IRC chat logs for #ltsp on irc.freenode.net (webchat)


Channel log from 17 May 2016   (all times are UTC)

02:28quinox has left IRC (quinox!~quinox@ghost.qtea.nl, Quit: :wq)
02:29quinox has joined IRC (quinox!~quinox@ghost.qtea.nl)
03:22adrianorg has left IRC (adrianorg!~adrianorg@189.58.238.199.dynamic.adsl.gvt.net.br, Ping timeout: 276 seconds)
03:24adrianorg has joined IRC (adrianorg!~adrianorg@189.58.237.237.dynamic.adsl.gvt.net.br)
03:51kjackal has left IRC (kjackal!~kjackal@athedsl-4547229.home.otenet.gr, Ping timeout: 260 seconds)
05:04alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 244 seconds)
05:55schlady has joined IRC (schlady!~schlady@ip1f111304.dynamic.kabel-deutschland.de)
06:05robb_nl has joined IRC (robb_nl!~robb_nl@62.235.103.195)
06:07ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz)
06:13mikkel has joined IRC (mikkel!~mikkel@mail.dlvs.dk)
07:02alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
07:19kjackal has joined IRC (kjackal!~kjackal@onopfy.static.otenet.gr)
07:57kjackal has left IRC (kjackal!~kjackal@onopfy.static.otenet.gr, Remote host closed the connection)
07:57schlady has left IRC (schlady!~schlady@ip1f111304.dynamic.kabel-deutschland.de, Remote host closed the connection)
07:57robb_nl has left IRC (robb_nl!~robb_nl@62.235.103.195, Ping timeout: 276 seconds)
08:10kjackal has joined IRC (kjackal!~kjackal@onopfy.static.otenet.gr)
08:41robb_nl has joined IRC (robb_nl!~robb_nl@62.235.103.195)
08:44schlady has joined IRC (schlady!~schlady@ip1f111304.dynamic.kabel-deutschland.de)
09:49mrtux has joined IRC (mrtux!b2cacee4@gateway/web/freenode/ip.178.202.206.228)
09:51
<mrtux>
Hi. Every time I try to build an image (i386) it fails due couldn't download package: xyz ... If I retry it I get the same error but a different that couldn't be downloaded.. I tried it several times now ... any suggestions? I'm on a amd64 Server Ubuntu 16.04
09:56book` has left IRC (book`!~book`@2607:5300:100:200::160d, Quit: Leaving)
10:20mrtux has left IRC (mrtux!b2cacee4@gateway/web/freenode/ip.178.202.206.228, Quit: Page closed)
10:23book` has joined IRC (book`!~book`@68.ip-149-56-14.net)
10:43book` has left IRC (book`!~book`@68.ip-149-56-14.net, Quit: Leaving)
10:49Phantomas has joined IRC (Phantomas!~ftsamis@ubuntu/member/phantomas)
10:55lbssousa has joined IRC (lbssousa!~lbssousa@177.143.31.184)
10:58book` has joined IRC (book`!~book`@68.ip-149-56-14.net)
11:31schlady has left IRC (schlady!~schlady@ip1f111304.dynamic.kabel-deutschland.de, )
11:37schlady has joined IRC (schlady!~schlady@ip1f111304.dynamic.kabel-deutschland.de)
11:50schlady has left IRC (schlady!~schlady@ip1f111304.dynamic.kabel-deutschland.de, )
12:02kjackal has left IRC (kjackal!~kjackal@onopfy.static.otenet.gr, Ping timeout: 276 seconds)
12:04GodFather has joined IRC (GodFather!~rcc@2600:1007:b12d:fdf9:8d58:8ea5:40b5:eecf)
12:11GodFather has left IRC (GodFather!~rcc@2600:1007:b12d:fdf9:8d58:8ea5:40b5:eecf, Remote host closed the connection)
12:27mmetzger has joined IRC (mmetzger!~mmetzger@172-125-4-17.lightspeed.rcsntx.sbcglobal.net)
