[03:09] good morning [03:29] "good morning" <- 👋 hi there, good morning [03:29] heya ilvipero o/ [11:18] yay lotuspsychje! [11:57] I've been using ubuntu for work for about 10+ years in large environments (thousands of VMs) and have historically been an advocate of it over more traditional subscription-based enterprise distributions [11:57] have a couple of pieces of feedback: [11:57] the installer for 22 is so bad that for our 22 VMs we've had to upgrade a 20 machine and use that as a template - there is no way to create a single partition install without swap or boot [11:57] the increasing reliance on snap is something that might cause us to abandon the distribution completely - at a bare minimum it should never be on the server version and as a desktop user it would be nice not to have to deal with it there or to have configurable options that avoid it [13:07] ni-ux, funny you say that, when we started 9y ago with snap development it was mainly for server, cloud, industrial, robotics etc ... [13:07] (and this is in fact where they are used in millions nowadays) [13:08] but if you dont like snap, just purge snpd (it will tear out all traces of snaps as well) and be done [13:08] *snapd [14:28] ogra: I can see that being the purpose, but involving people with real world experience would have helped with design decisions and made the technology more usable [14:29] the big no-go for enterprise use-cases is no control of updates, but the resource utilization is also not ideal [14:29] in a world where we have containers there just isn't a reason to use snap in its current form [14:30] I guess I should be asking how I could get involved with the work that goes into the installer on the iso - in the open source ecosystem if we want things changed we should be helping [14:31] ni-ux, nearly all snap development is driven by paying customers ... so i'd guess there are plenty of real-world people involved [14:32] not sure what you mean by "control of updates" ... you can schedule them, hold them, set them to never happen at all etc [14:32] but their use cases are not everyone's use cases... [14:33] especially for enterprises you can also do test installs from beta/candidate channels before moving to the actual thing etc etc ... [14:33] yes by virtue of being a worse version of containerization there are some benefits you would get like that, I do understand that [14:33] i know a lot of enterprises use snaps by utilizing all their features (see snap help for a very minimal subset of what snaps can do) [14:34] "being a worse version of containerization" ? i'd say wuite the opposite and big companies i work with would say that too (given they picked snaps over something like lxd or docker) [14:36] *quite the ... [14:36] I can see why someone developing something to be consumed by the public would be tempted to use snap but their interests are in direction opposition to the consumers of it [14:36] i dont know a single containerization implementation offering the same bits snaps offer really [14:37] what you're really saying is "we have people that pay us and they use this so we're stuck with it" and that probably means for the average user or enterprise consumer we may need to switch distros if it becomes difficult to make sure snap isn't used for core app management [14:37] which is good to know [14:37] paying customers certainly should drive development [14:38] starting from being cppletel read-only gpg signed (i.e. 100% tinker proof) over always compressed and never unpacked on disk .. to automated self-tests on upgrades and roll back (manual or automatic) to built-in fleet management ... to just name a few [14:38] oh, and indeed unlimited parallel installs (for failover and HA cases etc) [14:39] none of those things require the inability to control updates manually... [14:39] LOL ... we're not "stuck with it" ... it isnt like several 100 people inside canonical work on it because we are "stuck with it" ... we developed it for certain use-cases and it is a big hit in these areas [14:39] ni-ux: you can control the updates. Unlike apt you can schedule them but you can also disable them just like apt. What more control do you want? [14:40] control of the version that gets installed for one, not just "refreshing" to get the latest version [14:40] ni-ux, can you please elaborate ? you have full control over upgrades ... [14:40] the only difference is, unlike apt, automatic update is enabled by default. But you can disable them completely [14:40] to be fair, this is fairly new [14:41] ni-ux: https://ubuntuhandbook.org/index.php/2022/11/turn-off-automatic-updates-snap-apps/ [14:41] uh, but you cant do that with other packaging systems either ... if a version is in the archive you have to take it or not ... [14:41] ni-ux: maybe you just weren't aware that you could disable the automatic updates? [14:41] ... and one can roll back [14:41] I can mark, hold, pick specific versions of apt packages [14:41] leftyfb, you could always use the snap-store-proxy and turn it to airgapped mode ... enterprises have done that for years [14:41] it's cool that you can at