Wait for elasticsearch cluster to be up

[pavolloffay]

Resolves #216

Signed-off-by: Pavol Loffay [email protected]

[jpkrohling]

This change is 

[pavolloffay]

There is one minor is deployment.go deletes ES deployment once it's up, not sure why I will have to dig into it

[codecov]

Codecov Report

Merging #242 into master will decrease coverage by 0.28%.
The diff coverage is 61.29%.

@@            Coverage Diff            @@
##           master    #242      +/-   ##
=========================================
- Coverage   89.69%   89.4%   -0.29%     
=========================================
  Files          62      62              
  Lines        2717    2738      +21     
=========================================
+ Hits         2437    2448      +11     
- Misses        181     187       +6     
- Partials       99     103       +4

Impacted Files	Coverage Δ
pkg/controller/jaeger/jaeger_controller.go	33.33% <0%> (ø)	⬆️
pkg/storage/elasticsearch.go	79.81% <100%> (-0.19%)	⬇️
pkg/controller/jaeger/elasticsearch.go	59.09% <58.33%> (-4.55%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9302c08...2906cb6. Read the comment docs.

[jpkrohling]

Reviewed 2 of 3 files at r1, 2 of 2 files at r2.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @pavolloffay)

a discussion (no related file):

One side-effect to this PR is that the images for Jaeger components no longer happen in parallel to the bootstrap of Elasticsearch. It's not a big deal, as it makes a difference only for the first time each image is used, but it might be worth opening an issue to investigate if we can pull the images in advance, before we declare them in a deployment.

---

---

pkg/controller/jaeger/elasticsearch.go, line 71 at r2 (raw file):

			}
		}
		return available == expectedSize, nil

Could you add a Debug() message with the current state? When debugging issues, I found it helpful to know that it's waiting for (N) nodes to become ready. We have something like this in the deployment:

		if d.Status.ReadyReplicas != d.Status.Replicas {
			log.WithFields(log.Fields{
				"namespace": dep.Namespace,
				"name":      dep.Name,
				"ready":     d.Status.ReadyReplicas,
				"desired":   d.Status.Replicas,
			}).Debug("Waiting for deployment to estabilize")
			return false, nil
		}

---

pkg/storage/elasticsearch.go, line 114 at r2 (raw file):

				"app.kubernetes.io/part-of":   "jaeger",
				// we cannot use jaeger-operator because our component controllers removes all objects
				// created by ES operator

Do you mean that our Operator removes the Elasticsearch stateful sets?

[pavolloffay]

Do you mean that our Operator removes the Elasticsearch stateful sets?

There are no SS but yes. These objects would be queried by our controllers - which should not be bc they are managed by ES operator

[jpkrohling]

Across the code base, we are now using log as alias to this package.

[jpkrohling]

s/logrus/log