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


Channel log from 17 April 2012   (all times are UTC)

00:35Parker955_Away is now known as Parker955
00:46gentgeen__ has left IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net, Ping timeout: 246 seconds)
01:16highvolt1ge has joined IRC (highvolt1ge!~highvolta@modemcable104.153-23-96.mc.videotron.ca)
01:18highvoltage has left IRC (highvoltage!~highvolta@ubuntu/member/highvoltage, Read error: Connection reset by peer)
01:21highvolt1ge has left IRC (highvolt1ge!~highvolta@modemcable104.153-23-96.mc.videotron.ca, Ping timeout: 248 seconds)
01:22highvoltage has joined IRC (highvoltage!~highvolta@ubuntu/member/highvoltage)
01:29gentgeen__ has joined IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net)
01:30gentgeen__ has left IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net, Remote host closed the connection)
01:32Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
01:38Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
01:40alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
01:46vagrantc has left IRC (vagrantc!~vagrant@74-92-172-241-Oregon.hfc.comcastbusiness.net, Ping timeout: 240 seconds)
02:14
<alkisg>
Which one would be more handy? 1) SCREEN_07=rsync, to copy the fat client image to a local client disk, or 2) install ltsp-client over an existing installation, and that would add a grub entry for ltsp-style booting?
02:15
...this is to help in cases where the local network is very slow, but central user management + lts.conf options + central homes are desired
02:22* alkisg will probably implement (2)...
02:22
<alkisg>
...and leave the cloning for clonezilla
02:33Parker955 is now known as Parker955_Away
02:38
<alkisg>
Meh clonezilla has a lot of stupid limitations... source partition needs to be <= larger partition, no multicasting in the live cd edition, a DRBL server setup is needed for it.... :-/
02:38
We should tell them to use LTSP instead of DRBL, which can run from a live media too
03:06Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
03:06alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
03:34adrianorg__ has left IRC (adrianorg__!~adrianorg@187.58.155.242, Ping timeout: 260 seconds)
03:42roasted has joined IRC (roasted!6c0be84f@gateway/web/freenode/ip.108.11.232.79)
03:42
<roasted>
D
03:42
:D
03:53intelliant has joined IRC (intelliant!~Shrenik@122.160.53.16)
03:59monteslu_ is now known as monteslu
04:30staffencasa has left IRC (staffencasa!~staffenca@128-193-148-207.oregonstate.edu, Ping timeout: 244 seconds)
05:00alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
05:03monteslu_ has joined IRC (monteslu_!~monteslu@ip68-109-174-213.ph.ph.cox.net)
05:07monteslu has left IRC (monteslu!~monteslu@ip68-109-174-213.ph.ph.cox.net, Ping timeout: 265 seconds)
05:32roasted has left IRC (roasted!6c0be84f@gateway/web/freenode/ip.108.11.232.79, Quit: Page closed)
05:32risca_ has left IRC (risca_!~risca@wi-secure-3075.cc.umanitoba.ca, Quit: Lämnar)
05:32risca has joined IRC (risca!~risca@wi-secure-3075.cc.umanitoba.ca)
05:37risca has left IRC (risca!~risca@wi-secure-3075.cc.umanitoba.ca, Ping timeout: 244 seconds)
06:20bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl)
06:31risca has joined IRC (risca!~risca@wnpgmb0903w-ds01-249-233.dynamic.mtsallstream.net)
06:32bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, Remote host closed the connection)
06:33bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl)
06:38
<alkisg>
I just proposed several changes in the ltsp-developers mailing list, if anyone wants to comment, I'll read the irclogs... http://sourceforge.net/mailarchive/forum.php?thread_name=4F8D0439.8070604%40gmail.com&forum_name=ltsp-developer
06:41
<knipwim>
lot of reading, you've been busy :)
06:42
<alkisg>
I started writing a mail about merging ltsp-pnp, but then I realized that most of its code should be the default anyway, and ltsp-pnp should only become a small tool. ltsp-export-root, optionally used by whoever wants it...
06:43
vagrantc has mentioned about making dnsmasq the default many times already
07:09risca has left IRC (risca!~risca@wnpgmb0903w-ds01-249-233.dynamic.mtsallstream.net, Quit: Lämnar)
07:14
<alkisg>
bbl, will read logs
07:15alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
07:39bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at)
08:01dobber has joined IRC (dobber!~dobber@213.169.45.222)
08:24
<Hyperbyte>
mmmhmmm
08:47Gremble has joined IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com)
08:50monteslu_ has left IRC (monteslu_!~monteslu@ip68-109-174-213.ph.ph.cox.net, Ping timeout: 265 seconds)
08:54monteslu_ has joined IRC (monteslu_!~monteslu@ip68-109-174-213.ph.ph.cox.net)
08:57alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
09:02
<alkisg>
knipwim: if I want to have a quick look at how gentoo organizes /etc, e.g. install dnsmasq and look at its default configuration, would you recommend the gentoo live cd, the sabayon live cd...?
