Hi,
I have applications that should be running on at least 4 instances (a
webapp running in tomcat).
The reason is availability, it's a fairly common setup, 2 datacenters ,
each datacenter has multiple physical racks.
I am trying to figure out if and how that could be configured with regard
to locations.
My starting point is the below pasted yaml, with the following (test setup
)locations defined in brooklyn.properties:
brooklyn.location.named.dc1rack1=byon:(hosts="10.0.0.16{0-9}")
brooklyn.location.named.dc1rack2=byon:(hosts="10.0.0.17{0-9}")
brooklyn.location.named.dc2rack1=byon:(hosts="10.0.0.18{0-9}")
brooklyn.location.named.dc2rack2=byon:(hosts="10.0.0.19{0-9}")
Firing up the app initially should start one tomcat on each of the
locations, and if load goes up, an extra tomcat should be started on each
location.
Can this be done with one yaml, or do I have to deploy 4 yaml's, one for
each location ?
thanks,
Harry
my yaml:
name: tomcat8Cluster2203
services:
- type: org.apache.brooklyn.entity.webapp.ControlledDynamicWebAppCluster
name: MyTC8Cluster
initialSize: 4
location: dc1rack1,dc1rack2,????
brooklyn.config:
memberSpec:
$brooklyn:entitySpec:
type: org.apache.brooklyn.entity.webapp.tomcat.Tomcat8Server
wars.named:
- https://www.computerhok.nl/tmp/JSPWiki.war
shell.env:
LANG: en_US.UTF-8
jspwiki_pageProvider: VersioningFileProvider
jspwiki_fileSystemProvider_pageDir: jspwiki-pages
jspwiki_basicAttachmentProvider_storageDir: jspwiki-pages
jspwiki_workDir: jspwiki-work
jspwiki_baseURL: http://hardtopredict:8003/JSPWiki
brooklyn.policies:
- policyType: org.apache.brooklyn.policy.autoscaling.AutoScalerPolicy
brooklyn.config:
metric:
$brooklyn:sensor("org.apache.brooklyn.entity.webapp.DynamicWebAppCluster",
"webapp.reqs.perSec.windowed.perNode")
metricLowerBound: 3
metricUpperBound: 5
minPoolSize: 4
maxPoolSize: 16
resizeDownIterationIncrement: 1
resizeUpIterationIncrement: 1
resizeUpStabilizationDelay: 50000
resizeDownStabilizationDelay: 70000
