Schema of the architecture of pushpin ?

[inkhey]

Hello, i use pushpin for the https://github.com/tracim/tracim project.

As we get multiple log un pushpin related to the multiples components of it. It's a bit complicated to clearly understand what it mean. A schema giving explaination of the process (the goal of conduire, zurl, mongrel2, pusphin-handler, etc...), will be useful.

Is there already something like this ?

Thanks for pushpin, it's an awesome software !

[jkarneges]

Hmm, there isn't any documentation for this. I suppose it would be good to add.

In the meantime, here's an overview:

• condure: handles incoming connections
• zurl: handles outgoing connections
• pushpin-proxy: handles proxying logic
• pushpin-handler: handles pubsub/grip logic

Earlier versions used mongrel2+m2adapter for handling incoming connections, but this is deprecated in recent versions.

When looking at logs for debugging, the proxy logs are usually the best place to start, to see how a connection was accepted/routed. The handler logs are useful for knowing what channels are being subscribed and published to. The condure/zurl logs are useful for investigating connection issues.

By the way, it can be convenient to run pushpin with the --merge-output option. This causes all the logs to be written to stdout instead of files, with each log line prefixed with the name of the sending process. That way there is only one log to look at.