On 2/24/2021 12:42 PM, Sam Doran wrote:
Jimi actually did this a while ago <https://github.com/jimi-c/ansigo> to see if there were any benefits.
I wasn't aware of this. Thanks for the reference.
In the end, there was no real benefit to moving to Go since most of the time the control node is waiting for code to execute and return results from the managed nodes.
I'd like to learn how complicated the playbooks and host files were that he used for the experiment. As I said, in small environments I wouldn't expect much of an improvement. But, I once worked at a place that had ~1,000 hosts and a very complicated group structure where I think a compiled language would have made a difference on the controller.
T here are probably other details I'm forgetting since this was a while ago. I'll let Jimi comment further.
I'm looking forward to this. One other thing that's true now is the the Go linker has gotten ~30% faster.
It's possible to use Go modules with Ansible today since we support binary modules. But the problems with that were already outlined.
I hadn't known about binary modules (it's been a while since I used Ansible). That's an interesting idea. Given Go's ability as a cross compiler, I could see the controller compiling and linking modules in the proper format for a particular client. I suspect, though, that binary modules would be much larger than the amount of Go source code that would have to be transferred. Whether that's important is another question. Thanks for all the responses. Jon -- You received this message because you are subscribed to the Google Groups "Ansible Project" group. To unsubscribe from this group and stop receiving emails from it, send an email to ansible-project+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/ansible-project/s16fqp%24j73%241%40ciao.gmane.io.
