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


Channel log from 13 September 2009   (all times are UTC)

00:12F-GTSC has joined #ltsp
00:18FGXR6 has quit IRC
00:29chupacabra has joined #ltsp
00:30alkisg2 has joined #ltsp
00:41ifny has joined #ltsp
00:47F-GTSC has quit IRC
00:49F-GTSC has joined #ltsp
01:12alkisg2 has quit IRC
01:15alkisg has joined #ltsp
02:02yoshi__ has joined #ltsp
02:16vagrantc has quit IRC
02:18yoshi__ has quit IRC
02:42Ahmuck-Sr has quit IRC
02:57ogra has joined #ltsp
03:33cyberorg has joined #ltsp
03:58ogra_ has joined #ltsp
04:15ogra has quit IRC
05:16yanu has quit IRC
05:38yanu has joined #ltsp
05:52yoshi__ has joined #ltsp
06:05garymc has joined #ltsp
06:10alkisg has quit IRC
06:16garymc has quit IRC
06:19alkisg has joined #ltsp
06:19wimiko1 has joined #ltsp
06:28mikkel has joined #ltsp
06:32vvinet has joined #ltsp
07:17lieven1 has joined #ltsp
07:19
<lieven1>
Hi guys, anybody knowing which X driver module is needed in kvm guest (to use as testenvironment for ltsp-client)? Autodetect doesn't seem to work (Ubuntu Jaunty client image).
07:43vvinet has quit IRC
07:55ogra__ has joined #ltsp
08:14
<lieven1>
I have two kvm guests: one is the ltsp server, booting fine using cirrus X-driver. Other is ltsp client diskless guest booting of the ltsp server. X is not started on client. Xorg.log @ http://pastebin.com/m6d4ef576 Any suggestions?
08:16ogra__ has quit IRC
08:25wwx has quit IRC
08:27yoshi__ has quit IRC
08:33wwx has joined #ltsp
08:40pmatulis has joined #ltsp
08:45BugsBunnyBR has quit IRC
08:45alkisg has quit IRC
08:49
<lieven1>
ok, worked around problem by generating xorg.conf on another kvm-guest and passing it via lts.conf. Problem seems to be framebuffer module, which is autoloaded by default on autogenerated xorg.conf on ltsp client side... Might be nice to be able to blacklist this module in lts.conf.
08:51
got some issues with it before on physical ltsp clients too (with nvidia hardware)
09:10alkisg has joined #ltsp
09:41wwx has quit IRC
09:44wwx has joined #ltsp
10:20F-GTSC has quit IRC
10:23F-GT has joined #ltsp
11:16Egyptian[Home] has quit IRC
11:17Egyptian[Home] has joined #ltsp
11:31vagrantc has joined #ltsp
11:53rjune_ has quit IRC
11:54rjune_ has joined #ltsp
12:03rjune__ has joined #ltsp
12:05alkisg has quit IRC
12:16rjune_ has quit IRC
12:19alexqwesa has quit IRC
12:24alexqwesa has joined #ltsp
12:53alkisg has joined #ltsp
13:21garymc_ has joined #ltsp
13:38bobby_C has joined #ltsp
13:43garymc_ has joined #ltsp
13:43garymc_ is now known as gartmc
13:44gartmc is now known as Garty
13:46Ahmuck-Sr has joined #ltsp
13:47yoshi__ has joined #ltsp
13:53yoshi__ has quit IRC
13:58Garty has joined #ltsp
14:11Garty has joined #ltsp
14:12Garty has left #ltsp
14:12Garymc has joined #ltsp
14:15Garymc is now known as garymc
14:31atomic007za has joined #ltsp
14:41yoshi__ has joined #ltsp
14:41Selveste1 has joined #ltsp
14:45wwx has quit IRC
14:48yoshi__ has quit IRC
14:51pmatulis has quit IRC
14:53wwx has joined #ltsp
14:57PMantis has joined #ltsp
14:58
<PMantis>
This has to be something simple, but I's setting up LTSP5 for Myth, and it seems to be stopping at NFS. Last line I see is "Negotiation:"
14:58
Or is that nbd?
