00:40 | kjackal has left IRC (kjackal!~quassel@66.78.209.68, Ping timeout: 240 seconds) | |
00:53 | vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving) | |
01:03 | dgroos has joined IRC (dgroos!~dagro001@63.225.132.145) | |
01:11 | dgroos has left IRC (dgroos!~dagro001@63.225.132.145, Quit: dgroos) | |
01:40 | dsjii has joined IRC (dsjii!~david@047-134-241-234.res.spectrum.com) | |
02:07 | gdi2k has left IRC (gdi2k!~gdi2k@58.69.160.27, Read error: Connection reset by peer) | |
02:08 | gdi2k has joined IRC (gdi2k!~gdi2k@58.69.160.27) | |
03:04 | dsjii has left IRC (dsjii!~david@047-134-241-234.res.spectrum.com, Ping timeout: 276 seconds) | |
04:47 | kjackal has joined IRC (kjackal!~quassel@66-180-252-71.static-ip.telepacific.net) | |
05:33 | kjackal has left IRC (kjackal!~quassel@66-180-252-71.static-ip.telepacific.net, Remote host closed the connection) | |
05:37 | <alkisg> dgroos, are you talking about this one? https://github.com/epoptes/epoptes/issues/103
| |
05:37 | Test if it happens when you **haven't** used the epoptes "exec dialog" at all for a user
| |
06:04 | * alkisg starts testing https://github.com/ipxe/pipxe ... | |
06:07 | <alkisg> Woah it just worked out of the box!!!
| |
07:01 | uumas__ has left IRC (uumas__!uumas@kapsi.fi, Remote host closed the connection) | |
07:03 | <alkisg> ...but it needs an 64bit kernel... I just filed https://github.com/ipxe/pipxe/issues/1 for it
| |
07:43 | woernie has joined IRC (woernie!~werner@p57A0E783.dip0.t-ipconnect.de) | |
08:06 | GekkePrutser has left IRC (GekkePrutser!~gp@unaffiliated/gekkeprutser, Ping timeout: 245 seconds) | |
08:35 | sutula has left IRC (sutula!~sutula@184.100.153.93, Ping timeout: 276 seconds) | |
08:38 | sutula has joined IRC (sutula!~sutula@184.97.9.153) | |
10:06 | statler_ has joined IRC (statler_!~Georg@gwrz3.lohn24.de) | |
11:14 | georgene_ has joined IRC (georgene_!~georgeneo@217.27.33.213) | |
11:17 | <georgene_> good afternoon, alkisg, I have a question about user groups. currently I am using chrome group policies which uses a policy.json in /etc/opt/chrome/policies/managed, with my previous setup before LTSP, I used ansible to manually place this policy.json file for computers in different groups. with LTSP however, all computers use the same boot image, thus, the same policy.json file. my question, is there a way to specify either by MAC address or
| |
11:17 | something, which computers should get which variation of this policy.json file?
| |
11:18 | I am looking at creating virtualmachine image perhaps, one for each group, but again, how can I specify which computers should boot which image?
| |
11:37 | I think I found something, is this it? :D
| |
11:37 | [pc01]
| |
11:37 | INCLUDE=nvidia
| |
11:37 | [nvidia]
| |
11:37 | POST_INIT_LN_XORG="ln -sf ../ltsp/xorg-nvidia.conf /etc/X11/xorg.conf"
| |
12:17 | mgariepy has joined IRC (mgariepy!~mgariepy@ubuntu/member/mgariepy) | |
12:33 | <alkisg> Hi georgene, reading...
| |
12:35 | georgene, yes, something like that, you would use e.g.