12:46kjackal has joined IRC (kjackal!~kjackal@2a02:587:3117:9e00:58be:3396:de0c:ae9e)
13:10mmetzger has left IRC (mmetzger!~mmetzger@172-125-4-17.lightspeed.rcsntx.sbcglobal.net, Ping timeout: 240 seconds)
13:41
<sbalneav>
Man, writing a pam module is triiiiiicky
13:44adrianorg has left IRC (adrianorg!~adrianorg@189.58.237.237.dynamic.adsl.gvt.net.br, Ping timeout: 260 seconds)
13:46adrianorg has joined IRC (adrianorg!~adrianorg@179.187.29.104.dynamic.adsl.gvt.net.br)
13:48ben_roose has joined IRC (ben_roose!~roose@roose.cs.wichita.edu)
13:55gdi2k has left IRC (gdi2k!~gdi2k@49.151.17.102, Ping timeout: 260 seconds)
13:58gdi2k has joined IRC (gdi2k!~gdi2k@49.151.17.102)
14:07bennabiy has left IRC (bennabiy!~bennabiy@unaffiliated/bennabiy, Remote host closed the connection)
14:10mikkel has left IRC (mikkel!~mikkel@mail.dlvs.dk, Quit: Leaving)
14:37bennabiy has joined IRC (bennabiy!~bennabiy@unaffiliated/bennabiy)
14:44mrtuX has joined IRC (mrtuX!b2cacee4@gateway/web/freenode/ip.178.202.206.228)
14:45mrtuX is now known as meaurieati
14:47
<meaurieati>
Got some problems building i386 on amd64. Anytime I try to build an image some packages won't load. If i try it a second time other packages (some of them may be found and downloaded the last time) won't download. Any ideas on that?
14:48adrianorg has left IRC (adrianorg!~adrianorg@179.187.29.104.dynamic.adsl.gvt.net.br, Ping timeout: 260 seconds)
14:48adrianorg has joined IRC (adrianorg!~adrianorg@179.187.29.104.dynamic.adsl.gvt.net.br)
14:58
<||cw>
meaurieati: sounds like a connection issue. also, whatever ubuntu archive mirror I'm on today is running slow as hell, 152KB/s when I usually get 1MB/s or so (only a 10Mbit connection)
14:59
<meaurieati>
so I just should keep on trying? :)
15:00
<||cw>
I guess. it's one of the reasons I run apt-cacher-ng
15:01
I did a 386 client on amd64 server a couple weeks ago with no issues, so i don't think it's a software issue
15:01
also was on a different lan, no apt-cache there
15:02
<meaurieati>
okay. i'll keep on trying. perhaps tomorrow or later on it'll work fine for me.
15:16ben_roose has left IRC (ben_roose!~roose@roose.cs.wichita.edu, Remote host closed the connection)
15:17ben_roose has joined IRC (ben_roose!~roose@roose.cs.wichita.edu)
15:26adrianorg has left IRC (adrianorg!~adrianorg@179.187.29.104.dynamic.adsl.gvt.net.br, Ping timeout: 252 seconds)
15:26adrianorg has joined IRC (adrianorg!~adrianorg@177.18.182.123)
15:42adrianorg has left IRC (adrianorg!~adrianorg@177.18.182.123, Ping timeout: 260 seconds)
15:44adrianorg has joined IRC (adrianorg!~adrianorg@187.113.244.222)
16:02
<quinox>
I also encountered a really slow Ubuntu mirror yesterday
16:11
<alkisg>
meaurieati: ltsp-build-client --mount-package-cache <other-params>
16:11
This keeps the packages in cache, so you have better chanches each time it fails
16:18
<meaurieati>
I tried --keep-packages and --cache-packages ... thanks for your help.. will try it right now
16:27
alkisg thank you very much. obviously its working fine!
