MODCLUSTER-590 - workers array for Deterministic failover is allo…

[Karm]

…cated dynamically now

Before patch

Doesn't even compile.

mod_proxy_cluster.c(2203): error C2057: expected constant expression
mod_proxy_cluster.c(2203): error C2466: cannot allocate an array of constant size 0
mod_proxy_cluster.c(2203): error C2133: 'workers': unknown size

After patch

[Thu Jun 15 20:17:16.181198 2017] [mpm_winnt:notice] [pid 4732:tid 624] AH00455: Apache/2.4.23 (Win64) mod_cluster/2.0.0.Alpha1-SNAPSHOT OpenSSL/1.0.2h configured -- resuming normal operations
<SNIP>
[pid 7548:tid 1908] mod_proxy_cluster.c(2919): cluster: balancer://qacluster Found value JuJ8S_O87OvyrLtC4GM2YUSPAnwB3Jl0R9Hn3mFp.jboss-eap-7.1-3 for stickysession JSESSIONID|jsessionid
[pid 7548:tid 1908] mod_proxy_cluster.c(1057): update_workers_node starting
[pid 7548:tid 1908] mod_proxy_cluster.c(334): Created: reusing worker for ajp://172.16.27.132:8312
[pid 7548:tid 1908] mod_proxy_cluster.c(334): Created: reusing worker for ajp://172.16.27.132:8312
[pid 7548:tid 1908] mod_proxy_cluster.c(1069): update_workers_node done
[pid 7548:tid 1908] mod_proxy_cluster.c(2336): Using deterministic failover target: jboss-eap-7.1-4
<SNIP>

JIRA: MODCLUSTER-590

[Karm]

@pferraro, if you find it sane, I'll back-port.

Is this balancer->workers->pool the right pool to use? I guess so as it lives as long as workers...

Windows CI coming up on https://ci.modcluster.io/ any day now should be catching these things easy.