Use tmpfs for etcd in e2e test clusters

[pablochacin]

Description

Add configuration to the kind cluster for setting etcd storage path to a memory-mapped volume.

Fixes #179

Checklist:

• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works.
• I have run linter locally (make lint) and all checks pass.
• I have run tests locally (make test) and all tests pass.
• I have run relevant e2e test locally (make e2e-xxx for agent, disruptors, kubernetes or cluster related changes)
• Any dependent changes have been merged and published in downstream modules
  

[pablochacin]

Initial tests show no significant difference running disruptor e2e tests:

go test -tags e2e e2e/disruptors/pod_e2e_test.go

Before

ok  	command-line-arguments	137.136s

After

ok  	command-line-arguments	133.973s

This can be due to the disk not being a bottleneck in the test machine (which uses SSD)

[nadiamoe]

Even if there's not a big improvement, it seems nicer on the environment (by means of less trashed SSDs :P)
It would be nice to have this by default, specially if we don't foresee etcd to grow big enough to eat the whole RAM.

[nadiamoe]

How can I check if this is working?

I've tried running E2E_ETCD_RAMDISK=1 make e2e-disruptors, but I'm not seeing what I would expect. By describing the etcd pod in the cluster that gets created, I can see:

  volumes:
    - hostPath:
        path: /etc/kubernetes/pki/etcd
        type: DirectoryOrCreate
      name: etcd-certs
    - hostPath:
        path: /var/lib/etcd
        type: DirectoryOrCreate
      name: etcd-data

If I exec into the kindest/node, /var/lib/etcd is not mounted as tmpfs. From the code I guess path for the etcd-data volume should be /tmp/etcd instead. Am I doing something wrong?

[pablochacin]

Am I doing something wrong?

E2E_ETCD_RAMDISK=1 should be E2E_ETCD_RAMDISK=true or be omitted because it is the default.

[pablochacin]

@roobre You were right, I had a bug and wasn't passing the option from the e2e cluster config to the kind cluster config. Good catch.

I fixed it and checked that with the default settings the etcd pod mounts the etcd-data volume at /tmp/etcd and that the tmp directory is mounted on a tmpfs file system in the host node.

[nadiamoe]

👍🏻 I can see the expected volume in etcd now, LGTM. I also like this being on by default (otherwise I would surely forget to turn it on 😅)

Thanks for taking care of it!