16:31vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
16:39
<meaurieati>
is there any possibility for a client to choose between i386 and amd64. Is there any option to set up the dhcp.conf like if arc="i386" filename "/ltsp/i386/pxelinux.0"; else filename "ltsp/amd64/pxelinux.0"?
16:41
<vagrantc>
there's no option out of the box, although there is some kernel detection support for architectures that support both i386 and amd64 kernels
16:42
you can set up a custom pxelinux.cfg dir that will autodetect and allow you to select a different chroot/image to boot
16:42
using ifcpu4
16:42
and there is some support for doing that from dhcp as well, but i'd recommend using ifcpu64 instead.
16:44
<meaurieati>
thank you. I'll try it on tomorrow
16:48meaurieati has left IRC (meaurieati!b2cacee4@gateway/web/freenode/ip.178.202.206.228, Quit: Page closed)
16:52vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Ping timeout: 260 seconds)
16:54vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
17:37
<Leolo_2>
wooo! I got LDM working on CentOS 6 with NFS
17:37
it only took patching openssh :-)
17:37
now to see if I can get nbd working
17:37
<vagrantc>
Leolo_2: you're probably the first to do that in many years!
17:39
<Leolo_2>
the problem is that ssh needs to create the control socket in an atomic way. so it does bind/listen on a temporary socket, then it does link(temp,final); unlink(temp); but that fails with unionfs
17:39
but change the link/unlink to rename() and it works
17:40
<Hyperbyte>
Leolo_2, isn't another problem that CentOS 6 is quite... old?
17:41
<Leolo_2>
old, but not dead
17:41
but being old isn't a problem. I know my way around CentOS 6
17:41
<Hyperbyte>
Well sure
17:42
But for a good user experience, you'd need to keep things like Firefox, LibreOffice, etc up to date, no?
17:42
That's a lot of work.
17:43
<Leolo_2>
CentOS 6 still uses openoffice
17:43
and firefox is at 38.7.0
17:43
sorry, make that 45.1.0
17:43
<gehidore>
I tried to install open office just a week ago only to realize it had been so long since I used it that it had been forked, I assume, to libreoffice
17:44
Name : firefox
17:44
Version : 46.0.1-1
17:44
is 45.1.0 that old?
17:45
<Leolo_2>
no
17:45
<gehidore>
everything else is just old ?
17:45
:P
17:45
<Leolo_2>
I am currently 100% fine with my users using openoffice and not libreoffice
17:57pscheie has joined IRC (pscheie!8964891e@gateway/web/freenode/ip.137.100.137.30)
17:58gehidore has left IRC (gehidore!~username@unaffiliated/man, Quit: WeeChat 1.5)
18:00gehidore has joined IRC (gehidore!~username@unaffiliated/man)
18:42Phantomas has left IRC (Phantomas!~ftsamis@ubuntu/member/phantomas)
18:46
<maldridge>
gehidore: I don't think it was forked, I think they just changed the name to better reflect the project goals
18:48
<vagrantc>
it was definitely forked
18:49
oracle acquired openoffice.org and did some questionable licensing changes
18:50
<maldridge>
ah
19:01vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Ping timeout: 276 seconds)
19:03vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc)
19:16
<sbalneav>
!paste
19:16
<ltsp>
paste: try !pastebot
19:16
<sbalneav>
!pastebot
19:16
<ltsp>
pastebot: try !pastebin
19:16
<sbalneav>
BWAHAHA
19:16
!pastebin
19:16
<ltsp>
pastebin: the LTSP pastebin is at http://ltsp.pastebin.com. Please paste all text longer than a line or two to the pastebin, as it helps to reduce traffic in the channel. Don't forget to paste the URL of the text here.
19:19
<vagrantc>
!learn paste as the LTSP pastebin is at http://ltsp.pastebin.com. Please paste all text longer than a line or two to the pastebin, as it helps to reduce traffic in the channel. Don't forget to paste the URL of the text here.