| |
12:35 | [mac:address] POST_INIT_LN_POLICY="ln -sf /etc/ltsp/policy-group1.json /etc/opt/chrome/policies/managed"
| |
12:58 | kjackal has joined IRC (kjackal!~quassel@66-180-252-71.static-ip.telepacific.net) | |
13:00 | section1 has joined IRC (section1!~section1@178.33.109.106) | |
13:00 | georgene_ has left IRC (georgene_!~georgeneo@217.27.33.213, Quit: Textual IRC Client: www.textualapp.com) | |
13:01 | georgene_ has joined IRC (georgene_!~georgeneo@217.27.33.213) | |
13:33 | support^LTSP has joined IRC (support^LTSP!c5f5eb2a@dsl-197-245-235-42.voxdsl.co.za) | |
13:33 | <support^LTSP> Hi
| |
13:54 | support^LTSP has left IRC (support^LTSP!c5f5eb2a@dsl-197-245-235-42.voxdsl.co.za, Remote host closed the connection) | |
14:00 | spaced0ut has left IRC (spaced0ut!~spaced0ut@unaffiliated/spaced0ut, Remote host closed the connection) | |
14:02 | spaced0ut has joined IRC (spaced0ut!~spaced0ut@unaffiliated/spaced0ut) | |
14:07 | kjackal has left IRC (kjackal!~quassel@66-180-252-71.static-ip.telepacific.net, Ping timeout: 250 seconds) | |
14:26 | Faith has joined IRC (Faith!~Paty_@unaffiliated/faith) | |
14:32 | kjackal has joined IRC (kjackal!~quassel@66-180-252-71.static-ip.telepacific.net) | |
15:25 | georgene_ has left IRC (georgene_!~georgeneo@217.27.33.213, Quit: My MacBook has gone to sleep. ZZZzzz…) | |
15:32 | ltspuser5678 has joined IRC (ltspuser5678!d510b783@gateway/web/cgi-irc/kiwiirc.com/ip.213.16.183.131) | |
16:16 | kjackal has left IRC (kjackal!~quassel@66-180-252-71.static-ip.telepacific.net, Ping timeout: 240 seconds) | |
16:19 | dsjii has joined IRC (dsjii!~david@047-134-241-234.res.spectrum.com) | |
16:22 | georgene_ has joined IRC (georgene_!~georgeneo@31.216.123.235) | |
16:27 | georgene_ has left IRC (georgene_!~georgeneo@31.216.123.235, Ping timeout: 240 seconds) | |
16:36 | dsjii has left IRC (dsjii!~david@047-134-241-234.res.spectrum.com, Ping timeout: 240 seconds) | |
17:21 | statler has left IRC (statler!~Georg@p5B30EE2F.dip0.t-ipconnect.de, Remote host closed the connection) | |
17:24 | GodFather has joined IRC (GodFather!~rcc@d53-64-7-141.nap.wideopenwest.com) | |
17:54 | douglas_br has joined IRC (douglas_br!bb5f6554@187.95.101.84) | |
17:54 | <douglas_br> hello
| |
17:56 | please! Intel J3060 with 4 GB can be good Ltsp client?
| |
17:56 | 4 GB Ram
| |
17:58 | <ltspuser5678> cpubenchmark 978 - https://www.cpubenchmark.net/cpu.php?cpu=Intel+Celeron+J3060+%40+1.60GHz&id=2852
| |
18:01 | Whether it will work well for you, depends upon what applications you'll be running (locally).
| |
18:02 | <douglas_br> we are trying to found a good but not much expensive ltsp client
| |
18:04 | ltspuser5678 thank you. internet app: youtube, skype, libreoffice, educational games, not heavy apps
| |
18:04 | <ltspuser5678> Sure, but for which APPLICATION. It's probably fine for rdp/ica/etc, but not so good if you want to decode HD video.
| |
18:05 | <douglas_br> no no, decode HD video no
| |
18:06 | office and school use
| |
18:06 | <alkisg> douglas_br: how much money did you find this with?
| |
18:07 | 1000 cpu score can browse today, but it might not be able to browse in 5 years from now
| |
18:07 | So if you want it for many years, it might be best to buy something better
| |
18:08 | Imagine that laptops with 400 euros have 10.000 score, i.e. are 10 times faster than that
| |
18:08 | You can also buy refurbished PCs with 50 euros that have 2000 cpu score
| |
18:09 | <douglas_br> hummm I see
| |
18:10 | <alkisg> Also note that ltsp specifications are the same as "normal pc specifications", hardware isn't ltsp specific
| |
18:11 | This one, with windows and hard disk and all, is 239 euros, and 2600 score: https://www.e-shop.gr/innovator-3-office-basic-j4105-me-windows-10-p-PER.908453
| |
18:11 | So you should be able to find new ones with <= 200 euros with 2500+ score
| |
18:12 | <douglas_br> let me see alkisg
| |
18:12 | <alkisg> !cheap-client
| |
18:12 | <ltspbot> cheap-client: https://www.gearbest.com/tv-box-c_11262/?attr=2081-1279
| |
18:13 | Chakal has joined IRC (Chakal!b1b9f0d1@177.185.240.209) | |
18:20 | <mwalters> douglas_br: I've been buying gigabyte barebones mini computers & intel nucs
| |
18:20 | usually come out around $300usd (with memory, no monitor/keyboard/mouse)
| |
18:20 | 8gb memory
| |
18:21 | most of them generally have a celeron N3...something
| |
18:22 | 90% of our workload is gmail/webapps, though
| |
18:23 | <douglas_br> mwalters yes, the problem that here all expensive..... argh same the simple config
| |
18:23 | <mwalters> Brasil?
