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


Channel log from 30 December 2013   (all times are UTC)

00:10lifeboy has left IRC (lifeboy!~roland@105-236-159-32.access.mtnbusiness.co.za, Quit: Ex-Chat)
00:25gdi2k has joined IRC (gdi2k!~gdi2k@222.127.254.113)
00:26freedomrun has left IRC (freedomrun!~freedomru@unaffiliated/freedomrun, Quit: So long and thanks for all the fish)
01:17Phantomas1 has joined IRC (Phantomas1!~Phantomas@ubuntu/member/phantomas)
01:18Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 265 seconds)
01:42gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com)
01:47gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 252 seconds)
02:01Phantomas1 has left IRC (Phantomas1!~Phantomas@ubuntu/member/phantomas, Quit: Leaving.)
02:21
<gdi2k>
I have fat clients set up on 12.04, but performance is really not great when starting new applications - for example, opening libreoffice writer takes about 27 seconds. is there any way I can improve that?
02:21
(performance on 13.10 was much better, but it was very buggy)
02:25
network is 1gb from server to switch, 100mb from switch to client. I am seeing data being pulled from the server at a rate of 10 mb / sec, which isn't that far off wire speed
02:29Enslaver has left IRC (Enslaver!~Enslaver@fedora/Enslaver, Read error: Connection reset by peer)
02:32Enslaver has joined IRC (Enslaver!~Enslaver@fedora/Enslaver)
02:59gdi2k has left IRC (gdi2k!~gdi2k@222.127.254.113, Remote host closed the connection)
03:45gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com)
03:49gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 252 seconds)
03:55alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
04:00
<alkisg>
Good morning
04:14* vagrantc waves
04:14
<vagrantc>
alkisg: should i hold off, or upload ltsp as is?
04:14
i remembered, and then forgot, one more feature...
04:14
<alkisg>
vagrantc: meh I just sent a launchpad build, I should have waited for you to tag it :D
04:15
There was udhcp, but for a later release
04:15
!alkisg-todo
04:15
<ltsp>
alkisg-todo: (#1) ltsp-update-image.symlinks, or (#2) upload fixed nbd-server package in the PPA, or (#3) implement LDM_[PRE|IN|POST]SESSION_CMD, or (#4) put IPAPPEND support in ltsp-update-kernels, or (#5) prefer kernels from img rather than dir, or (#6) LDM password hash, or (#7) check remoteapps in fat clients, or (#8) launch ltspfsd/cdpinger upon first media insertion, or (#9) Check ltsp bugs in debian.org, or (#10) (1 more message)
04:16
<vagrantc>
!vagrant-todo
04:16
<ltsp>
vagrant-todo: (#1) pxemenus, or (#2) tweak debian/rules to install appropriate update-kernels.conf, or (#3) update-kernels fix memtest support or drop it
04:16
<vagrantc>
!forget vagrant-todo 3
04:16
<ltsp>
The operation succeeded.
04:16
<vagrantc>
!forget vagrant-todo 2
04:16
<ltsp>
The operation succeeded.
04:17
<vagrantc>
!vagrantc-todo
04:17
<ltsp>
vagrantc-todo: Merge LinuxMint changes... (Subliminal)
04:17
<vagrantc>
!forget vagrantc-todo
04:17
<ltsp>
The operation succeeded.
04:17
<vagrantc>
!vagrantc-todo
04:17
<ltsp>
I do not know about 'vagrantc-todo', but I do know about these similar topics: 'vagrant-todo'
04:17
<vagrantc>
!vagrant-todo
04:17
<ltsp>
vagrant-todo: pxemenus
04:17
<vagrantc>
that's a work in progress, but made some decent progress
04:19
<alkisg>
vagrantc: I could enable translation commits for a few hours in case there were translation commits... although I think we did that recently so there won't be many, not worth it
04:20
<vagrantc>
yeah, not worth it
04:22
<alkisg>
!alkisg-todo
04:22
<ltsp>
alkisg-todo: (#1) upload fixed nbd-server package in the PPA, or (#2) implement LDM_[PRE|IN|POST]SESSION_CMD, or (#3) LDM password hash, or (#4) Check ltsp bugs in debian.org, or (#5) nbd-disconnect isn't called at all because the cmdline changed, no nbdroot=
04:22* alkisg checks bugs at debian.org...
04:25
<alkisg>
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=718727 is probably the same as my request to use binary mode in tftp in launchpad, which sbalneav turned down...
04:25
<vagrantc>
alkisg: you got my main regarding aoe in initramfs-tools?
04:25
<alkisg>
mail, yup
04:25
Thanks, I'll subscribe to that bug
04:26
<vagrantc>
the proposed patch was a lot more complicated than what you have
04:27
<alkisg>
Yeah there's no need for a "BOOT=aoe" construct, when plain root=path is enough...
04:27
But again, what I'd like is a single root=mount-request1,mount-request2 parameter, nothing else
04:27
<vagrantc>
alkisg: i'm tempted to put configure_networking into aoe_ltsp
04:28
<alkisg>
why?