19:19
<ltsp>
The operation succeeded.
19:19
<vagrantc>
!paste
19:19
<ltsp>
paste: (#1) try !pastebot, or (#2) the LTSP pastebin is at http://ltsp.pastebin.com. Please paste all text longer than a line or two to the pastebin, as it helps to reduce traffic in the channel. Don't forget to paste the URL of the text here.
19:19
<vagrantc>
!forget paste 1
19:19
<ltsp>
The operation succeeded.
19:19
<vagrantc>
!paste
19:19
<ltsp>
paste: the LTSP pastebin is at http://ltsp.pastebin.com. Please paste all text longer than a line or two to the pastebin, as it helps to reduce traffic in the channel. Don't forget to paste the URL of the text here.
19:19
<vagrantc>
!forget pastebot
19:19
<ltsp>
The operation succeeded.
19:19
<vagrantc>
!forget pastebin
19:19
<ltsp>
The operation succeeded.
19:19
<sbalneav>
vagrantc: http://pastebin.com/M3gDgzFx
19:20
<vagrantc>
!learn pastebot as try !paste
19:20
<ltsp>
The operation succeeded.
19:20
<vagrantc>
learn !pastebin as try !paste
19:20
!pastebin
19:20
<ltsp>
I do not know about 'pastebin', but I do know about these similar topics: 'pastebot'
19:20
<vagrantc>
!pastebot
19:20
<ltsp>
pastebot: try !paste
19:20
<vagrantc>
!learn pastebin as try !paste
19:20
<ltsp>
The operation succeeded.
19:20
<vagrantc>
!pastebin
19:20
<ltsp>
pastebin: try !paste
19:20
<vagrantc>
!pastebot
19:20
<ltsp>
pastebot: try !paste
19:20
<vagrantc>
!paste
19:20
<ltsp>
paste: the LTSP pastebin is at http://ltsp.pastebin.com. Please paste all text longer than a line or two to the pastebin, as it helps to reduce traffic in the channel. Don't forget to paste the URL of the text here.
19:20
<vagrantc>
sorry for the noise
19:21
been meaning to do that for ages
19:21
!past
19:21
<ltsp>
I do not know about 'past', but I do know about these similar topics: 'paste', 'pastebot', 'pastebin'
19:22
<vagrantc>
sbalneav: definitely looking interesting!
19:23
sbalneav: it's a little frightful in that it's surely enough rope to hang oneself with :)
19:29
<sbalneav>
Well, I'm going to try to hit someone up that I know to do a code review.
19:30
Conceptually speaking, so long as this *module* is free of errors, I can't see that handing the pam auth off to an external program that communicates via stdin and stdout is any more dangerous than trying to write pam module from scratch.
19:32
Plus it significantly lowers the bar to someone wanting to write a custom pam or nss service, since they can now write it in a (potentially) safer scripting language than diddling around directly in C, with all it's inherent problems.
19:40robb_nl has left IRC (robb_nl!~robb_nl@62.235.103.195, Remote host closed the connection)
19:41pscheie has left IRC (pscheie!8964891e@gateway/web/freenode/ip.137.100.137.30, Ping timeout: 250 seconds)
19:43
<quinox>
I've made one PAM module in Bash, that was okay
19:44
it allows skipping 2FA for LAN IPs so that fat clients can still login
19:44
<vagrantc>
sbalneav: yeah, i get that
19:46
sbalneav: though i do forsee some people well in over their heads writing a one-off where they wouldn't even try otherwise ... but that's ust a matter of documenting correct behavior in the readme or something
19:46
sbalneav: maybe with a "here be monsters" clause
19:46
<sbalneav>
Oh yes.
19:47
<vagrantc>
there are a lot more people hacking up pam modules for LTSP than i thought...
