383671+00:00 0:02:59.675328
2020-06-09 15:05:01.197224+00:00 0:03:00.047818
2020-06-09 11:15:01.543906+00:00 0:02:59.555052
2020-04-28 01:36:01.764504+00:00 0:02:59.101809

[balchua]

Thanks @tomtor for reporting this. Will probably do a PR soon.

[balchua]

The PR is merged it should be available in latest/edge within the day.