| |
18:23 | <douglas_br> yes
| |
18:24 | eddyTV has joined IRC (eddyTV!~eddyTV@unaffiliated/eddytv) | |
18:24 | <mwalters> Yeah, I used to work for a company with an office in Sao Paulo, it was always hard to find good equipment
| |
18:25 | (locally for them, I mean)
| |
18:25 | <douglas_br> so all time we need to do magic for something run well
| |
18:26 | mwalters you are right
| |
18:27 | <mwalters> Half the time we ended up shipping most stuff from the USA ;)
| |
18:28 | buying it locally then shipping it was usually cheaper
| |
18:30 | <eddyTV> alkisg (or anybody else who might know): I've installed `lightdm` with `--no-install-recommends` and now autologin has stopped working. I'm guessing I need to manually install a "greeter", like `lightdm-autologin-greeter`?
| |
18:30 | <alkisg> eddyTV: if you don't have a greeter, yes, you need one
| |
18:32 | vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc) | |
18:35 | <eddyTV> Alright I will give the autologin one a try.
| |
18:36 | ltspuser5678 has left IRC (ltspuser5678!d510b783@gateway/web/cgi-irc/kiwiirc.com/ip.213.16.183.131, Ping timeout: 240 seconds) | |
18:36 | <eddyTV> Did you see my question from a couple days ago about the "best practice" for adding options to the kernel boot cmdline? (I want to try `quiet splash` and hide as much of the "computery" boot stuff as possible until the X session starts.)
| |
18:36 | Chakal has left IRC (Chakal!b1b9f0d1@177.185.240.209, Ping timeout: 260 seconds) | |
18:36 | douglas_br has left IRC (douglas_br!bb5f6554@187.95.101.84, Ping timeout: 260 seconds) | |
18:37 | douglas_br has joined IRC (douglas_br!bb5f6554@187.95.101.84) | |
18:39 | <alkisg> eddyTV: KERNEL_PARAMETERS should be make to work there
| |
18:39 | it only works per mac currently
| |
18:42 | * vagrantc is very happy to see more activity again | |
18:42 | <vagrantc> noticing a lot more emails actually discussing improving things rather than just workarounds
| |
18:42 | <douglas_br> mwalters aboyt
| |
18:43 | mwalters about your text - yes I believe you
| |
18:44 | <alkisg> vagrantc: indeed; I feel bad because I have to work less on LTSP for the next year and I won't be able to help jumpstart the new community as much as I'd like to
| |
18:45 | Hopefully I'll make it up next year
| |
18:45 | <eddyTV> Out of curiosity, has there been any performance testing done on which distro gets clients up and in to X the fastest, or other ways to optimize an LTSP chroot to minimize startup times?
| |
18:45 | <vagrantc> alkisg: you've been doing plenty thus far :)
| |
18:45 | <eddyTV> Agreed! The new LTSP stuff is really nice.
| |
18:46 | <alkisg> eddyTV: eh, around Ubuntu 10.04 they focused on startup times, so I was able to netboot clients in 10 secs; but both Ubuntu and me lost interest in fast startup times, so now it's more like 30 secs...
