amazon web services - Can ReadWriteOnce PVC be rejected by AWS EBS/GCP PersistentDisk upon Pod Rescheduling? -
assume have 2 pvcs readwriteonce, claimed pod #1 , pod #2 respectively. both running on node #1.
next, pod #2 updated newer docker image. however, @ same time pod #3 run , allocated node #1. since node #1 full, pod #2 allocated kubernetes node #2.
since aws ebs , google persistentdisk mounted on single node, pod #2 become unable connect the-previously-claimed pvc?
if yes, how avoid having issue?
yes, downside of pv/pvc current storage deliveries of aws , gce.
to avoid this, have use different storage infrastructure not have limitation. possibilities ceph, gluster, scaleio (and others). these solutions abstract storage away disks , provide storage layer not node dependent anymore.
Comments
Post a Comment