09:29ogra_ has left IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de, Read error: Operation timed out)
09:32ogra_ has joined IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de)
09:41bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, Ping timeout: 276 seconds)
10:04Trixboxer has joined IRC (Trixboxer!~Trixboxer@115.124.115.71)
10:11bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl)
10:45Briareos1 has joined IRC (Briareos1!~quassel@94.136.98.13)
10:46Briareos1 has left IRC (Briareos1!~quassel@94.136.98.13)
10:48[GuS] has left IRC ([GuS]!~gustavo@unaffiliated/gus/x-663402, Remote host closed the connection)
10:52
<alkisg>
stgraber: isn't putting "1.2.3.4 server" in /etc/hosts enough for ssh to connect by hostname? Why is a DNS entry required? (also replied in the ML)
11:07
<stgraber>
alkisg: it'd be indeed, though then it really wouldn't make any difference from what we currently have as you'd still need to rebuild the nbd image when you change your server's ip
11:07
<alkisg>
stgraber: that entry is automatically created on client boot
11:08
Also, I just tested kvm + dnsmasq vs tftpd-hpa, both need 4 seconds for a 15MB initrd
11:08
<stgraber>
alkisg: also, ssh tracks host => ip records and stores the key for both the host and ip, if any of the two changes it re-prompts, so I'm not sure what difference it'd make
11:08
<alkisg>
On the next client boot, the IP entry won't be there
11:08
<stgraber>
hmm, I'd have to check my exact kvm config but I can reliably reproduce the issue for at least a year here ;)
11:09
<alkisg>
So once that part is solved, we're good to go for making ltsp-server-standalone recommend: dnsmasq in Ubuntu? :)
11:09
I think vagrantc was eager for that too...
11:10
<stgraber>
right, so ssh will just add the match then, I guess that'd work, not sure if our expect script is picky about the extra line being printed by ssh but I guess not
11:10
well, before moving to dnsmasq we still need to see how we deal with upgrades ;)
11:10
<alkisg>
It's not, I've tried it with LDM_SERVER=server + changing my server IP, it works fine
11:10
<stgraber>
anyway, got to run
11:11
<alkisg>
To test kvm, I did the following: install kvm, run kvm -boot order=n, press ctrl+b in ipxe because I don't have bridged networking etc configured, set next-server, run imgfetch initrd
11:11
Both with dnsmasq and kvm I got 4 secs
11:11
About the upgrade path, I think we can solve that in the packaging, when 12.10 starts being developed
11:12
Thanks, bb
11:12
<stgraber>
alkisg: for kvm, both client and server are using virtio. Client is using ipxe
11:12
alkisg: I can reproduce it every time when booting from ltsp-live with an edubuntu dvd (using dnsmasq)
11:12
<alkisg>
So I boot one kvm guest with the edubuntu live, and netboot another?
11:12
OK, will try
11:13
<stgraber>
yeah, boot the edubuntu dvd, start ltsp-live from there, then boot a client. That's the setup I have here that's hanging.
11:13
in the exact same setup, post-install (so using dhcp-server then), it boots immediately as expected
11:13Gremble has left IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com, Quit: I Leave)
11:15Gremble has joined IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com)
11:22markit has joined IRC (markit!~marco@151.78.74.112)
11:25
<markit>
hi, kubuntu 11.04, only thin clients, recently I've updated the system and installed hplip-gui to better control the multifuncion hp laser printer connected with usb to the server. After that sometime a children prints (I suppose from writer or impress) and ... THE SERVER FROZES (reisub is the only way to solve, seems)
11:25
I've removed the hplip-gui but the problem stays :(
11:26
anyone heard of such a problem and has a solution? and in a more general way.. how can froze the whole server??? (also thin clients are frozen at the same time)
11:26
this recalls me Windows 3.1 stability and task isolation.....
11:26
(server 64 bit, chroot 32, it it matters)
11:38
<alkisg>
Check dmesg?
11:38
If a kernel module crashes, of course it takes the whole kernel+server with it
11:39Parker955_Away is now known as Parker955
11:50
<vmlintu>
alkisg: with dnsmasq we experienced problems in bigger setups as it could not handle more than one dhcp request at a time
11:51
alkisg: when it receives a dhcp request, it pings the ip it's about to give out for 3 seconds. During this time it does not handle other requests
11:51
<alkisg>
vmlintu: that's quite strange, when did you have that problem? With recent versions of dnsmasqs, or very older?
11:51
Ah, that's the default for dhcpd too
11:51
You can disable that in both
11:51
Ooops
11:51
"it does not handle other requests" ??
