Launching agents via SSH is a reasonable approach also for normalizing
configuration of agents across multiple platforms.
My company's build environment uses multiple Linux flavours, a Windows
client (configured using OpenSSH), and a NonStop - all as agents, with a
Controller running under Ubun
For situations where it is reasonable, I agree with Slide’s recommendation: #1
Launch agents via ssh
It’s a good implementation, fairly straightforward, and I think it leads to
better reliability and fewer broken connections. I don’t have any data on that,
though, so I’d also love to hear from
FYI, even though the term is used in some places in the docs and code
(because it's hard to change everything), "agent" is the requested term
instead of "slave". That being said, I prefer the "Launch agents via ssh",
I've never had a problem with reconnection after a reboot of the agent
computer.
Any inputs on this please?
On Tuesday, December 4, 2018 at 9:03:37 AM UTC-5, gotvi...@gmail.com wrote:
>
> What is the best or standard way to configure a Linux slave to the
> Master? These are the options that comes while we try to do the
> configuration
>
> 1. Launch agent agents via ssh
> 2.
What is the best or standard way to configure a Linux slave to the Master?
These are the options that comes while we try to do the configuration
1. Launch agent agents via ssh
2. Launch agent via Java Web Start
3. Launch agent via execution of command on master
4. Let Jenkins control this Window
