After running several swarm nodes for about a week I’m getting these errors in swarm logs:
level=error msg="HTTP error: Error response from daemon: couldn't get vxlan id for \"10.0.52.0/24\": failed to allocate vxlan id: no bit available" status=500
Then no new overlay network can be created.
I observed few unused network and removed them, but that not cured situation, errors start appear 4 hour later and that time no unsued networks that can be deleted remains.
I used to clear KV storage and restart docker on all nodes.
This happened several times, I upgraded etcd, then changed KV from etcd to zookeeper and still getting same errors.
Number of nodes is from 5 to 7, number of used networks is up to 85, so overall limit isn’t hit.
How can I look that resource is used up?
Is there a way I can monitor current usage and limit?
After all that is a ‘bit’ ?