11:51
<vmlintu>
isc dhcpd can handle multiple pings in parallel
11:52
<alkisg>
That would be a very big problem, I wonder why Simon hasn't solved it yet... when did you experience this?
11:52
<vmlintu>
When schools start in the morning, we had in some cases 100 thin clients booting within minutes
11:52
This is with 10.04
11:53
Quite a few PXE roms wait only ~15 seconds before giving up
11:54
So with dnsmasq there was one dhcp response every 3 seconds, but isc dhcpd could handle them in parallel
11:54
<alkisg>
Could you ask in the dnsmasq list about that? Usually Simon replies within a few days, and he's very helpful, in some cases he solved problems upstream immediately
11:55
There's a --no-ping flag of course, but that's not the optimal configuration...
11:56
Or, have you seen any bug reports about that problem?
11:56gentgeen__ has joined IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net)
11:56
<vmlintu>
let me check if I have any notes left
11:58
<alkisg>
Ah, found it in http://www.thekelleys.org.uk/dnsmasq/docs/FAQ
11:58
The wait for a reply is between two and three seconds. Because the DHCP server is not re-entrant, it cannot serve other DHCP requests during this time. To avoid dropping requests, the address probe may be skipped when dnsmasq is under heavy load.
11:59
<vmlintu>
yes, I just found the same document
12:00adrianorg__ has joined IRC (adrianorg__!~adrianorg@187.58.155.242)
12:00
<alkisg>
So yeah this is a significant problem... if it did what it claims, i.e. at least not ping under heavy load, that would be workable
12:00
E.g. only "cache" 2 dhcp requests, not more => 5-6 seconds
12:01
<vmlintu>
We were using dnsmasq for some time, but switched back to isc dhcpd because of this
12:02
We turned off the pings when we were using static IPs for the thin clients, but now that we use dynamic addresses + ddns, we really need the ping checks
12:03
Nowadays the ltsp servers are also serving separate wlan vlans that have a lot of dhcp requests all the time
12:05
<alkisg>
Here we're using a router as the dhcp server, and dnsmasq only sends the boot filename (proxydhcp mode), so that wasn't ever a problem
12:05
But it is indeed a significant problem, we probably need a separate package for dnsmasq, and leave the default to isc-dhcp for that problem
12:08Gremble has left IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com, Quit: I Leave)
12:08
<vmlintu>
Most installations can probably use dnsmasq with no problems as also most of our installations worked without problems
12:09
We are still using dnsmasq for DNS (with bind for ddns)
12:10
<alkisg>
vmlintu: on second thought, the clients are supposed to quickly answer the ping request, in just a few msec
12:10
So that 2-3 sec delay would only be there if the client doesn't reply, i.e. if the previous lease was too big (e.g. still valid from yesterday)
12:11
Maybe the problem could be solved by using a smaller lease, e.g. 8h? /me looks up the default...
12:12
"If not given, the default lease time is one hour." ==> do you remember what lease time did you declare when the problem happened?
12:14
<vmlintu>
let me check if I can find it from the repository
12:15
actually, I think it sends the pings for every request to make sure that nobody is really using it
12:15
<alkisg>
Btw, you wouldn't mind using dnsmasq as the dns+tftp server, alongside isc-dhcp, right?
12:16
<vmlintu>
We have a wrapper for tftpd-hpa that dynamically gets configuration from LDAP and that cannot be done with dnsmasq if I remember correctly
12:17
<alkisg>
Ouch
12:17brunolambert has joined IRC (brunolambert!blambert@nat/revolutionlinux/x-iahbccplalhtqpwq)
12:18
<vmlintu>
basically the wrapper reads configuration from LDAP and creates client specific PXE boot files
12:18
This way we can configure e.g. kernel parameters in LDAP
12:18[GuS] has joined IRC ([GuS]!~MysT@213-117-16-190.fibertel.com.ar)
12:18[GuS] has joined IRC ([GuS]!~MysT@unaffiliated/gus/x-663402)
12:20
<vmlintu>
Looks like dnsmasq was using the default lease time
12:20
<alkisg>
dnsmasq does have a --dhcp-script option, maybe that could be used, but then you'd hit the other dhcp problem
12:21
--tftp-max=<connections>
12:21
Set the maximum number of concurrent TFTP connections allowed. This defaults to 50.
12:21
...maybe you were hitting that, instead of the 2 secs problem?
12:22
<vmlintu>
we never used dnsmasq for tftp
12:23
<alkisg>
...so it appears that a separate package will be needed, that defaults to using dnsmasq, instead of putting that as the default. Thank you vmlintu, most appreciated.
