Fix bugs in Lua implementation of sticky sessions

[ElvinEfendi]

What this PR does / why we need it:
I've detected two bugs with current implementation of sticky sessions in Lua:

• it does not re-set the cookie after setting it for the first time this should be the case - just make it a session cookie
• it uses shared lua dictionary which is per replica - so as soon as next request from same cookie goes to another replica it will potentially be proxied to a different upstream

The PR adds regression test and fix the bugs.

Which issue this PR fixes (optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close that issue when PR gets merged): fixes #

Special notes for your reviewer:

[andrewloux]

It seems like the only reason this or is there is to not repeat fetching the endpoint (L57 - L60).

You will have some repeated code here - but I would separate the conditions (is_sticky(backend) and backend["upstream-hash-by"]) to keep it easy to read. This way you could also take out the condition on L62.

[andrewloux]

Seems more appropriate to default to route higher up no?https://github.com/kubernetes/ingress-nginx/blob/master/internal/ingress/controller/controller.go#L501

[andrewloux]

Would be nice to have a helper method here that generates the digested endpoint string rather than doing it over here.

[codecov-io]

Codecov Report

Merging #2484 into master will decrease coverage by 0.09%.
The diff coverage is n/a.

@@            Coverage Diff            @@
##           master    #2484     +/-   ##
=========================================
- Coverage   41.66%   41.57%   -0.1%     
=========================================
  Files          74       74             
  Lines        5282     5270     -12     
=========================================
- Hits         2201     2191     -10     
+ Misses       2783     2782      -1     
+ Partials      298      297      -1

Impacted Files	Coverage Δ
internal/file/bindata.go	61.34% <ø> (-1.06%)	⬇️

---

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 94198fc...7ac4e1d. Read the comment docs.

[ElvinEfendi]

CI is failing because if a flaky test that's fixed at #2526. I'll rebase once that gets merged.

[ElvinEfendi]

/assign @aledbf

[aledbf]

/approve
/lgtm

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: aledbf, ElvinEfendi

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [aledbf]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment