kklimonda

I mean, Linux package managers are so great that we have at least 2 different ways of delivering software (especially GUI software) to Linux distributions that depends on "app images". To me that shows that none of those approaches are solving 100% of problems that you encounter in the wild.

This means installing software at scale (any number of systems), or the question how to cleanly uninstall software it not a question you should ever ask in a Linux environment.

And yet this is a problem that so many third-party vendors who try to support multiple Linux distributions have been struggling for years.

Tools like gsettings are culturally alien to the unix world.

Sure, Linux and UNIX are coming from different roots, but "cultural" means nothing in large organizations, where computers are basically tools not that far from printers, projectors, even hammers. A way to do someone's job. I may hate locked systems, but then I don't have to support users who cannot find their trash bin on the desktop anymore.

You can seed dotfiles for all users, but you can't really enforce that user cannot for example move his taskbar from bottom to the top of the screen without policy enforcement. gsettings/dconf may be culturally alien to this world, but it is (or at least was) solving an actual problem. A problem we may not care about, but some companies do.

Now, I think there is an interesting discussion here to be had - given this latest push from Windows to Linux, as a way of distancing Europe from US, would adding features that bridge this policy enforcement gap between Linux and Windows is desirable?

15-20 years ago I was going to say yes, but back then I cared so much more about Linux as Windows alternative for office use. Today I actually prefer Linux Wild West and how hard it is to lock it into any sort of MDM.

Isn't WinRM/PowerShell/RDP equivalent of SSH, and dpdk/apt-get is basically .msi with group policies for installation? This has been there for decades probably?

Group Policies also allow you to enforce things like browser configuration (proxy, homepage, search engine etc.) wallpapers, screen locks etc.

Can this be done on Linux? Honestly, I have no idea - I think gnome with gsettings/dconf can do that, but can KDE?

Lixnux version of AD is FreeIPA, with group policies translating to dconf - at least that was the way "enterprise" linux vendors (like RH or Canonical) were moving towards.

Now, how well is dconf integrated with all the software you want to run is another thing (it was done by GNOME, and ignored by KDE), and whether this is still the way they are all moving is yet another question but the infrastructure was being built.

Ceph can scale to pretty large numbers for both storage, writes and reads. I was running 60PB+ cluster few years back and it was still growing when I left the company.

datablocks.dev has a page explaining what white label and recertified disks are [1]. Those are not disks used for years under heavy load.

1: https://datablocks.dev/blogs/news/white-label-vs-recertified...

For me, and probably a lot of other people who moved from other search engines, long-term viability of Kagi is important - heck, that's the reason I've decided it's worth paying some money for search. Given that, I'd expect them to be very frugal with their spendings. Burning money on T-shirts, on another Browser, AI "improvements", Kagi Email (wtf? first time I've heard of it) show that they have incredibly startupy mindset, and will end up like every other company that takes VC money - bloated, money focused and deaf to their community.

You can't drive two LG 5k screens with a single cable, due to it lacking DSC support.

So this is not about forcing Apple to make clients for competing platforms, but to allow businesses to send spam to more users? Well, thanks Google.

Some people enjoy the internet of old, where you had to put some effort into finding venues for collaborating. There is nothing wrong with some places prioritizing building communities over algorithmical reach of the larger platforms.

Immigrants are indeed more willing to do the jobs that domestic workers are no longer interested in doing, but their effect on wages is minimal, mostly affecting lower class, mostly other immigrants. It's unclear if there is anything that can be done to incentivise people to have more children, to the extent that it makes a difference on the macro level. And this is kind of moot point anyway, most western countries that deal with lack of labor force don't have time to wait for children, if their citizens want to keep their level of support.

"The people" are fine with lowering age of retirement, and in general not that interested in rising enough new people to sustain economy and social policies of their countries. Granted, immigration policies in most of the western countries has been a disaster, but those did not arrive out of nowhere. It would be great to see discussions and planning on how to shape policies, but this would only hopefully change the reception of immigrants, and not the fact that they're there to stay.

It's always been like this in GPU space - all reviews have always mentioned number of compute units (be it SMS or "cuda cores"), and the total available for the given architecture is also known. A lot can be told about relative performance of two cards based on that, so this information is useful not only to the investors.

