I sometimesi have the issue that instances ramped up via econe get automatic_requirements set that contradict sched_requirements and thus the instances stay pending. I have multiple instances that get started and it’s not like an image or template is affected as the same image via the same template works 99% of the time but sometimes my default cluster is not included int the automatic requirements.
The interesting part of onevm show:
SCHED_REQUIREMENTS=“CLUSTER_ID=0 & ! ID=34 & !ID=35”
VIRTUAL MACHINE TEMPLATE
AUTOMATIC_DS_REQUIREMENTS="“CLUSTERS/ID” @> 102 | “CLUSTERS/ID” @> 103 | “CLUSTERS/ID” @> 106"
AUTOMATIC_REQUIREMENTS="(CLUSTER_ID = 102 | CLUSTER_ID = 103 | CLUSTER_ID = 106) & !(PUBLIC_CLOUD = YES)"
You see that the instance is bound to cluster 0 but the automatic requirement does not include cluster 0 so the instance never gets scheduled. On the instances that work (same image, same template) cluster 0 is included and everything is fine. It looks a bit like it’s load related but can’t really put my finger on it yet, it just looks like that happens more often if a lot of instances get requested in a short period of time.
I’m still on 5.14.15 as i did not yet find the time to upgrade (i will as soon as possible) and i’m running a 3 node ha setup. Any ideas would be highly appreciated.