HNHacker News
TopNewBestAskShowJobs

bfgpereira

16 karma · joined June 20, 2019

submissionscomments
bfgpereira··on TOP500 at ISC’26: We have a New Number 1 Supercomputer
The TOP500 rank is not a count of nodes, GPUs, or aggregate FLOPS.

Many systems have the node count to be able to run such benchmarks, but are not optimized or even capable of running them. Having these systems run these large calculations in a sustained way, and performing, is what separates a bunch of nodes together in a data center from an actual cluster that is able to run a benchmark like HPL or HPCG.

To sustain 70 - 80% of peak performance across hundreds of thousands of cores, you need a real low-diameter, high-bandwidth, low-latency fabric and a balanced memory subsystem, running on a system w almost no failures or network issues. A loosely-coupled cluster with an oversubscribed fat-tree will 'run Linpack' and then post an Rmax that's a small fraction of its naive peak.

Also, have a look at the Green500, and the systems there. This is not about bragging rights vendors, this is about placing commodity hardware, tuning it, and bringing it up to health in a way that squeezes all of that last performance possible out of the clusters on those lists. That's the opposite of vendor flexing - it's a craft that you cannot see in a simple node count, as some have been comparing here.

If you ever worked on this field, and with the vendors at this scale, you would know. Its not easy, its actually very hard.

... and imagine you need to deploy this, systems at this scale, w technologies that are sometimes just emerging and sometimes even need proper field testing * every 6 months * to be able to reach the scale and stability to land on these lists.

bfgpereira··on Pwning your web server the easy way or why exposing –/.ssh/ is a bad idea
You describe precisely one of my points there. Its incredible how comfortable people are touching components that they don't fully understand. :D
bfgpereira··on Pwning your web server the easy way or why exposing –/.ssh/ is a bad idea
> Only time will tell if I am in fact right.

I guess you prove a point here: you trust that the deployment of your GKE module allows you to be safe, so your investment vs risk trade seems to satisfy you. But you, yourself, cannot even predict how insecure you are at the moment due to the complexity of the software solutions you are using.

> I suppose we'll see if companies with sysadmins have more breaches than the guys who run their own ops using container orchestration etc.

Oh, I do agree, at least with the current state of sysadmins out there.

But the problem is that you loose control of the integrity of your system once you reach a point where software complexity becomes your security entry point.

If you are sure the containers you will be using are secure, have sane defaults, are up to date, etc, then fine, good job! I just don't trust that most people will be able to reassure me that. And please keep in mind that in your case, the target is not your container platform, the target would be the containers in that platform, and the services they run.

bfgpereira··on Pwning your web server the easy way or why exposing –/.ssh/ is a bad idea
I am agreeing with what you wrote, though. Or at least I am trying to.

I have seen numerous times the results of that, where for instance, a developer creating a tool decides that his interpretation of a bad requirement is satisfied in a poor way. Or a sysadmin deciding that a default configuration is good enough because he did `mv conf.example to conf`, and it works.

I guess what I am trying to say is that the learning curve given the complexity of software / systems now a days, and the lack of judgement and training by the users / developers / sysadmins of those systems results in decisions where risk is not taken in to account.

I miss the time where people genuinely knew what they were doing, and had a mindset that allowed them to avoid / prevent risk in the decisions they take during their daily tasks.

bfgpereira··on Pwning your web server the easy way or why exposing –/.ssh/ is a bad idea
This is what I have learned in many years of work: people who know systems should be let to handle those systems.

This is what happens when a developer is left to do the work that a system administrator should be trained to do - not all are -.

For a developer, in most cases, "just works" is the end goal, when referring to systems. Not "how it works", and what are the implications of making it work like this.

This really makes me sad.

bfgpereira··on Ramroot – Run Arch Linux Entirely from RAM (2017)
This is nothing new, or exciting. Most Linux distros will do this either via PXE, with NFS root, or rsync a rootfs to RAM and then boot, etc etc etc. There are literally so many ways of doing this, it would take me eons to stop the creativity kill.

What do so something cool? Boot a whole computer cluster using BitTorrent as a backend, diskless, diskfull, at lightning speeds: https://github.com/dchirikov/luna