| |
18:46 | It's more important if it's fast after boot
| |
18:46 | Clear Linux claims very fast startup times
| |
18:47 | <eddyTV> True. I just keep hoping there are things that can be disabled / pre-optimized / etc. to eek out those few extra seconds of dead time
| |
18:47 | <alkisg> For example, if you cache kernel/initrd locally, it should make things a lot faster
| |
18:47 | Than trying to squeeze 1-2 seconds out of systemd
| |
18:48 | The equivalent of IPAPPEND 3 would also help to save the second dhcp request
| |
18:48 | <eddyTV> I was going to ask about that next :)
| |
18:48 | <vagrantc> and then kernel/initrd updates need some mechanism to auto-update and reboot on updates ... which will double the boot time for a new kernel
| |
18:48 | <eddyTV> Is that on the roadmap for LTSP19?
| |
18:49 | (passing IP config to the kernel from the pre-boot environment)
| |
18:49 | <vagrantc> i think you can just pass the arguments via ltsp.ipxe script, no?
| |
18:49 | ltspuser5678 has joined IRC (ltspuser5678!d510b783@gateway/web/cgi-irc/kiwiirc.com/ip.213.16.183.131) | |
18:49 | <alkisg> eddyTV: you can just pass ip=${ip}:${srv} etc, yeah
| |
18:49 | <vagrantc> e.g. ip=${ip}:...
| |
18:49 | <alkisg> vagrantc: did you see the bug about ipxe supporting rpi3?
| |
18:49 | <eddyTV> oh really? is that documented anywhere?
| |
18:50 | <vagrantc> that sounds like a nice small contribution someone could implement to get their feet wett :)
| |
18:50 | <alkisg> I saw the new ltsp19 menu in rpi3, it was awesome
| |
18:50 | <vagrantc> alkisg: oh, you mean an upstream ipxe bug?
| |
18:51 | <alkisg> https://github.com/ltsp/ltsp/issues/68#issuecomment-554922796
| |
18:51 | <ltspuser5678> I read some Github comments by Alkis a few days ago, regarding DMs (lightdm, gdm, sddm etc) and he mentioned nodm
| |
18:51 | <vagrantc> alkisg: i surely read it, but kind of glazed over it probably :)
| |
18:51 | <alkisg> vagrantc: I think that tomorrow I will be able to make the kernel boot too, it just needs a veeeery recently patched kernel
| |
18:51 | <vagrantc> i've got an rpi3b+ i could try it on
| |
18:52 | <ltspuser5678> nodm might be a good choice for all those people using ltsp for kiosks. I know its author stopped updating it, but I guess it still works ok for X
| |
18:52 | <alkisg> Great, will document it in that bug report
| |
18:52 | ltspuser5678: lightdm-gtk-greeter should be fine too
| |
18:53 | <vagrantc> looks like the primary ipxe author is working on it ... wonder if it will go upstream :)
| |
18:53 | <alkisg> And it should also work for wayland
| |
18:53 | statler_ has left IRC (statler_!~Georg@gwrz3.lohn24.de, Remote host closed the connection) | |
18:53 | <alkisg> vagrantc: upstream where? the kernel patch went upstream, tianocore supports rpi3 upstream, and mcb30 is upstream for pipxe... everything is fine
| |
18:53 | <vagrantc> i guess it's just the ipxe arm64 target
| |
18:54 | ltspuser5678 is now known as kostas5678 | |
18:54 | <vagrantc> alkisg: yes, i was kind of joking :)
| |
18:54 | <alkisg> whoops :)
| |
18:54 | <vagrantc> alkisg: but it seems a separate project from upstream
| |
18:54 | <alkisg> Ah you mean integrated into ipxe
| |
18:54 | kostas5678: any reason to prefer nodm over lightdm-gtk-greeter? Just size?
| |
18:55 | <vagrantc> yeah, but sounds like it's just using the normal arm64-efi build of ipxe
| |
18:55 | <alkisg> He said it can boot rapsbian, so I don't care how he does it :D
| |
18:55 | (with 32bit userspace)
| |
18:55 | <vagrantc> would kind of be nice to skip the EFI part...
| |
18:56 | <alkisg> If tianocore can support all boards, it would be a nice layer of compatibility for projects like ipxe to sit upon
| |
18:56 | <vagrantc> looks like i should get an arm64-efi build in the debian ipxe packages...
| |
18:56 | no word from waldi the debian ipxe maintainer... :/
| |
18:56 | <alkisg> Did you manage to get feedback from debian ipxe maintainer?
