Fix router bug where cookie is not set correctly when dynamic changes are enabled

[ramr]

fixes bugz #1610712 - lil' bit of caveat here mentioned in the bugz.

/cc @openshift/sig-network-edge

[ironcladlou]

@pravisankar PTAL

[pravisankar]

Changes looks good but for the caveat, do you think that will be an issue when client caches cookie derived from blueprint route(pre-reload)?
Is this feasible?

• disable dynamic-cookie backend <blueprint-route>
• set dynamic-cookie-key backend <route-backend> ...
• enable dynamic-cookie backend <route-backend>

[ramr]

@pravisankar so the problem is the cookie name is tied to the backend and we can't change it dynamically(and there's the value part issue too).
When the route is added, a candidate route from the pool of available matching blueprint routes is picked up - let's say the cookie name N1 is basically the one associated with that backend _hapcm_blueprint_<name>_route_<nn>.
Example in bugz: d4bee9152b34dbe72ebd486a984a2809

Now when the router reloads, a new actual backend is added for the route with a different
cookie name say N2.
Example in bugz: 7ffbbf18807d0b625cf2fe2e2fe0ccf4

N1 != N2
And similarly for the value of that cookie -- which is really the hash for the backend server/pod which the session will stick to - that would change as well.

As re: the changes suggested - disabling the dynamic cookies on the blueprint route disable dynamic-cookie backend <blueprint-route> will make the session go to any of the backend server pods (remember the route is associated with a service and a service can have n pods backing it).
So that disables session stickiness while we are using a blueprint pooled route (for whatever the commit interval is set to - default 1 hour). And only enables it past the reload.

So my 2c here were:

1. Whilst we are using a blueprint route - let the session be sticky with one server pod.
2. And once the commit/reload happens, the session will still be sticky (but it could be to another server).
   The good part on 2 is that the stickiness name and value stays forever across any future
   reloads. Could change if a pod dies/ip changes etc but that's as normal.

[pravisankar]

/lgtm

Thanks Ram for the detailed explanation!

[ironcladlou]

/lgtm

[openshift-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: ironcladlou, pravisankar, ramr

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

The pull request process is described here

Details

Needs approval from an approver in each of these files:

• images/router/OWNERS [ironcladlou]
• pkg/router/OWNERS [ironcladlou]

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

[pravisankar]

/retest

[ironcladlou]

/retest

[ramr]

/retest

[ironcladlou]

/retest