But this seems like a perfectly valid argument to have. I hate that I currently have at least 5 different "usb-like" cable types in my household: usbc to usbc, usbc to lightning, usba to usbc, usba to lightning, usba to microusb, and I probably still have some mini USB cables in some boxes.

I'd really hope that if something new is proposed it will be only once every player have decided to put their weight behind it to propose and push for a change.

How would they do that? Those are shipped directly from Amazon, and don't have any external markings that could be used to link specific card to amazon account. Unless the idea is that vouchers arrive at amazon in some additional packaging and then are repackaged after linking voucher to the account.

By the end of the day I agree, if you have any "real" reason for using VPN you pretty much have to implicitly trust your provider to not keep any traffic flows and connections that could correlate traffic to your IP, but not even sending money in envelope saves your from that.

They also have vouchers you can buy from Amazon, which I find a nice alternative to sending cash in an envelope.

Given that NATO is a defensive alliance, I don't see how such a rule would make much sense. Can you spell that rule out, and link to the official document?

It depends on your bank - mine has an application for phones that is used to interact with 3D Secure, confirm money transfers etc.

Probably all of them? Jakt repo isn't even 3 months old.

My argument, based on this thread alone, would be that Codium will always be... held back by its roots in open core project, with some amazing features (like remote editing) or plugins being unavailable. Codium developers do not "own" their core platform (they can't really fork it without breaking plugins), and they don't have much (if any) say in the future direction of it. Given that it's not unreasonable to think that VS Code is not really that open, and a more desirable outcome would be for developers to support more open alternatives.

vim bindings in Emacs are so great that when something isn't implemented properly (eg. editing macros, global marks) it hurts so much more.

I can't really tell if this is some sort of satire, or a serious post I should actually put some thought into...

Aren't most stores insured, and if so wouldn't this data be available due to insurance claims?

q3k has mentioned metallb+bgp, which is basically in-cluster implementation of LoadBalancer Service type (bgp speakers are running on k8s nodes and announce /32 routes to nodes based on configuration), but it doesn't provide an answer for "stabilizing" ecmp connections when there are changes to backends. There has to be something "behind" metallb[1] that will handle not only stable hashing for connections, but keep forwarding "in-flight" flows (like established tcp sessions) to correct backends, even if packets arrive on different ingress nodes. It seems cilium has some solution for that[2] (by both bundling metallb, and having maglev-based loadbalancer implementation) but I haven't had time to dig into it, so I was curious if someone else has solved it and would be willing to share stories from the front. This is one of those rough edges around kubernetes deployments in bare metal environments and I'd love to see what can be done to make it more robust.

[1] metallb only really announces IPs so that "behind" is probably just CNI that actually handles traffic [2] https://cilium.io/blog/2020/11/10/cilium-19#maglev

How do you keep ecmp hashing stable between rollouts?

Wasn't this the case with Fortnite, which originally was available only as a side loaded app, and now you need a separate store app to install it? Or am I misremembering it?

Just because CentOS Stream has "CentOS" in its name does not mean it's the same thing - CentOS, as is known currently, will effectively stop existing once CentOS7 is no longer supported (not to mention CentOS8 will not even be supported for previously announced period).

Alpine is still linked against musl and not glibc, right? That is a rather large departure, and potentially more problematic than small package repository (usually the smaller repository, the easier it is to package your own stuff from scratch - I assume it beats rpm and deb at that).

RedHat provides developer subscriptions to their products: https://developers.redhat.com/articles/getting-red-hat-devel...

DO has upgraded their Ceph cluster to luminous back in 2018 according to this blog post: https://blog.digitalocean.com/block-storage-volume-performan...

Assuming unlimited budget for hardware and licenses, then building a team that can deploy and manage VMWare on 40 nodes will be much easier - however raw VM is not really comparable to what k8s is giving you, so you'll be also solving reproducible deployments, load balancing traffic to your cluster etc. Still, with enough money those can be solved by purchasing more hardware&software, and you'll have easier time finding people who can maintain that over on-premise kubernetes.