least disable the updates now but that's still less granular [14:42] ni-ux: "less granular"? How so? [14:42] ni-ux, what kind of granularity are you talking about ... [14:42] ni-ux: you can do it per snap package if you like [14:42] ni-ux, what can you do with a deb from the archive you can not do with a snap in that area ? [14:42] is it now possible to pick a particular version of a snap to update to? or if you enabled the auto update scheduling you'll just get latest [14:43] ni-ux, can you show me how to do that with a deb ? [14:43] eh, that's trivial [14:43] is it ? [14:43] there is usually only one supported deb in the archive [14:43] and that is the latest one uploaded [14:44] not different in any way how snaps behave [14:44] apt install package=version [14:44] ni-ux: you know that only works for the versions available right? As in 2 for a short period of time [14:44] yeah [14:45] you can have multiple repositories with each their own version... [14:45] maybe, but I've used it before to pick specific versions where needed successfully, and if you wanted to you can setup a PPA to help [14:45] same as "snap revert " ... just that it will always work for a snap [14:45] (since the one you revert to is local) [14:46] are these customers publishing apps as snaps? or just consuming them internally [14:46] both [14:46] if they're publishing them...yes of course, it's much easier [14:46] but consuming them vs deb there is no advantage only disadvantages [14:46] for instance, Mozilla publishes the Firefox snap [14:46] any of the virtues you're extolling are better used with containers [14:46] on the consuming side [14:47] yes - I saw that this morning when updating my desktop personal system - it's what led me here [14:47] this cancer has made its way to firefox now too? [14:47] very frustrating [14:47] ni-ux: ok, you don't like snaps and no amount of disproving your misconceptions and perceptions is going to change your opinion. Good luck. [14:47] there are already work-around articles for it like this one: [14:47] https://www.omgubuntu.co.uk/2022/04/how-to-install-firefox-deb-apt-ubuntu-22-04 [14:47] mozially asked for it [14:47] *mozilla [14:48] unlike debs, Mozilla can publish their snap and have the package updated to the latest across all OS's that support snaps, including non-deb-based distro's [14:48] leftyfb: not at all - if we fixed it so a snap isn't an isolated wasteful duplication of libraries and fixed it so updates work the same way as apt without just grabbing "latest", I'd be on board [14:48] package versions in snaps are distro-agnostic [14:49] sure - because they wastefully just include libraries in the package [14:49] it's sheer laziness [14:49] geez [14:49] ni-ux: hence containerized/secure [14:49] can you stop spreading FUD you read in forum comments [14:49] just like flatpak [14:50] and ... what's the one other one ... imagesomething? [14:50] snaps have had shared library snaps since close to the start in 2014 [14:50] I dont' really participate in forums much these complaints are the result of using this OS as my personal desktop for decades, and making leadership decisions in large companies that use the software [14:50] you can dismiss them as "oh this guy just doesn't understand snap" and that's ok - but many have these complaints [14:50] well, if you used them and looked close enoug you will see that libs are mostly shared ... particulary for desktop apps [14:51] that "many have these complaints" does not make them true 🙂 [14:51] snaps or anti-snap are not religions ... just look at the code, or how it functions on your system ... it is all opensource [14:51] I wonder why people hate snaps, but trust apt packages over snaps.. [14:52] its been some good years - we'll continue to route around snap as damage and if that becomes impossible or effort-prohibitive at some point I guess we'll jump ship [14:52] or flatpak, with disadvantages also.. [14:52] ni-ux: good luck [14:52] if there is a way to help the team that does the installer I would appreciate it - that it still assumes you will have separate partitions for /boot, swap, etc... is silly [14:52] the 20 installer allowed a single LVM partition for everything, which is what you want for public or private cloud [14:53] * ogra has admittedly not tried the upcoming 23.04 installer yet ... but what yu complain about has nothing to do with snap packaging [14:53] some native support for a swapfile would be good too but it's not hard to throw one in after the fact [14:53] ni-ux: feel free to file bugs [14:54] yeah it's a separate issue - we're not likely to agree on snap and you've already let us know paid customers like it and app developers are pushing for it [14:54] !bug | ni-ux [14:54] ni-ux: If you find a bug in Ubuntu or any of its official !flavors, please report it using the command « ubuntu-bug » - See https://help.ubuntu.com/community/ReportingBugs for other ways to report bugs. [14:54] we do default to swapfiles since 22.04 IIRC ? [14:54] it's not really a bug in the installer it's just lacking some features of the old installer [14:54] this is for server