19:55
<quinox>
http://pamtester.sourceforge.net/ <-- just FTY, this tool made debugging my stuff way easy
19:55
s/FTY/FYI/
20:00kjackal has left IRC (kjackal!~kjackal@2a02:587:3117:9e00:58be:3396:de0c:ae9e, Ping timeout: 260 seconds)
20:11lbssousa has left IRC (lbssousa!~lbssousa@177.143.31.184, Quit: Leaving)
20:12kjackal has joined IRC (kjackal!~kjackal@2a02:587:3117:9e00:b036:4169:f788:fad6)
20:14ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving)
20:31kjackal_ has joined IRC (kjackal_!~quassel@2a02:587:3117:9e00:b036:4169:f788:fad6)
20:34kjackal_ has left IRC (kjackal_!~quassel@2a02:587:3117:9e00:b036:4169:f788:fad6, Remote host closed the connection)
20:35kjackal_ has joined IRC (kjackal_!~quassel@2a02:587:3117:9e00:b036:4169:f788:fad6)
20:39kjackal has left IRC (kjackal!~kjackal@2a02:587:3117:9e00:b036:4169:f788:fad6, Quit: Leaving)
21:02bennabiy has left IRC (bennabiy!~bennabiy@unaffiliated/bennabiy, Read error: Connection reset by peer)
21:32ben_roose has left IRC (ben_roose!~roose@roose.cs.wichita.edu, Remote host closed the connection)
22:18kjackal_ has left IRC (kjackal_!~quassel@2a02:587:3117:9e00:b036:4169:f788:fad6, Ping timeout: 260 seconds)
22:21
<Leolo_2>
ok, now I have LTSP working on CentOS 6 with NDB and LDM !
22:22
<vagrantc>
ACRONYM POWER!
22:22
<Leolo_2>
question : how do I aleviate the need for a 15 minute ltsp-update-image every time I want to tweak something in the client?
22:23
2nd question : every time a client boots, I get 10 NDB connect/disconnects
22:23
is this normal?
22:24
<vagrantc>
Leolo_2: regarding the first question ... reduce the size of your image or get faster hardware
22:24
<Leolo_2>
is there no other way?
22:24
<vagrantc>
Leolo_2: it will likely get at least one NBD connection attempt to check for newer images when LDM starts
22:25
Leolo_2: use NFS>
22:25
<Leolo_2>
I thought NFS was the old, deprecated way and NBD was the New Hotness
22:25
<vagrantc>
Leolo_2: though that's getting harder and harder to do, as various unionfs/aufs/overlayfs sometimes have compatibility issues with NFS
22:25
NBD is pretty old hotness at this point
22:26
<Leolo_2>
what's the New Hotness then?
22:27
<vagrantc>
nothing solid has come along, but some folks have experimented with AoE
22:27
<Leolo_2>
I thought there was some loopback + ext4 magic that could make updating the .nbd less painful
22:27
I have AoE running for my CORAID
22:27
I thought AoE server on Linux was sub-optimal
22:27
<vagrantc>
yeah, you can use the chroot with a real filesystem, and then updates go live sort of immediately
22:28
<Leolo_2>
oh
22:28
<vagrantc>
i know some folks have played with btrfs and such
22:28
<Leolo_2>
so /opt/ltsp/x86_64 is it's own FS. and NBD would just serve the underlying device
22:28
its
22:28
<vagrantc>
that's one theory of operation, sure.
22:29
<Leolo_2>
that would require everyone to log out if I want to play in the chroot
22:29
at least, safely
22:29
<vagrantc>
gets a little complicated, in that some filesystems don't assume the underlying FS will change underneath a running filesystem
22:29
i suspect btrfs snapshots might be able to work around that issue
22:30
i think alkisg even has experimented with directly exporting the server's host filesystem
22:30
<Leolo_2>
I'm pretty sure I don't want to drink the btrfs coolaid
22:30
just yet
22:30
<vagrantc>
there are countless options to explore, but the default out of the box experience is NBD+squashfs
22:30
<Leolo_2>
what advantage does NBD have over NFS?