14:59
<vagrantc>
NBD
14:59
<PMantis>
Ahhhhh, right. tcpdump shown cisco-sccp, which lists in /etc/services as port 2000
14:59
Ok...
15:00
<vagrantc>
i think the default LTSP port for NBD was picked out of thin air
15:00
<PMantis>
Heh
15:01
netstat -anp | grep 2000
15:01
oops
15:01
Whaaaaaaaa?
15:01
tcp 0 0 0.0.0.0:2000 0.0.0.0:* LISTEN 11685/asterisk
15:01
THERE's the problem!
15:03
rebooting...
15:04
Funny, many times, I fix my own problems, but not until I explain it to someone else do I realize I know the answer. LOL
15:05
Oh, past that, but new error.
15:06
nbd0: unknown partition table
15:06
Done.
15:06
mount calll failed - server replied: Permission denied.
15:07bobby_C has quit IRC
15:11
<PMantis>
I think that's an image problem.
15:12Ahmuck-Sr has quit IRC
15:14johnny has joined #ltsp
15:40Ahmuck has joined #ltsp
15:51mikkel has quit IRC
16:15wimiko1 has quit IRC
16:16
<alkisg>
johnny, vagrantc: the included example dhcpd-dnsmasq included with ltsp-server isn't quite right, could it maybe be replaced with the one I proposed here? https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/428760
16:17
(I'm asking you 'cause I think you're the only devs using dnsmasq...)
16:18
<vagrantc>
alkisg: it worked when i committed it ... but of course we can add improvements to it
16:19
<alkisg>
vagrantc: it works, but lts.conf doesn't reach the clients when one uses nbd
16:19
...because it doesn't send the filename dhcp option to ipconfig
16:20
...and so it can't locate the lts.conf path
16:20
<vagrantc>
ooooh! port=0
16:21
<alkisg>
Yeah I thought it would be nice to have it as a default, saves a lot of problems :)
16:22* vagrantc notes the difference between old and broken dhcp clients
16:23
<alkisg>
Many comments are from the original dnsmasq.conf from the dnsmasq package :)
16:25
<vagrantc>
alkisg: although calling "Linux ipconfig" a pxe client seems to be a hack for working around stupid assumptions in the initramfs scripts
16:26
<alkisg>
I totally agree. :)
16:27
<vagrantc>
alkisg: would seem more clear to specify lts.conf when "Linux ipconfig" is the client, at least on an LTSP network
16:27
<alkisg>
Btw, is rootpath needed for non-netbooted clients? Why do we send the filename only for netbooted clients, but we _always_ send the rootpath?
16:28
<vagrantc>
non-netbooted?
16:28
<alkisg>
dhcp-boot=net:pxe,/ltsp/i386/pxelinux.0 ==> we only send a filename for net: pxe class
16:28
We don't have a similar "if" statement for rootpath...
16:29
<vagrantc>
oh, you mean a non-LTSP machine that happens to be on the network?
16:29
<alkisg>
Yes.
16:29
<vagrantc>
simplicity
16:29
it doesn't hurt anything to send it, does it?
16:29
<alkisg>
Why not do the same for "filename" then?
16:29
Only if the client doesn't have a big enough buffer, but I don't think that's gonna happen..
16:29
<vagrantc>
because at the time i wrote it, it didn't have the ability to set a default filename, and then override it
16:30
<alkisg>
I think the minimum buffer defined by the specs is 512bytes
16:30
<vagrantc>
but if you can write it up so that it does, i'd be fine with that.
