BTW, if you have an application that relies upon the "CF_INSTANCE_ADDR", "CF_INSTANCE_IP", or "CF_INSTANCE_PORTS" environment variables for direct communication between apps not in the same space or n

This is not an idle timeout but a max timeout. For my deployment we do disable the timeout by setting a value of 0 relying on our load balancer's idle timeout capabilities and it hasn't caused us any

Thanks Amit, In the past we've typically used the bosh version deployed to PWS as an indication of bosh version that has gone through some real use. I understand the desire to not publish "recommended

We are preparing to release 225 and noticed the release notes don't list a bosh and stemcell version. Does anyone have that info? Mike

Thanks for the update Will. I'll keep waiting patiently. :)

I am interested in better understanding this situation. In CF, when Diego removes an app instance it sends a TERM to the process in I'm assuming this CF behavior is the same for TCP and Http routing?

As an F5 and CF user we wouldn't have an issue changing this setting. I wonder if F5 had a good reason to set this to "Remove" by default? Mike

Yes Will, that summary is essentially correct. But, for even more clarity let me restate the complete story again and reason I want 92085170 to work across stemcell upgrades. :) Today if NATS goes dow

Hi Will, Though I see the main reason for the issue assuming a healthy running environment I've also experienced a deploy related issue that more unique port assignment could help defend against. Duri

Since I cannot comment in tracker I'm starting this thread to discuss story: https://www.pivotaltracker.com/n/projects/1158420/stories/92085170 Some comments I have: * Although I can see how a rolling

That answers my question. Thanks Will. Mike

This is awesome dr. max! Easily the most requested feature in my enterprise. I'm super excited to see this proposal. I'll give it a review next week. Mike wrote:

Thanks for the response Will. You are correct I'm not interested in hacking anything up myself. Just looking for what I should tell my customers. Based on your response there is certainly no way to do

This is a Diego buildpack app. Mike wrote:

Looking at the service broker API it doesn't appear as though it is possible for a service broker to provide tags as part of the provisioning of a service instance. I think it would be nice for a brok

I have some customers who would like to be able to execute a tcpdump on their container interface. I made a quick attempt: /usr/sbin$ ip link show 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue

Sorry, after looking at our manifests it appears we actually set it to 0 not -1. properties: request_timeout_in_seconds: 0 Then you need to set CC timeout to something else since nginx doesn't like 0.

It appears my issue was caused by this uaa issue: https://github.com/cloudfoundry/uaa/issues/223 Now to figure out the best way to work around it. Thanks for your help Matt. Mike

This is what I'm now seeing in the logs: ssh-proxy: {"timestamp":"1446134601.007453442","source":"ssh-proxy","message":"ssh-proxy.authentication-failed","log_level":2,"data":{"error":"no auth passed y

I think I'm getting closer. In UAA I now get the error: TokenEndpoint: Handling error: InvalidGrantException, Invalid authorization code: ad1o9o This must be someone trying to redeem the auth code. Mi