mind you [14:55] pretty sure no debian-based installer, desktop nor server ever allowed you to specify a swap file as opposed to a swap partition [14:55] yeah a lot of the installers for all distros are stuck in 1995 it's not an ubuntu issue [14:56] well, ubuntu isnt anymore with the upcoming release 🙂 [14:56] ogra: ? [14:56] server switched in 22.04 ... desktop does in 23.04 [14:56] leftyfb, stuck in 1995 [14:56] ah, subiquity [14:57] well, curtin ... [14:57] I'll file a bug and describe what I'm trying to do vs what actually happens and see from there [14:57] 👍 [14:57] good move [14:58] that 22.04.1 bug on installing over a BMC(iDrac) was nasty. Good thing 22.04.2 fixed it. Made my life of documenting the installation for a team much easier without having to write up a workaround [14:58] yeah, bugs are awesome ... when they get fixed 😉 [14:59] well, the Firefox snap is *still* broken... (although I see someone recently marked one of the biggest bugs as "in progress" recently, after years of doing nothing, so let's see...) [15:00] JanC: broken how? [15:00] I explained that here in the past [15:00] Again, fresh install of Ubuntu 22.04 desktop last night, Firefox worked perfectly fine for me to download and install Chrome :) [15:00] there are bug reports (in Launchpad & Mozilla's bugtracker) [15:01] (and AFAIK the Chromium snap has had the same bug, since 2019 or so) [15:01] if one needs an OTA device with firefox snap, yes, it is broken [15:04] JanC, "the bug" is really a bit broad ... my firefox has not had any bugs for me in ages ... nor does my chromium ... [15:06] oerheks, whats an OTA device ? (i only know OTA devices from the embedded world as over the air updating devices) [15:07] in Belgium, one needs a special device to contact Taxservices. [15:07] because you don't (have to) use some features of Firefox does not mean other people don't... [15:07] wasnt that fixed with the native-messaging support being added ? [15:07] it's not just for tax services, it's pretty much all government services, healthcare, social security, etc. [15:07] (which was now ... what 2y ago ?) [15:07] not yet ogra [15:08] Not AFAIK [15:08] oh ! [15:08] that bug still not fixed [15:08] i thought it was using that tech [15:08] im using chrome to bypass that [15:08] and also in some other countries (I think Estonia, ...) [15:08] yeah, i heard that but was under the impression native messaging support had solved it [15:09] it's also not a special device, it's astandard smartcard device [15:09] bug #1741074 [15:09] -ubottu:#ubuntu-discuss- Bug 1741074 in firefox (Ubuntu) "[snap] chrome-gnome-shell extension fails to detect native host connector" [High, In Progress] https://launchpad.net/bugs/1741074 [15:09] ogra: there are countries in Europe that use "the internet" to contact the government. i know for us that is hard to believe :P [15:09] in nl we have DigiID, that is just a 2 factor auth, a simple sms on your phone [15:09] SMS... how secure :P [15:09] ravage, you mean to send around scanned FAX documents to print tem out at the receiver ? [15:09] 😛 [15:09] lol [15:10] ogra: yep :D [15:10] ogra: but this is about all-digital stuff, which they got working a couple decades ago... [15:13] JanC, yeah, not in germany where the biggest digital innovation is still sending docs as beeps over phone lines 😛 [15:14] I'm sure that happens too sometimes (not just at governments, I know for a fact the 2nd biggest telco here worked like that internally until at least 10 years ago), but these online things actually work now [15:15] telco/IT services company really, setting the example... [15:16] My bank still works with this, for over 10y years now https://cdn.nos.nl/image/2015/04/10/147991/xxl.jpg [15:16] same here oerheks [15:16] there is lots to be said about these online government projects (always delivered late, always delivered at huge budget overruns, etc.), but after a while they actually start to work well :) [15:17] that's not bad tech oerheks [15:17] My complaint; i can only set a 4 digit pincode, not 3, not 5-9 digits [15:17] better than SMS or a PIN code... [15:17] yeah, somtimes i suspect there are still sitting wimen in 50s dresses somewhere in the basements of german telcos ... plugging headphone plugs into matrixes of headphn jacks [15:17] oerheks: but you login is blocked after 3 failed attempts [15:18] jups, and i disabled contact-payments. [15:18] no pulling money from my account, by payments of max 25 euro [15:18] eh, I bet half the people did that after a consumer TV show showed how easy it is to steal money with contact-free payments :P [15:19] here [15:19] and no payment by smartphone [15:20] no e-sim too [15:27] BTW, I think (some of) those 2FA devices banks use can do 6-number PINs too [15:28] but banks probably can't enable that until all devices in use (these 2FA devices, ATMs, web/phone apps, ...) support that [15:29] might even be that it is held back to make sure your card still works in US ATMs or something like that :) [15:30] I mean, they used magnetic strips for more than a decade after almost everyone else switched to smartcards too...