04:28
<vagrantc>
alkisg: nothing else brings up networking otherwise
04:28
<alkisg>
We should fix that ^
04:28
<vagrantc>
maybe so
04:29
<alkisg>
in aoe, networking should be initialized in the real system
04:29
Omitting the "eth0 manual" entry should be enough to do so
04:29
dhclient and everything, no need for static IPs
04:29
<vagrantc>
no, need to add an "iface eth0 inet dhcp" ... unless there's network-manager or wicd or something like that
04:30
<alkisg>
Hmmm another use case there,
04:31
for local boot, networking again won't be available from the initramfs, but it needs to be in init-ltsp.d in order to fetch lts.conf
04:32
So we can't wait until network manager etc bring up the interface... we at least need ipappend 3
04:32* vagrantc nods
04:32
<vagrantc>
setting "ipappend 3" then requires nbdroot= for NBD
04:32
<alkisg>
It doesn't require it for me here...
04:33
<vagrantc>
so it broke my beautiful menu that supports all three with a single setup
04:33
alkisg: how does it figure out the root-path if you're using ipappend 3 ?
04:33
<alkisg>
From dhcp
04:33
Ah
04:34
It's just the default, /opt/ltsp/<arch>
04:34
<vagrantc>
when you're using ipappend, it doesn't pass root-path on the commandline
04:34
not in my setup...
04:34
<alkisg>
Hmm checks if he puts that in udhcp...
04:34
*me
04:34
<vagrantc>
still don't use udhcp
04:34
myself
04:35
<alkisg>
vagrantc: I think I should split udhcp to 2 scripts
04:35
The first part would be "try to sanitize networking, apply things from ipappend etc"
04:35
I use that ^ part
04:36
The second part would be "use udhcp"
04:36
I don't care at all about that part
04:36* vagrantc nods
04:36
<vagrantc>
you use ipappend exclusively?
04:36
<alkisg>
Yes
04:37
<vagrantc>
with AoE we're back to using specific numbers for each image
04:37
i.e. no correlary to named exports
04:37
still pretty impressed with it, though... haven't tried to put it through it's paces
04:38
did have a mysterious lockup at some point, though
04:38* alkisg wants to test flow control issues with aoe...
04:38
<vagrantc>
and of course, now i'm curious about iSCSI, although it looks a lot more complicated to set up... still probably worth exploring, though
04:39* alkisg would prefer exploring dynamic local caching of the image :)
04:40
<alkisg>
I saw 2 mainline methods to do that...
04:41
http://www.phoronix.com/scan.php?page=news_item&px=MTM4ODA
04:44
It would allow fat client labs with 100 mbps networking to be able to support many many clients...
04:46
<vagrantc>
alkisg: i still don't get how you get nbd working with ipappend without specifying nbdroot or something similar
04:46
<alkisg>
vagrantc: check udhcp,
04:46
# Ensure a default rootpath if it doesn't exist
04:46
if [ -z "$rootpath" ]; then
04:46
if [ -f /conf/arch.conf ]; then
04:46
rootpath=$(. /conf/arch.conf; echo "${DPKG_ARCH:+/opt/ltsp/$DPKG_ARCH}")
04:46
fi
04:46
rootpath=${rootpath:-/opt/ltsp/i386}
04:46
fi
04:46
<vagrantc>
aha
04:46gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com)
04:46
<vagrantc>
that doesn't seem to happen for me
04:47
where's arch.conf comee from?
04:47
<alkisg>
initramfs-tools
04:48gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Read error: Operation timed out)
04:49
<alkisg>
vagrantc: it should fall back to /opt/ltsp/i386 if it doesn't exist, so I think that your problem may be that udhcp isn't called _at all_....
04:56Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
04:57
<vagrantc>
alkisg: except it was /opt/ltsp/i386-jessie
04:57
<alkisg>
Ah, ok
04:57
<vagrantc>
so i broked it.
04:58
<alkisg>
vagrantc: if we use linux.c32, we can use ipappend 3 *and* have all the dhcp info, without a dhcp request
04:58
<vagrantc>
i bet if i symlink i386.img to i386-jessie.img it'd work ... i seem to recall that working fine in the past
04:58
alkisg: i've thought about using some of the fancier syslinux features
04:59
can even write something up in lua
04:59
<alkisg>
Adding kernel parameters dynamically based on the client hardware... would be very nice
05:26alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection)
05:31cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 272 seconds)
05:47Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 245 seconds)
05:48gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com)
05:52Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
05:53gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 272 seconds)
05:59Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 264 seconds)
06:53work_alkisg has joined IRC (work_alkisg!~alkisg@plinet.ioa.sch.gr)
06:55
<work_alkisg>
vagrantc: are you going to make an ltsp release today?
06:55work_alkisg is now known as alkisg
07:02
<vagrantc>
alkisg: probably wait till tomorrow
07:03
alkisg: started looking at http://bugs.debian.org/src:ltsp
07:04
<alkisg>
Ah, nice
07:04
vagrantc: I don't see ltspfs 1.3 in your launchpad packaging mirror...