12:39mgariepy has joined IRC (mgariepy!mgariepy@nat/revolutionlinux/session)
12:39mgariepy has joined IRC (mgariepy!mgariepy@nat/revolutionlinux/x-gulwnkcvzivnkoaw)
12:39mgariepy has joined IRC (mgariepy!mgariepy@ubuntu/member/mgariepy)
12:40
<alkisg>
Meh, another incompatibility between tftpd-hpa and dnsmasq... if I omit "--secure" from tftpd-hpa configuration, then I have to specify the whole tftp path, /var/lib/tftpboot/ltsp/i386/vmlinuz, instead of the relative, /ltsp/i386/vmlinuz...
12:42khildin has joined IRC (khildin!~khildin@ip-80-236-223-35.dsl.scarlet.be)
12:45wim_ has joined IRC (wim_!~chatzilla@WEGC203035.UNI-GRAZ.AT)
12:51
<markit>
alkisg: back now. I'm not on the pc and hte teacher has left, in any case is not a crash but a frozen. So maybe HURD is not that a bad idea :)
13:03* markit reading about dhsmasq problem... I've only 28 clients max, but thinks "faster is better" :)
13:08markit has left IRC (markit!~marco@151.78.74.112, Quit: Konversation terminated!)
13:15dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
13:20brunolambert has left IRC (brunolambert!blambert@nat/revolutionlinux/x-iahbccplalhtqpwq, Ping timeout: 245 seconds)
13:20alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
13:25bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, Quit: Leaving.)
13:29artista-frustrad has joined IRC (artista-frustrad!~fernando@200.247.43.2)
13:32brunolambert has joined IRC (brunolambert!blambert@nat/revolutionlinux/x-zivsorcbhfwgwmml)
13:35
<dobber>
!usb
13:35
<ltsp`>
dobber: usb: LTSP only supports usb filesystem devices, for other projects like http://usbip.sourceforge.net/ probably this isn't the correct channel to ask, not many people have mentioned it
13:37
<dobber>
is there a documentation for booting ltsp over usb stick
13:39wim_ has left IRC (wim_!~chatzilla@WEGC203035.UNI-GRAZ.AT, Remote host closed the connection)
13:40
<||cw>
dobber: I think you can get etherboot on USB sticks
13:40
<dobber>
what is etherboot ?
13:41
<||cw>
basically PXE loader for systems that don't have it built-in
13:41
it has ROM versions for nic that need that, and hdd and floppy versions
13:41
which I think can be used for cdrom and usb too
13:42
<dobber>
cool that, but my fat client image is 1.6gb ;(
13:42
<||cw>
http://www.rom-o-matic.net/etherboot/etherboot-5.4.4/contrib/rom-o-matic/
13:42
<dobber>
can't use that outside the office
13:42
<||cw>
oh you want to boot your image via usb?
13:42Phantomas has joined IRC (Phantomas!~Phantomas@athedsl-130374.home.otenet.gr)
13:43Phantomas has left IRC (Phantomas!~Phantomas@athedsl-130374.home.otenet.gr, Changing host)
13:43Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
13:43
<dobber>
i want to boot ltsp image over usb - outside of lan.
13:43
<||cw>
hm. I don't know how feasible that is
13:43
<dobber>
also i'll have to hack the network up process to start vpn
13:43
<||cw>
well, then t's not LTSP anymore, it's just an image
13:43artista-frustrad has left IRC (artista-frustrad!~fernando@200.247.43.2, Ping timeout: 276 seconds)
13:43
<dobber>
yes
13:43
it's the ltsp image
13:44
<||cw>
so basically, you want a liveusb and an ltsp in one image
13:44
<dobber>
yes
13:44
<||cw>
where's the fat client writable mount going to come from?
13:44artista-frustrad has joined IRC (artista-frustrad!~fernando@200.247.43.2)
13:45
<||cw>
on usb too? to sync back to the server how?
13:45
<dobber>
i'll have to start a vpn, the mount is from central server
13:45
<||cw>
more I think about it, the methods are too different
13:45
<dobber>
cw
13:46
<||cw>
it would probably be easier to just maintain 2 images. once setup, it doens't take much effort
13:47
<dobber>
the usb stick should be the client image - like /opt/images/oneiric.img
13:47
./opt/ltsp/images/
13:47
so boot the image from usb, instead of tftpd
13:49
<||cw>
your starting point will need to be a howto for making your own liveusb, then see if it'll take the ltsp image instead of the casper/squashfs
13:49
it might be just a matter of editing the syslinux.cfg to point at your image
13:49
<dobber>
ok
13:49
i've actually never made liveusb
13:50
so it's a goot time to stare at the docs
13:50
<||cw>
then you'll need scripts in your image to detect when it's usb booted, or just not on your home lan, to start the vpn
13:51
I've never made a custom liveusb either, but I have edited syslinux.cfg's to change behaviors and it's pretty easy
13:51
<dobber>
ok, i'm looking at ti now
13:53
http://live.learnfree.eu/
13:56
one last question
13:56
what partition type/fs should i use
14:00
<||cw>
on what, the usb stick? fat32
14:01
that url seems like it'll only build one from the repo's for you, you'll need manual instructions
14:01
<dobber>
are there symlinks on fat32 ?
