How etcd data is backed up and restored？

[AllenZMC]

How etcd data is backed up and restored？

[dhellmann]

Today you can make a copy of /var/lib/microshift. That will include etcd, the certificate authorities, and other dynamically generated data. I recommend stopping microshift before making the copy, to ensure that nothing writes to the database while you are backing it up.

[AllenZMC]

@dhellmann Thanks.
BTW, how to restore the backup, copy it directly to the /var/lib/microshift directory?

[AllenZMC]

@dhellmann How many container resources does microshift support, and how many resources does etcd support at most?
Is the storage limit 4G?

[AllenZMC]

How many pods does microshift support?

[dhellmann]

Yes, to restore the backup you would remove /var/lib/microshift and replace it with the backup content.

The storage limit for etcd is 8GB. Using that much storage does have an impact on RAM consumption, though.

We are optimizing for a small footprint, so the scale testing we are doing focuses on making things smaller, not necessarily supporting a large number of resources or pods. We have not imposed any hard limits, but would expect to have a very small number in a typical workload.

Can you tell me more information about your use case for MicroShift?

[AllenZMC]

@dhellmann Because ocp single node supports 250 pods, so I want to know whether microshift also supports 250 pods.

[dhellmann]

The cap is set to 250 for MicroShift as well. As I said, though, we are not focusing on testing up to that scale. If you are running on a host that can run 250 pods, single-node OpenShift is probably a better fit.