| |
18:56 | :(
| |
18:57 | (08:54:59 PM) alkisg: kostas5678: any reason to prefer nodm over lightdm-gtk-greeter? Just size? ==> whoops sorry I meant lightdm-autologin-greeter there
| |
18:57 | From the nodm author himself
| |
18:59 | Display managers need to do fancy stuff and set environment variables etc that always change year after year; we've seen that in ldm too, that's why we stopped maintaining it, like nodm's author did too
| |
19:04 | <kostas5678> alkisg regarding nodm, not just speed/size, it seems to have all features needed for a good autologin https://github.com/spanezz/nodm
| |
19:06 | <alkisg> kostas5678: "unmaintained" really means unmaintained; i.e. it may stop working at any time with no support
| |
19:06 | If the author himself decided to implement another DM and abandon it,there was a good reason for it
| |
19:07 | https://github.com/spanezz/lightdm-autologin-greeter
| |
19:09 | <kostas5678> I had noticed that ... Anyway nodm is included in repos (with some patches) https://packages.debian.org/search?keywords=nodm
| |
19:09 | I assume due to Wayland support?
| |
19:09 | <alkisg> lightdm-autologin-greeter is included too in repos; and possibly nodm might be removed in the future, as it's abandoned upstream
| |
19:18 | uumas_ has joined IRC (uumas_!uumas@kapsi.fi) | |
19:28 | GekkePrutser has joined IRC (GekkePrutser!~gp@unaffiliated/gekkeprutser) | |
19:37 | <kostas5678> Apparently people are still inquiring about nodm https://github.com/spanezz/nodm/issues/19
| |
19:38 | <vagrantc> and there's some maintenance in the debian packaging as of this year ... but it hasn't seen a new upstream release in many years
| |
19:39 | <alkisg> Btw, why all the talk about nodm, does anyone need ltsp do to something about it? Isn't lightdm support enough?
| |
19:40 | <kostas5678> Mostly academic interest at this point, since you mentioned it as an option some days ago.
| |
19:40 | <vagrantc> surely lightdm works sufficiently, nodm is much smaller than lightdm ... but maintained vs. unmaintained or barely maintained seems like an important consideration :)
| |
19:41 | <eddyTV> Yeah that was probably my fault. I brought up that adding `lightdm` brought in an additional 270+ packages because "recommended" packages get installed by default.
| |
19:41 | <fiesh> according to https://wiki.archlinux.org/index.php/Nodm, nodm is being superseded by lightm...
| |
19:41 | <eddyTV> Alkis pointed out I should install with `--no-install-recommends`, which indeed solved the problem and added only 2 packages
| |
19:42 | So I now it sounds like I just need to manually add `lightdm-autologin-greeter` and I'll be all set
| |
19:42 | <alkisg> eddyTV: btw, I don't think anyone has tried lightmd-autologin-greeter yet; if it does need any configuration, ltsp could surely do that
| |
19:43 | AFAIK the lightdm configuration done by ltsp should be enough though
| |
19:43 | Also, I think it's just a few KB, so we might recommend it as the default for anyone needing autologin
| |
19:43 | <kostas5678> Well, I had kept nodm in the back of my mind as an option for autologin in constrained ram / cpu hardware (like the t5745)
| |
19:44 | <alkisg> The greeters don't stay in RAM after login
| |
19:44 | So that would only affect people needing to cache the squashfs image to RAM
| |
19:45 | <kostas5678> Right. As I said, it's purely academic for me right now. And if needed, I guess one just has to edit 55-display-manager.sh ?
| |
19:46 | <alkisg> 55-display-manager handles the display managers it knows about
| |
19:46 | Since it doesn't know about nodm, I don't think it needs to be edited
| |
19:46 | If someone wants to add support for nodm, he'd write 55-nodm...
| |
19:46 | <kostas5678> Ah, OK, thx.
| |
19:47 | <alkisg> Or just a POST_INIT_xx command
| |
19:51 | Change of subject... About printing, I'm thinking this: CUPS_SERVER="$SERVER" by default (that's the IP there), meaning that administrator should enable printer sharing on the server and clients would use the server CUPS, disabling their own CUPS,
| |
19:51 | we also want CUPS_SERVER=localhost, which would allow a client to print to its own printer, after automatically installing it on boot (or get printer.conf via some POST_INIT_xx command),
| |
19:51 | and finally, CUPS_SERVER="", would disable all printer handling, to allow administrators to handle printing however they like
| |
19:51 | ...does that CUPS_SERVER="" make any sense in a configuration file? Does it feel like "don't touch my cups settings"?