14:01
<||cw>
no
14:01
you are puting the img file on it
14:01
what distro are you using?
14:02
<dobber>
ubuntu lucid on the server
14:02
oneiric no fat client
14:02
cw, you mean, i will just copy the squashfs image to the usb ?
14:03
<||cw>
that's the idea
14:03
<dobber>
cook
14:03
cool
14:04
<||cw>
and a kernel and initrd, then you setup syslinux to load the kernel and "mount" the img
14:04
it actually works a lot like pxe booting, and syslinux and pxelinux I think were derived from the same project
14:05khildin has left IRC (khildin!~khildin@ip-80-236-223-35.dsl.scarlet.be, Ping timeout: 276 seconds)
14:05
<||cw>
I could be wrong, but it's possible that you can simply copy your image over top of the normal liveusb img, renaming to keep the name, and it may work
14:07
<dobber>
all documents i found are for OS isos :(
14:14
<andygraybeal>
i got an 84 day uptime on my system, one of my best :)
14:14
well, ltsp system, my firewall never gets shutdown
14:15
<||cw>
andygraybeal: no security updates needed?
14:15
<andygraybeal>
eh, probably
14:15
<||cw>
rare to go that long with no kernel update
14:15
<andygraybeal>
yea, i need to reboot for kernel udpates
14:16
<||cw>
unless you're running something so old it no longer gets updates :)
14:16
<dobber>
i've got a few 1300s and a few close behind
14:16
<andygraybeal>
||cw, it's 10.04 so i still get updates, and yes your correct -- i just put the reboots off!
14:17
<||cw>
I do sometimes too, but it often leads to instability
14:17
especially when libc gets an update with it
14:17
<andygraybeal>
yea, my kvm/libvirt starts to cough at me.
14:18
<dobber>
you lucky guys
14:18
i'm still fighting with sarge machines
14:19
<||cw>
ha
14:19
I still have a couple sarges
14:20
<andygraybeal>
ah, that is before my time
14:20
<||cw>
one's a fax server, stupid old multiport fax driver doens't work quite the same on 2.6 kernels
14:20
<andygraybeal>
errr.. what was 8.10?
14:20
er no 8.04
14:20adrianorg_ has joined IRC (adrianorg_!~adrianorg@177.18.49.236)
14:21khildin has joined IRC (khildin!~khildin@ip-80-236-223-35.dsl.scarlet.be)
14:21
<dobber>
andygraybeal: it's before ubuntu
14:21
<||cw>
big time
14:21
<andygraybeal>
aah okay
14:21adrianorg__ has left IRC (adrianorg__!~adrianorg@187.58.155.242, Read error: Connection reset by peer)
14:21
<andygraybeal>
yea, debian sarge :)
14:21
<||cw>
sarge came with a 2.4 kernel initially, then had a 2.4 or 2.6 install option
14:21
or did it start out with the option?
14:21
man that was a long time ago
14:22
<andygraybeal>
i started with 8.04 :)
14:23
you guy sare old!
14:23
i actualy might be older than you age wise though :P
14:23
<||cw>
this fax server seems to have been installed in 2007
14:23
uh, yeah...
14:23
maybe
14:23
my first linux was a 2.2 kernel
14:23
slackware even
14:24
<dobber>
my first was a slackware 3.1, back in 98
14:24
it had a name for the root account
14:24
root of all evil
14:24
in the passwd file
14:35mgariepy has left IRC (mgariepy!mgariepy@ubuntu/member/mgariepy, Ping timeout: 245 seconds)
14:36brunolambert has left IRC (brunolambert!blambert@nat/revolutionlinux/x-zivsorcbhfwgwmml, Ping timeout: 245 seconds)
14:36Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas, Ping timeout: 252 seconds)
14:40brunolambert has joined IRC (brunolambert!blambert@nat/revolutionlinux/x-waefcshooegqbrel)
14:40
<wafflecone>
redhat 5.2
14:41mgariepy has joined IRC (mgariepy!mgariepy@ubuntu/member/mgariepy)
14:43
<wafflecone>
used it as a router for my home network ... and a workstation ... what was it before ipchains? ipwad or something?
14:47
<andygraybeal>
i bought redhat, probably sometime in 98... i had the actual box.
14:47
i think i also bought freebsd back then too or some type of bsd.
14:48
about 17 yrs ago
14:48
or 15 or somethign like this.