16:31
<alkisg>
I'd imagine that the following would work, but it would need testing:
16:31
dhcp-boot=/ltsp/i386/pxelinux.0
16:31
dhcp-boot=net:etherboot,/ltsp/i386/nbi.img
16:31
<vagrantc>
when i tested that, it didn't work. but that was 3 years ago.
16:32
<alkisg>
Anyway, I think I like the dhcp-boot=net:ltsp,/ltsp/i386/lts.conf directive that you just proposed better :)
16:32
...where ltsp==('Linux ipconfig' or whatever udhcp sends....)
16:32* vagrantc suspects "Linux ipconfig" is pretty particular to ipconfig
16:33
<vagrantc>
but yes.
16:33* vagrantc never liked putting lts.conf in tftp anyways
16:34
<alkisg>
I think getting lts.conf is the only reason to have a customized initramfs in ubuntu/ltsp...
16:35
otherwise the stock ubuntu initrd.img works just fine for booting over nfs or nbd...
16:35
So yup, I totally agree again :)
16:35
<vagrantc>
not with the aufs overlay
16:35
the unionfs-type overlay is the main reason it was written
16:36
<alkisg>
Does that really have better performance than mounting nbd as read/write?
16:36
(with the writes discarded, of course...)
16:36
<vagrantc>
how do you discard the writes?
16:36
<alkisg>
nbd-server -c or something
16:36
<vagrantc>
that feature is broken for our needs
16:37
<alkisg>
Ah
16:37
<vagrantc>
there's at least one bug on nbd in debian about it
16:38
<alkisg>
OK... Thanks for your time vagrantc!
16:39
<vagrantc>
i know fedora fetches the lts.conf later than the initramfs ... somewhere in the boot or init scripts
16:39
or at least, they did at one point.
16:41
alkisg: what's the dhcp-no-override thing do? that sounds like a feature i *liked* about dnsmasq's dhcp implementation
16:41
<alkisg>
I tried to make my own script to netboot a client, and I was able to do everything (bind-mounts, hostname etc) from /etc/rcS.d/S00my-script. I think it should be possible to use an S00 script to fetch lts.conf...
16:42
vagrantc: there are 2 ways to send the boot filename, I mean 2 different dhcp options for it
16:42
<vagrantc>
yeah, but aufs requires to be mounted in the initramfs
16:42
<alkisg>
The first option is "embedded" in the dhcp struct. The other one is on an "options list" that follows the struct
16:43
Both should be accepted normally, but ipconfig *doesn't* accept the filename from the options list
16:43
<vagrantc>
but the servername one as well?
16:43
<alkisg>
Yes
16:43
That's also both in the struct and in the options list
16:43
<vagrantc>
so it still defaults to servername normally
16:44
<alkisg>
I had sent a patch to ipconfig for this as well... but I had put a lot of "extras" in that patch and they didn't accept it because it was too big :)
16:47
<vagrantc>
alkisg: committing with some minor changes
16:47
<alkisg>
Merci!!! :)
16:47
<vagrantc>
alkisg: just to reduce the diffs, i reverted some capitalization changes, and left tftp enabled by default
16:48
<johnny>
sweet
16:49
<vagrantc>
alkisg: do you know if the ubuntu script will work with filename=/ltsp/i386/lts.conf ?
16:49
<alkisg>
Yes, it will
16:50
vagrantc: Uhm in ubuntu it's quite difficult to get rid of tftp-hpa...
16:50
<vagrantc>
ah, it should work, yes.
16:50
<alkisg>
I know that ubuntu packaging is not really related to upstream work, but it would be nice if tftp was disabled by default,
16:50
<vagrantc>
alkisg: well, you have to manually configure dnsmasq anyways
16:51
<alkisg>
because dnsmasq won't start because the port is taken...
16:51
<vagrantc>
alkisg: we can disable it later, but i'll keep only relevent changes for this commit
16:51
<alkisg>
...or at least a small note on top, e.g. "other tftp servers should not be running" etc
16:52
Okey :)
16:52
Thanks again!
