Hi,
The key was to set the *haproxy.router.openshift.io/disable_cookies
<http://haproxy.router.openshift.io/disable_cookies>* annotation to true.
Regards,
Marvin
On Mon, Dec 16, 2019 at 9:53 AM Just Marvin <
marvin.the.cynical.ro...@gmail.com> wrote:
> Dan,
>
> Thanks for the pointer. It didn't solve my problem. Whats worse, is
> that your link is to the 3.11 docs, and I tried the same annotation on a
> 3.11 install, and it didn't work there either. I have to demonstrate the
> load balancing to a customer, and my cheat was to use the
> *haproxy.router.openshift.io/pod-concurrent-connections
> <http://haproxy.router.openshift.io/pod-concurrent-connections>* annotation
> - when running "ab" with 10 concurrent connections, and
> "pods-conccurrent-connections" set to 6. I do see the spill-over traffic
> going to the second pod.
>
> But this is definitely a cheat for me. I would like to get the
> "balance" annotation, or whatever is the right approach, working. Is there
> a way to troubleshoot this and figure out what is going wrong?
>
> Regards,
> Marvin
>
> On Mon, Dec 16, 2019 at 8:22 AM Dan Mace <dm...@redhat.com> wrote:
>
>>
>>
>> On Mon, Dec 16, 2019 at 8:12 AM Just Marvin <
>> marvin.the.cynical.ro...@gmail.com> wrote:
>>
>>> Hi,
>>>
>>> In the openshift-ingress/router-default pod, in
>>> /var/lib/haproxy/conf/haproxy.config I see:
>>>
>>> backend be_edge_http:session-persistence:song-demo
>>> mode http
>>> option redispatch
>>> option forwardfor
>>>
>>> * balance leastconn*
>>> timeout check 5000ms
>>> http-request set-header X-Forwarded-Host %[req.hdr(host)]
>>> http-request set-header X-Forwarded-Port %[dst_port]
>>> http-request set-header X-Forwarded-Proto http if !{ ssl_fc }
>>> http-request set-header X-Forwarded-Proto https if { ssl_fc }
>>> http-request set-header X-Forwarded-Proto-Version h2 if { ssl_fc_alpn
>>> -i h2 }
>>> # Forwarded header: quote IPv6 addresses and values that may be empty
>>> as per https://tools.ietf.org/html/rfc7239
>>> http-request add-header Forwarded
>>> for=%[src];host=%[req.hdr(host)];proto=%[req.hdr(X-Forwarded-Proto)];proto-version=\"%[req.hdr(X-Forwarded-Proto-Version)]\"
>>> cookie 3e15d2ed54afa750c6c99f3e7974d5f8 insert indirect nocache
>>> httponly secure
>>> server pod:song-demo-6-tw5qr:song-demo:10.128.0.46:9080
>>> 10.128.0.46:9080 cookie 88434433dbc1caefd053e8d5252218f0 weight 256
>>> check inter 5000ms
>>> server pod:song-demo-6-nz8c2:song-demo:10.128.0.47:9080
>>> 10.128.0.47:9080 cookie c67ffb0cd9ed5115a63c2f6b0a39151c weight 256
>>> check inter 5000ms
>>> sh-4.2$
>>>
>>> I probably want "roundrobin" instead of that, though I would argue
>>> the system isn't doing what is implied by "leastconn" either. Is there a
>>> way to change that setting to what I need?
>>>
>>> Regards,
>>> Marvin
>>>
>>
>>
>> Marvin,
>>
>> You can try the "haproxy.router.openshift.io/balance" annotation on your
>> Route resource to configure the balancing strategy[1].
>>
>>
>> https://docs.openshift.com/container-platform/3.11/architecture/networking/routes.html#route-specific-annotations
>>
>>
>> On Sun, Dec 15, 2019 at 8:21 PM Just Marvin <
>>> marvin.the.cynical.ro...@gmail.com> wrote:
>>>
>>>> Hi,
>>>>
>>>> I'm using code-ready containers and I've scaled my pod to do, but
>>>> when I hit it form a for loop in a shell script, I get back results
>>>> indicating that its only being routed to one of the pods. Is there some
>>>> setting that I need to tweak to make things load balanced even at low load?
>>>>
>>>> [zaphod@oc6010654212 session-song-demo]$ oc describe svc song-demo
>>>> Name: song-demo
>>>> Namespace: session-persistence
>>>> Labels: app=song-demo
>>>> app.kubernetes.io/component=song-demo
>>>> app.kubernetes.io/instance=song-demo
>>>> Annotations: openshift.io/generated-by: OpenShiftNewApp
>>>> Selector: deploymentconfig=song-demo
>>>> Type: ClusterIP
>>>> IP: 172.30.140.43
>>>> Port: 9080-tcp 9080/TCP
>>>> TargetPort: 9080/TCP
>>>> Endpoints: 10.128.0.42:9080,10.128.0.43:9080
>>>> Port: 9443-tcp 9443/TCP
>>>> TargetPort: 9443/TCP
>>>> Endpoints: 10.128.0.42:9443,10.128.0.43:9443
>>>> Session Affinity: None
>>>> Events: <none>
>>>> [zaphod@oc6010654212 session-song-demo]$ oc describe route song-demo
>>>> Name: song-demo
>>>> Namespace: session-persistence
>>>> Created: 5 hours ago
>>>> Labels: app=song-demo
>>>> app.kubernetes.io/component=song-demo
>>>> app.kubernetes.io/instance=song-demo
>>>> Annotations: <none>
>>>> Requested Host: song-demo.apps-crc.testing
>>>> exposed on router default (host apps-crc.testing) 5 hours ago
>>>> Path: <none>
>>>> TLS Termination: edge
>>>> Insecure Policy: <none>
>>>> Endpoint Port: 9080-tcp
>>>>
>>>> Service: song-demo
>>>> Weight: 100 (100%)
>>>> Endpoints: 10.128.0.42:9443, 10.128.0.43:9443, 10.128.0.42:9080 + 1
>>>> more...
>>>> [zaphod@oc6010654212 session-song-demo]$
>>>>
>>>> Regards,
>>>> Marvin
>>>>
>>> _______________________________________________
>>> users mailing list
>>> users@lists.openshift.redhat.com
>>> http://lists.openshift.redhat.com/openshiftmm/listinfo/users
>>>
>>
>>
>> --
>>
>> Dan Mace
>>
>> Principal Software Engineer, OpenShift
>>
>> Red Hat
>>
>> dm...@redhat.com
>>
>>
>>
_______________________________________________
users mailing list
users@lists.openshift.redhat.com
http://lists.openshift.redhat.com/openshiftmm/listinfo/users