14:48
time fades
14:51
<dobber>
redhat box ?
14:52
you mean the server or the paper box
14:53
http://farm3.staticflickr.com/2746/4451716090_e4fa6ab0a5_z.jpg
14:53Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
14:58
<dobber>
||cw: i managed to boot from usb into ltsp
14:59
<andygraybeal>
dobber, paper box!
14:59
cardboard
15:00
<dobber>
i'd love to have a few of these on my desk
15:00
<andygraybeal>
at the time, you jus tthink they are in the way :)
15:00
<||cw>
yeah? what's the trick?
15:01
when you have a CRT, they are
15:02
<dobber>
no trick, just craft a syslinux.cfg and put the images
15:02Gremble has joined IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com)
15:09
<||cw>
cool
15:12
<dobber>
ah there is a trick
15:12
you have to use syslinux from the chroot
15:13bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Ping timeout: 276 seconds)
15:15Trixboxer has left IRC (Trixboxer!~Trixboxer@115.124.115.71, Quit: "Achievement is not the end, its the beginning of new journey !!!")
15:21staffencasa has joined IRC (staffencasa!~staffenca@128-193-148-207.oregonstate.edu)
15:41dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection)
16:06bengoa has joined IRC (bengoa!~bengoa@2001:1291:229:2:216:cbff:feab:6cc9)
16:11intelliant has left IRC (intelliant!~Shrenik@122.160.53.16, Quit: Leaving.)
16:26dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Leaving...)
16:31dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
16:43intelliant has joined IRC (intelliant!~Shrenik@59.93.202.137)
16:49dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Computer has gone to sleep.)
16:49Lns has left IRC (Lns!~Lns@pdpc/supporter/professional/lns, Quit: Leaving)
16:59alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
17:03monteslu__ has joined IRC (monteslu__!~monteslu@ip68-109-174-213.ph.ph.cox.net)
17:04dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
17:07intelliant has left IRC (intelliant!~Shrenik@59.93.202.137, Read error: Connection reset by peer)
17:07monteslu_ has left IRC (monteslu_!~monteslu@ip68-109-174-213.ph.ph.cox.net, Ping timeout: 265 seconds)
17:09intelliant has joined IRC (intelliant!~Shrenik@59.93.200.73)
17:42khildin has left IRC (khildin!~khildin@ip-80-236-223-35.dsl.scarlet.be, Quit: I'm gone, bye bye)
17:44vmlintu has left IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi, Quit: Lhdss)
18:11Gremble has left IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com, Quit: I Leave)
18:19Phantomas1 has joined IRC (Phantomas1!~Phantomas@unaffiliated/phantomas)
18:19Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas, Disconnected by services)
18:19Phantomas1 is now known as Phantomas
18:34vagrantc has joined IRC (vagrantc!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net)
18:36mhterres has joined IRC (mhterres!~marcelo@2001:1291:229:2:20c:6eff:fe23:acbc)
18:36
<mhterres>
hey guys
18:37vagrantc_ has joined IRC (vagrantc_!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net)
18:37vmlintu has joined IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi)
18:42staffencasa_ has joined IRC (staffencasa_!~staffenca@128-193-148-207.oregonstate.edu)
18:42vagrantc_ has left IRC (vagrantc_!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net, Quit: leaving)
18:43staffencasa_ has left IRC (staffencasa_!~staffenca@128-193-148-207.oregonstate.edu, Read error: Connection reset by peer)
18:45staffencasa has left IRC (staffencasa!~staffenca@128-193-148-207.oregonstate.edu, Ping timeout: 244 seconds)
18:46staffencasa has joined IRC (staffencasa!~staffenca@128-193-148-207.oregonstate.edu)
18:47rthomson has left IRC (rthomson!~rthomson@mars.pet.ubc.ca, Read error: Connection reset by peer)
18:56risca has joined IRC (risca!~risca@wnpgmb0903w-ds01-249-233.dynamic.mtsallstream.net)
19:00* alkisg filed a bug about tftpd-hpa not working correctly with symlinks (--secure omitted): http://www.syslinux.org/archives/2012-April/017487.html
19:02
<vagrantc>
oh fun
19:02
it would be so nice to be able to just use symlinks.
19:03
i use atftp or dnsmasq on a few servers to be able to use symlinks more easily
19:03* ||cw <3 dnsmasq
19:03
<vagrantc>
alkisg: overall, i'd welcome the switch to dnsmasq.
19:04
<alkisg>
vagrantc: unfortunately vmlintu mentioned a severe problem: http://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/2012q2/005767.html
19:07
:( hpa reply from #syslinux: (10:04:22 μμ) alkisg: hpa: what I'm looking for is to export /var/lib/tftpboot while allowing symlinks to point outside that dir, but I don't want to have to specify /var/lib/tftpboot from my clients. If I understood correctly, that's neither possible nor in the future plans, right?