| |
19:52 | <vagrantc> it feels a bit like something's missing :)
| |
19:52 | e.g. many systems replace empty values with default values
| |
19:53 | <alkisg> We could also special-case CUPS_SERVER=ignore or disabled... or have yet another variable, like DONT_TOUCH_CUPS=1 :D
| |
19:54 | douglas_br has left IRC (douglas_br!bb5f6554@187.95.101.84, Remote host closed the connection) | |
19:54 | <vagrantc> well, disabled would suggest it might actually do things to actually disable a configured cups environment
| |
19:54 | so ignore seems best? maybe?
| |
19:54 | * alkisg googles for synonyms of "ignore" :D | |
19:56 | <vagrantc> CUPS_SERVER=stay-out-of-the-way
| |
19:56 | CUPS_SERVER=manual ?
| |
19:57 | <alkisg> unmanaged?
| |
19:57 | <vagrantc> that could work too
| |
19:57 | unspecified?
| |
19:58 | basically, this variable is weather ltsp does anything specific regarding cups configuration, yes?
| |
19:58 | leave-it-alone or server/localhost ?
| |
19:58 | <alkisg> Hrm... actually I'm trying to remember what was the difference between "localhost", which is actually the debian/ubuntu default, and "unspecified", where we don't do anything
| |
19:59 | I don't remember what I was planning to do for localhost, if anything
| |
20:01 | <vagrantc> i guess it would be ok for it to be set to empty if the default was empty ... but then overriding a set value with the empty value looks weird
| |
20:03 | <alkisg> There are 2 things to do: (1) set /etc/cups/client.conf so that the client uses the server cups, and (2) disable the cups service.
| |
20:03 | With CUPS_SERVER=$SERVER, we need both. With CUPS_SERVER=unspecified==localhost, we need none.
| |
20:03 | ...the weird case was, when we need (1) but not (2): in this case, the client would still use the server cups, but it would also run its own cups, exposing its printers to the network, and the server would see that printer and make it available to all the clients
| |
20:03 | I think that was what I was trying to remember...
| |
20:05 | Then maybe KEEP_SYSTEM_SERVICES=cups is the way to describe that
| |
20:06 | section1 has left IRC (section1!~section1@178.33.109.106, Quit: Leaving) | |
20:12 | <alkisg> Or maybe CUPS="use-case-or-server" is better than CUPS_SERVER. I.e. CUPS="$SERVER", CUPS="unmanaged", CUPS="share-printers"
| |
20:13 | <vagrantc> that does seem a bit cleaner...
| |
20:13 | <alkisg> Or 3 variables, CUPS_SERVER, CUPS_ENABLE, CUPS_SHARE. Meh. Too late to think, later :)
| |
20:32 | kostas5678 has left IRC (kostas5678!d510b783@gateway/web/cgi-irc/kiwiirc.com/ip.213.16.183.131, Remote host closed the connection) | |
20:53 | mgariepy has left IRC (mgariepy!~mgariepy@ubuntu/member/mgariepy, Remote host closed the connection) | |
20:58 | woernie has left IRC (woernie!~werner@p57A0E783.dip0.t-ipconnect.de, Remote host closed the connection) | |
20:59 | eddyTV has left IRC (eddyTV!~eddyTV@unaffiliated/eddytv, Quit: Not legit enough.) | |
22:13 | dsjii has joined IRC (dsjii!~david@047-134-241-234.res.spectrum.com) | |
23:12 | Faith has left IRC (Faith!~Paty_@unaffiliated/faith, Quit: Leaving) | |
23:33 | bengoa has left IRC (bengoa!~alberto@194.50.55.200, Ping timeout: 246 seconds) | |
23:45 | GodFather has left IRC (GodFather!~rcc@d53-64-7-141.nap.wideopenwest.com, Ping timeout: 240 seconds) | |
23:52 | bengoa has joined IRC (bengoa!~alberto@194.50.55.200) | |