i have setup deployment (dep1) in k8s , there's persistent volume associated with access mode readwriteonce. have horizontalpodautoscale resource dep1 auto-scales deployment if cpu consumption above threshold. happen if deployments deployed across multiple nodes? auto-scaling work? in general, there concern regarding autoscaling when using pvs in readwriteonce mode?
i didn't try myself yet guess depends on type of volume work with. if work ebs pv, mounts across nodes @ same time aren't possible , pods fail if scheduled on different nodes. if use nfs type of volume on other hand, mounts on multiple nodes possible.
that being said, think have overthink access mode policy. readwriteonce per definition allows mount single node. think in scenario need change readwritemany.
Comments
Post a Comment