19:07
(10:04:55 μμ) hpa: Correct
19:07
So no symlinks possible either, if we don't want to have to specify /var/lib/tftpboot in the client command line
19:09* alkisg now thinks that a new ltsp-server-dnsmasq package would be better instead of moving ltsp-server-standalone from isc-dhcp/tftpd-hpa to dnsmasq... :-/
19:10
<vagrantc>
isc-dhcp-server also does the ping trick ... although it still handles requests better?
19:10
<alkisg>
It's reentrant, so it can handle multiple pings in parallel
19:12
<vagrantc>
that's unfortunate...
19:12
although, when it switches to it's failure mode, it merely doesn't check for the ip in use already ... or it actually fails to work?
19:13
<alkisg>
vmlintu said that the clients wouldn't boot
19:13
So I'm guessing the failure mode doesn't work so well, it loses packets
19:13
<vagrantc>
if the failure mode had a configurable threshold and you could get it to kick in easier
19:13
<elias_a>
alkisg: Do I understand correctly that vmlintu is also testing your stuff?
19:14
<alkisg>
elias_a: no, he just had a bad experience with dnsmasq and reported it
19:14
<elias_a>
alkisg: Ok. Thanks!
19:15
alkisg: I am just interested in how you two get along.
19:15
<alkisg>
vagrantc: so basically you'd be OK if I started committing the stuff I mentioned in the ML, but with the additional change that dnsmasq would now be a different package instead of the default one?
19:15
<elias_a>
The benefit would imho be mutual....
19:18
<vagrantc>
alkisg: i think it could still be useful in trunk, yes.
19:19
alkisg: though it might still make sense as it's own source package.
19:19
depends on what you want to do with it.
19:19
<alkisg>
The ltsp-pnp "essense" is the ltsp-export-root tool, which wouldn't harm if it was part of ltsp-server anyway,
19:20
so for the new package I was asking for, I basically want to have an easy way to install ltsp in proxydhcp mode, which is frequently used
19:20
The rest, symlinks, tftpd replacement etc, are just "nice to have", nothing important
19:21
<vagrantc>
sure.
19:21
<alkisg>
(10:19:03 μμ) alkisg: What new does syslinux 4.10 offer?
19:21
(10:19:11 μμ) hpa: http/ftp support
19:21
(without gpxelinux.0, sounds cool!)
19:21
<vagrantc>
nice.
19:25brunolambert has left IRC (brunolambert!blambert@nat/revolutionlinux/x-waefcshooegqbrel, Ping timeout: 245 seconds)
19:26risca has left IRC (risca!~risca@wnpgmb0903w-ds01-249-233.dynamic.mtsallstream.net, Quit: Lämnar)
19:32intelliant has left IRC (intelliant!~Shrenik@59.93.200.73, Quit: Leaving.)
19:38brunolambert has joined IRC (brunolambert!blambert@nat/revolutionlinux/x-ybwwtfyxphratlkf)
19:49[GuS] has left IRC ([GuS]!~MysT@unaffiliated/gus/x-663402, Remote host closed the connection)
20:01alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Ping timeout: 260 seconds)
20:07mhterres has left IRC (mhterres!~marcelo@2001:1291:229:2:20c:6eff:fe23:acbc, Quit: Leaving)
20:08
<vmlintu>
alkisg: we didn't do much testing with the failure mode as the networks always seem to have clients that use an IP without a lease for it
20:09
e.g. printers
20:09
<elias_a>
vmlintu: Hi. Shall we meet in Hämeenlinna?
20:11
<vmlintu>
elias_a: I'll be attending ITK at least on Thursday
20:12
<alkisg>
vmlintu: I think the problem isn't there when another client has the lease; it's the opposite, if no client has the lease (e.g. when the clients first boot in the morning), then noone answers the ping, so dnsmasq waits 2 secs for all of them (and obviously goes back to the problematic failure mode)
20:13
<elias_a>
vmlintu: Ok. See you there!
20:14
<vagrantc>
alkisg: ah, in our thin client environment, the leases tend to last beyond the expire time?
20:19
<vmlintu>
alkisg: yes.. the failure mode can cause all kinds of problems if someone is already using the IP that dnsmasq would be handing out
20:24alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47)
20:31
<alkisg>
vagrantc, vmlintu, no what I'm saying is that the problem mostly exists when no clients have that lease nor that IP address
20:31
E.g. one first starts dnsmasq and starts a single client
20:31
dnsmasq will ping for the lease
20:31
And since noone will reply, it'll take 2-3 seconds
20:32
If a client has that IP address and replies (no matter if it has a lease or not), then it should be almost instant, but of course dnsmasq would still ping for the next lease it would try
20:33alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Remote host closed the connection)
20:33alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47)
20:35
<alkisg>
vmlintu: vagrantc: answer from the dnsmasq developer: " Why is using --no-ping risky for LTSP? DHCP clients will still do address-in-use checks and dnsmasq handles DHCPDECLINE messages resulting from those checks happily. I've had reports of successfully netbooting hundreds of machines with --no-ping set."
