Closed cjabrantes closed 7 months ago
Hi all,
I was testing your operator that really seems to make our life easier (thanks for that). I have some doubts.
I m testing a setup where i have 3 box, with 3 sentinels and 2 redis. In my redisfailover.yaml i have
podAntiAffinity: requiredDuringSchedulingIgnoredDuringExecution: - labelSelector: matchExpressions: - key: app.kubernetes.io/component operator: In values: - sentinel topologyKey: kubernetes.io/hostname
To avoid having the sentinel is the same node.
The problem is when i try to make some changes the upgrade gets stuck:
rfs-redis-69c47c54fc-6kbv4 1/1 Running 0 24m 10.42.4.42 k3s-agent-1 rfs-redis-69c47c54fc-bbh4q 1/1 Running 0 23m 10.42.5.129 k3s-agent-2 rfs-redis-69c47c54fc-xw4kn 0/1 Pending 0 23m rfs-redis-6f6d45449c-42x8p 1/1 Running 0 5d23h 10.42.3.237 k3s-agent-3
I believe the reason for this is maxSurge default of 25%, meaning that when the new pod is scheduled there is no place for it to run.
I was not able to find in the CRD a way to configured the maxSurge, is it possible? or any workaround?
Thanks,
This issue is stale because it has been open for 45 days with no activity.
This issue was closed because it has been inactive for 14 days since being marked as stale.
Hi all,
I was testing your operator that really seems to make our life easier (thanks for that). I have some doubts.
I m testing a setup where i have 3 box, with 3 sentinels and 2 redis. In my redisfailover.yaml i have
To avoid having the sentinel is the same node.
The problem is when i try to make some changes the upgrade gets stuck:
rfs-redis-69c47c54fc-6kbv4 1/1 Running 0 24m 10.42.4.42 k3s-agent-1
rfs-redis-69c47c54fc-bbh4q 1/1 Running 0 23m 10.42.5.129 k3s-agent-2
rfs-redis-69c47c54fc-xw4kn 0/1 Pending 0 23m
rfs-redis-6f6d45449c-42x8p 1/1 Running 0 5d23h 10.42.3.237 k3s-agent-3
I believe the reason for this is maxSurge default of 25%, meaning that when the new pod is scheduled there is no place for it to run.
I was not able to find in the CRD a way to configured the maxSurge, is it possible? or any workaround?
Thanks,