17:03alkisg has quit IRC
17:26epsas has quit IRC
17:29epsas has joined #ltsp
18:04map7|away has quit IRC
18:07map7 has joined #ltsp
18:14Selveste1 has quit IRC
18:22petre has joined #ltsp
18:37map7 is now known as map7|away
18:45map7|away is now known as map7
18:47garymc has quit IRC
18:56vagrantc has quit IRC
19:15map7 is now known as map7|away
19:20Faithful has joined #ltsp
20:07Faithful has quit IRC
20:08Faithful has joined #ltsp
20:19Ahmuck has quit IRC
20:43Faithful has quit IRC
21:15FGXR6 has joined #ltsp
21:27PMantis has quit IRC
21:29F-GT has quit IRC
21:30FGXR6 has quit IRC
21:30F-GT has joined #ltsp
22:01petre has quit IRC
22:11Ahmuck has joined #ltsp
22:15Ahmuck has quit IRC
22:16muh2000 has joined #ltsp
22:16
<muh2000>
hi
22:16
<sbalneav>
Hello
22:22
<jammcq>
Scotty !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
22:22
<sbalneav>
Hey jammcq!
22:28
<muh2000>
is there a way to install ltsp stuff on debian without pulling dhcp-server?
22:28
have to reinstall...
22:29
<sbalneav>
I beleive if you just install ltsp-server, instead of ltsp-server-standalone.
22:33
<muh2000>
yeah thnx :)
22:45alkisg has joined #ltsp
22:46chupacabra has quit IRC
22:49johnny has left #ltsp
23:06warren has joined #ltsp
23:06
<warren>
hey buddies
23:13
<muh2000>
where can i find current lts.conf parameters?
23:13
<jammcq>
warren: hey
23:14
<warren>
jammcq: hey
23:14
question
23:14
jammcq: I'm building a spam trap network.
23:14
<jammcq>
oh
23:14
<warren>
jammcq: I need people who have their own domain names that are popular web sites, but they do not use MX at all.
23:14
<jammcq>
hmm
23:14
<warren>
jammcq: MX can point at me, so I receive all the spam.
23:15
<jammcq>
how does that work?
23:15
you'd get ALL mail
23:15
<warren>
jammcq: I receive all mail delivered to that domain's MX record
23:16
jammcq: so only point MX at me in cases where the entire domain should have zero ham
23:16
<jammcq>
I think i've got MX's for all the domains that i'm involved with
23:16
<warren>
how many actually use incoming mail on those MX records?
23:16
<jammcq>
I think all of them
23:17
<warren>
oh
23:17
<jammcq>
how else would I receive email at those domains?
23:17
<alkisg>
muh2000: zcat /usr/share/doc/ltsp-server/lts-parameters.txt.gz | less
23:17
<warren>
right... I'm asking more domains not using their own MX
23:17
<muh2000>
thnx
23:17
<jammcq>
but who would have a domain that doesn't want any email?
23:17
<warren>
domains not using their own MX are ideal to point at my spam trap
23:18
a popular website might be popular, but they never turned on thier MX or do mail at that domain name
23:18
<jammcq>
hmm
23:18
<warren>
spammers try to send mail to that domain anyway
23:18
even though it goes nowhere
23:19
jammcq: one of eharrison's mail servers was decomissioned 10 years ago, but it still receives tons of spam
23:20
jammcq: he's going to redirect that monster MX to me so I can rapidly collect spam
23:21
<jammcq>
hmm
23:21
what are you gonna do with the spam?
23:28
<muh2000>
reinstalling ltsp does good wonders :D
23:28yanu has quit IRC
23:29
<warren>
jammcq: statistical analysis to improve spamassassin
23:30alkisg has quit IRC
23:30yanu has joined #ltsp
23:43dkostousov has quit IRC
23:43dkostousov has joined #ltsp
23:44
<dkostousov>
ping
23:53Sarten-X has quit IRC