20:35
<vagrantc>
sounds like an answer to me.
20:36Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas, Ping timeout: 248 seconds)
20:37
<alkisg>
If the clients do an address-in-use check, then yup sounds good to me too
20:37* vagrantc wonders if that's what made dnsmasq slow for some kvm users
20:37* alkisg checks the dhcp rfc...
20:38
<alkisg>
And about the cache poisoning problem, that only is an issue for local machine resolving, so we can still use it for thin client but not for their sessions (unless someone is willing to take the risk, where he could manually configure it)
20:39
<vagrantc>
how big of a problem is the cache poisoning? can setting some of the ttl variable options mitigate it?
20:39alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Ping timeout: 244 seconds)
20:40
<alkisg>
A script running in localhost can run netstat | grep dnsmasq, see the source port it uses for an upstream dns request, and answer it instead of letting the upstream server answering it
20:41
So if that script runs netstat every second, the ttl should be even lower....
20:41
<vmlintu>
would be interesting to know how many PXE roms check if the address is in use
20:44alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47)
20:44
<alkisg>
From the DHCP RFC: "when allocating a new address, servers SHOULD check that the offered network address is not already in use" (not MUST).
20:44
The client SHOULD perform a final check on the parameters (e.g., ARP for allocated network address). If the client detects that the address is already in use (e.g., through the use of ARP), the client MUST send a DHCPDECLINE message to the server and restarts the configuration process.
20:44
And of course dnsmasq honors that DHCPDECLINE message... but yeah that still doesn't tell us if all PXE stacks do that
20:44* alkisg checks with gpxe first...
20:45
<vagrantc>
could test by setting up a bunch of static ip addresses, and setting up dnsmasq to give out leases on all those ip addresses.
20:47
<alkisg>
The client should ping no matter if the IP already exists or not, so /me is testing with a single client, his router as the dhcp server, and wireshark...
20:48
<vagrantc>
won't test if the decline works on both ends
20:48
<alkisg>
vmlintu: if you are ever given the chance, it would be nice if you could check if with --no-ping your network is still stable
20:49
True, but I don't worry much about the dnsmasq side, only about the client side... and basically I don't want to mess up with my local network while testing... :D
20:51
<vagrantc>
some testing is generally better than none :)
20:51
<vmlintu>
alkisg: all production systems are now running with isc dhcpd, so testing dnsmasq is quite difficult there
20:52alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Remote host closed the connection)
20:52Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
20:52alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47)
21:10bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at)
21:12brunolambert has left IRC (brunolambert!blambert@nat/revolutionlinux/x-ybwwtfyxphratlkf, Quit: brunolambert)
21:15Roasted has joined IRC (Roasted!~jason@unaffiliated/roasted)
21:15
<Roasted>
pardon me
21:15
!gnome-classic
21:15
<ltsp`>
Roasted: gnome-classic: To select gnome-classic as your default session, put this line in your lts.conf: LDM_SESSION="gnome-session --session=gnome-classic"
21:15
<Roasted>
:)
21:19
<Hyperbyte>
All hail to the ltspbot!
21:26alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Read error: Connection reset by peer)
21:26alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
21:27artista-frustrad has left IRC (artista-frustrad!~fernando@200.247.43.2, Quit: Leaving)
21:34alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Ping timeout: 264 seconds)
21:34
<Roasted>
gotta love it!
21:48alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 244 seconds)
21:56markit has joined IRC (markit!~marco@88-149-177-66.staticnet.ngi.it)
22:08dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Leaving...)
22:12bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Ping timeout: 246 seconds)
22:18alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
22:27
<alkisg>
vagrantc, vmlintu, bad news, my tests show that the clients don't check if the assigned address is in use, so using --no-ping is risky. But let's wait for Simon's reply to make sure.
22:31
<markit>
hi alkisg :) btw, my bug about squid3 being killed by dnsmasq seems has been confirmed
22:40
<alkisg>
Hi markit, nice, hope they solve it for you :)
22:53risca has joined IRC (risca!~risca@wi-secure-3075.cc.umanitoba.ca)
23:36alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
23:44bengoa has left IRC (bengoa!~bengoa@2001:1291:229:2:216:cbff:feab:6cc9, Quit: Leaving.)
23:59loather has joined IRC (loather!~khudson@wsip-98-175-250-115.sd.sd.cox.net)