22:31
<vagrantc>
i resisted for ages with debian, using NFS+aufs, but aufs support was dropped in favor of overlay FS, and that has incompatibilities with NFS
22:31
<Leolo_2>
another thing I notice is that mtab is mostly empty on the client. there's no mount -f going on to flush it out
22:32
<vagrantc>
Leolo_2: NBD is much faster boot times, and load times, which really gets important with LTSP Fat Clients
22:32
<Leolo_2>
overlay FS is another name for unionfs?
22:32
<vagrantc>
overlay FS is another union-style filesystem implementation
22:33
unionfs, aufs, overlayfs, and it's close cousin overlay fs.
22:33
all union-style filesystems.
22:33
overlay fs is the first to get into the mainline kernel.
22:35
<Leolo_2>
on DRBL if you have a signle system image, they have one NFS root but they also have template_{etc,root,var}.tgz which is used to modify the root at boot time
22:36
that could be easily set up. NBD root and then patches on top of it
22:36
"patches"
22:36
and when done tweaking, a 15 minute ltsp-update-image
22:37
<vagrantc>
you would need a filesystem that supported all that
22:37
<Leolo_2>
uhm
22:37
<vagrantc>
aufs + NFS was working for Debian for several years ... but support has been dropped
22:38
oh, you mean a writeable NFS root?
22:38
<Leolo_2>
NO !
22:38
writable NFS ? *shudder*
22:38
<vagrantc>
yes
22:38
<Leolo_2>
I mean, NBD + unionfs, then grab a "tweak.tgz" (over tftp?) and unpack that on the client
22:39
<vagrantc>
ah, sure.
22:39
well, except for unionfs
22:39
is there a working unionfs for any recent kernel version?
22:40
there's also support in squashfs to simply append to the existing squashfs ... but i don't think it supports deletions
22:40
so you can add or replace files, but ... that gets messy.
22:41
at any rate, we're always happy to review working patches!
22:41
<Leolo_2>
2.6.32-573 has unionfs which is good enough for my purposes
22:42
<vagrantc>
that must be centos/redhat patches on top of 2.6.32
22:42
<Leolo_2>
yes
22:42
does LTSP 5 and Ubuntu not use unionfs?
22:42
<vagrantc>
not for many years
22:42
if ever...
22:43
<Leolo_2>
huh
22:43
<vagrantc>
they used aufs, maybe unionfs-fuse for a blink of an eye, and then overlayfs ... and now overlay fs.
22:43
<Leolo_2>
I ask because I have trouble figuring out where LTSP ends and CentOS begins
22:43
wait, is fuse.unionfs distinct from unionfs?
22:44
<vagrantc>
LTSP5 is just a collection of code to glue together the distro of choice to support LTSP network booting
22:44
<Leolo_2>
yeah
22:45
well, it's a bit more then that because it also mounts /home/USER and starts up X and so on
22:45
<vagrantc>
sure, there are some LTSP specific parts
22:45
we leave most of the implementation up to the distro developers
22:46
which have been largely absent for redhat/fedora/centos, unfortunately
23:03
<Leolo_2>
apart from the ssh socket bug, it works
23:21
oh wait, there's the inifite "Newer NBD image exists. Reboot?" bug
23:24
<vagrantc>
well, that would explain your numerous NBD connects/disconnects
23:32
<Leolo_2>
it didn't do the inifite loop thing the first time I booted
23:42
<vagrantc>
basically, it connects to the NBD server with a second NBD device, and compares the headers of the two NBD devices to check for differences
23:43
is it connecting to the correct server, with the correct NBD image name?
23:43
or is it using the deprecated NBD port method to select images?
23:50
Leolo_2: good luck!
23:50* vagrantc waves
23:50vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving)
23:55
<Leolo_2>
dedicated NBD port
23:56
IIRC this was one bug that pushed me to use NFS back in the day