Set shards and replicas based on redundancy policy

[pavolloffay]

https://issues.jboss.org/browse/TRACING-386

Signed-off-by: Pavol Loffay ploffay@redhat.com

[jpkrohling]

This change is 

[codecov]

Codecov Report

Merging #229 into master will increase coverage by 0.12%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #229      +/-   ##
==========================================
+ Coverage   90.35%   90.48%   +0.12%     
==========================================
  Files          59       59              
  Lines        2645     2680      +35     
==========================================
+ Hits         2390     2425      +35     
  Misses        164      164              
  Partials       91       91

Impacted Files	Coverage Δ
pkg/storage/elasticsearch.go	78.43% <100%> (+11.26%)	⬆️

---

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 0337cc9...98fa19d. Read the comment docs.

[jpkrohling]

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

---

pkg/storage/elasticsearch.go, line 55 at r1 (raw file):

			"--es.token-file="+k8sTokenFile,
			"--es.tls.ca="+caPath)
		if !containsPrefix("--es.num-shards", p.Containers[0].Args) {

Could you add a test ensuring that the user's value always has precedence?

[pavolloffay]

I have added more tests and slightly changed ES CR. If node count is higher than 3 (N) it creates CR with 3 master nodes and N-3 client,data nodes.

It's not recommended to deploy more than 3 master nodes in general and ESO even does not allow it:

time="2019-02-25T15:55:43Z" level=error msg="error syncing key (myproject/elasticsearch): Failed to reconcile Elasticsearch deployment spec: Invalid master nodes count. Please ensure there are no more than 3 total nodes with master roles"

[pavolloffay]

At the moment there is a bug in ES operator, it does not create deployment correctly for 3 master and N other nodes https://jira.coreos.com/browse/LOG-340

[pavolloffay]

@jpkrohling could you please re-review and/or merge?

[jpkrohling]

, just fix the one typo and it's good to be merged. Bonus points if you could add the link to the bug preventing more than 3 masters.

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

---

pkg/storage/elasticsearch.go, line 133 at r2 (raw file):

func getNodes(es v1alpha1.ElasticsearchSpec) []esv1alpha1.ElasticsearchNode {
	if es.NodeCount <= 3 {

Could you add a reference to the bug? Is there a public place where the bug can be seen? I tried opening the JIRA, but I don't seem to have access to it O_o

---

pkg/storage/elasticsearch_test.go, line 39 at r2 (raw file):

				RedundancyPolicy: esv1alpha1.FullRedundancy,
				Storage: esv1alpha1.ElasticsearchStorageSpec{
					StorageClassName: "floppydisk",

How do you even know this exists :D

---

pkg/storage/elasticsearch_test.go, line 99 at r2 (raw file):

	tests := []struct {
		pod      *v1.PodSpec
		extected *v1.PodSpec

s/extected/expected ?

[pavolloffay]

How do you even know this exists :D

I heard old dudes talking about it... I thought that I will give it a try on k8s!

Could you add a reference to the bug? Is there a public place where the bug can be seen? I tried opening the JIRA, but I don't seem to have access to it O_o

It's private jira, once it's fixed we don't have to change anything as it is the correct topology. I have just added a comment here to let people know.