On using legacy tooling in modern HPC systems
By joe- 2 minutes read - 377 words
Or as House MD may have put it
So there you are, working on a system with a group, when you realize that something is out of kilter. And you think to yourself …
It’s not DNS
There is a no way it’s DNS
It was DNS
So your team works on resolving the issue. And the tooling they use … the tooling.
Is from the late 80s/early 90s.
There are so many … better … easier to use … tools. Things that, from an HPC perspective, just work.
Of course, you are under horrific time pressure to get this done. This system is why you’ve not had a day off, or even less than 12 hours per day, everyday for the last month.
The core of the system depends upon tools that are actively hard to configure correctly. Previous version was based upon a completely broken variant (new thing if you can imagine). But a pivot. And now this.
I can’t describe what this is like. Ok, maybe.
So you need to make sure your machines have the needed users on them. And your tool to do this is an ssh loop and some useradd magic.
Its like that.
For dns and dhcp (and tftp boot) at a fairly large scale, if you are doing something other than dnsmasq, you are probably making your life simply too hard. And the cool thing about this you can replicate this, and subnet it if you wish, so you can have many servers if you think you need them (rare that you will).
Absolutely trivial to configure. Autogenerates things you have to manually generate (and check) by hand. Does dns, dhcp (correctly, versus, say, kea), tftp (versus say, tftpd). With minimal effort you can let it tftp an iPXE payload, which then chain boots to an S3 like repository. This is what I do in my home lab with tiburon . Node configuration post boot, indexed by mac or other things.
When I ran Scalable Informatics, this is how our systems worked. And why we could adapt so quickly to new needs. We didn’t waste time with crappy infrastructure/hardware/software. We focused on being able to move, very quickly, correctly.
I want to help … I really want to help.