6 ms·
systemd-networkd DHCP performance
- Fasebook 13y agothankfully we have systemd to manage these kinds of module dependency startup problems in linux now.
- wmf 13y agoRelated classic blog post: how do Macs get on the network so fast? http://cafbit.com/entry/rapid_dhcp_or_how_do http://cafbit.com/entry/rapid_dhcp_or_how_do
- JoshTriplett 13y agoIt's nice to see this specifically treated as an optimization goal. This is one of those cases (much like git for version control) where sufficiently good performance enables new uses that would not otherwise be possible. For instance, consider getting a decent network connection over an intermittent link.
- zokier 13y agoIt is bit odd that network management is one of those things that gets iterated on fairly often. What makes this something that has difficulties converging? It seems like almost every distro has it's own solution, or several of them.
- superbaconman 13y agoNetworks are distributed systems, thus a pain in the ass. SDN will hopefully save us from a lot of existing pain, but it'll be few years (plus a few failed network companies) before things get easier.
- asabjorn 13y agoFortunately this network management solution is now part of systemd, and systemd has been adopted as default by all major distributions except gentoo.
- simula67 13y agoFor those who missed it, Ubuntu is also moving to systemd as default : http://www.markshuttleworth.com/archives/1316 http://www.markshuttleworth.com/archives/1316
- josteink 13y agoConsider me one of those. Here I thought Ubuntu would be starting a conflict with its Debian-based roots and had me worried about how it would all play out. In short: Thanks for posting.
- dredmorbius 13y agoI was really hoping Shuttleworth would hold out on this. I've got serious concerns about systemd still. Ted Ts'o's recent G+ post, while generally supportive, gets into many of the issues involved: A realization that I recently came to while discussing the whole systemd controversy with some friends at the Collab Summit is that a lot of the fear and uncertainty over systemd may not be so much about systemd, but the fear and loathing over radical changes that have been coming down the pike over the past few years, many of which have been not well documented, and worse, had some truly catastrophic design flaws that were extremely hard to fix. Quite. https://plus.google.com/117091380454742934025/posts/4W6rrMMvhWU https://plus.google.com/117091380454742934025/posts/4W6rrMMv...
- groovy2shoes 13y agoThankfully Slackware is still running sysvinit.
- zanny 13y agoAnd even Gentoo has reasonable systemd support if you want to use it, and Sabayon (the largest Gentoo derivative) uses systemd.
- dsr_ 13y agoThe most interesting bit, to me, is actually in the comments. You don't have to recognize Ted Lemon's name, or know of his work, in order to read the tag that G+ put next to his query. But even if you completely ignore that and treat him as a random person, the LMGTFY response was rather rude. Is this endemic to the systemd project? Because I keep seeing similar rudeness whenever I look at their mailing lists.
- scrollaway 13y agoIt's endemic to a lot of open source projects. But yes, it's very present around systemd/gnome. I'm not a huge gnome fan but I love systemd and I hugely respect both projects in many respects. I've met several of the people in both projects as well and I like some of them (not all, but that's not to be expected). The attitude though, oh god the attitude. In those very comments, Ted Lemon put it extremely well into words: "I've worked with people who respond to questions this way. It makes for a stressful work environment—you're always wondering whether they're going to try to score points off you when you ask a question." I don't know how to qualify this attitude as. I'd say "holier than thou" but that kind of lost its meaning along the way. At their core, most of the people working on these projects expect near-perfection from others and are not willing to assist those they demand perfection from. Anyone who fails to achieve their demands simply gets dismissed. Very few of the people who are interested in the project end up contributing because of the toxic attitude... it's all some sort of project-wide social filter bubble. This seemingly has the effect of getting brilliant people interested in those projects, but just as many fall through the cracks. I've not thought about all of this long enough to talk about the long term consequences but what I do see is that while GNOME may be a great feat of desktop engineering, it's meaningless if everyone despises the developers (and consequentially, the name). Funny, too, UX people don't really fit into this monoculture. This really breaks my heart. I see it as a form of bullying... You either help, or you don't help (don't reply). You don't go around making people feel like they are lesser men because they don't have the knowledge you nurtured over years or even decades. I bet Mozart sucked with computers, too. PS: I apologise for going off-topic. This work on dhcp performance is really damn awesome and those issues have nothing to do with the topic at hand.
- 13y ago
- dscrd 13y agoPiece by piece, Linux is becoming an OS that we don't constantly have to defend for its small deficiencies. Loving it.