07:04
<vagrantc>
namely 701647, 613988, 704830 and 706812 (thinking of switching to http.debian.net)
07:04
alkisg: i'll make sure it's pushed
07:05
<alkisg>
vagrantc: re: #701647, are they using bind-mounts?!
07:07
<vagrantc>
alkisg: who knows.
07:07
<alkisg>
Hmm anyway I can look at bugs from home too, I should take advantage of the real clients I have here at work to do some aoe testing...
07:07
<vagrantc>
cool.
07:08
i've got a few errands to do tomorrow, but hope to get an upload end by the end of the day
07:08
<alkisg>
vagrantc: I want to start by installing a debian kernel alongside my trusty kernel, for clients without cmov (that I have here),
07:08
would you suggest some specific kernel? the -486 one?
07:08
...one that would target older clients mostly...
07:08
<vagrantc>
for maximum backwards-compatibility
07:08
<alkisg>
Cool, thanks
07:08
<vagrantc>
yeah, -486
07:09
hopefully your userspace cooperates
07:15
alkisg: did you pull ltspfs into ubuntu?
07:15prem_ has joined IRC (prem_!~prem@218.248.24.19)
07:15
<prem_>
hi all
07:15* vagrantc waves
07:16
<prem_>
is it possible to setup LTSP client to boot from wireless?
07:16
<alkisg>
vagrantc: no, either stgraber did that or it was done automatically
07:16
<prem_>
this is required for a library setup here on high priority
07:16
<alkisg>
Booting from wireless sucks, copy the ltsp image at some local media
07:17
<prem_>
i want to boot the ltsp clients over wireless and boot the client machine
07:17
<vagrantc>
alkisg: http://bugs.debian.org/613988, i was thinking of just adding ${SERVER_NAME:-"server"} to init-ltsp.d/09-hostname for the /etc/hosts editing
07:17* alkisg reads...
07:17
<vagrantc>
but it's not clear that's all they need
07:18
!wireless
07:18
<ltsp>
wireless: http://wiki.ltsp.org/twiki/bin/view/Ltsp/WirelessLtsp
07:18
<vagrantc>
oh wow, that's an old URL
07:18
prem_: the short of it is, it's not really a good idea, but for 1 client it might work
07:18
prem_: as long as nobody else is using the wireless much
07:18
<prem_>
vagrantc, am plannign to have atleast 10 clients simultaenoously
07:19
am planign to use that router only for pxe booting
07:19
not any other normal dhcp services will be there
07:19
<vagrantc>
prem_: then you either need to run a local OS, or string some wires.
07:19
prem_: or try wireless for yourself and see how poorly it performs
07:20
<alkisg>
vagrantc: SERVER_NAME would be a workaround, but the real solution would be to do a reverse dns query?
07:20
<vagrantc>
alkisg: it's not clear to me... guess i should follow-up.
07:21
<alkisg>
Because we could just do a reverse query, and if it works, use it, otherwise use "server"...
07:21* vagrantc wonders if DNS is setup at that point
07:22
<alkisg>
host $SERVER $SERVER should work even if DNS is unset...
07:23
host $DNS_SERVER $SERVER or something
07:25
<vagrantc>
right
07:28
<prem_>
vagrantc, is there any way to enhance the wireless setup in ltsp
07:29
<alkisg>
prem_: do/can the clients have local storage? internal disks, usb sticks etc?
07:29
<prem_>
alkisg, no we are planning to ban the
07:29
them*
07:29
if thats very much necessary., we ll try to include them vagrantc
07:30
alkisg,
07:30
<alkisg>
A thin client needs 50 MB to boot. 10 clients need 500 MB. On a wifi that would need 5+ minutes
07:30
So if you can have some local storage in the clients, you copy transfer /opt/ltsp/images/i386.img there and tell them to boot from that first, and THEN connect to the LTSP server for authentication + shared /home
07:32
<prem_>
alkisg, how this can be achieved ?
07:32
booting from local storage and authenticating from ltsp server ?
07:33
<alkisg>
!local-boot
07:33
<ltsp>
local-boot: If you want LTSP fat clients on a low-speed network, you can put i386.img on e.g. C:\Boot\LTSP\i386.img and use this command line in pxelinux.cfg: APPEND ro initrd=ltsp/i386/initrd.img init=/sbin/init-ltsp root=/dev/sda1 rootflags=ro loop=/Boot/LTSP/i386.img; IPAPPEND 3
07:34
<prem_>
alkisg, how about booting from a usb stick
07:34
<alkisg>
...something like that... I don't think anyone has written good documentation for it, you'd need to ask here again when you've done some steps
07:34
It's the same, usb or internal disk
07:35* alkisg wonders if squashfs supports writing a boot sector in the image in order to make it directly dd'able to USB sticks
07:35
<prem_>
alkisg, say i copied i386.img to usb and booted from that, then how come the image knows the ltsp server and get authenticated ?
07:36
<alkisg>
It does a dhcp request
07:36
Or you can also pass it to the command line
07:36
Those are details, first think if it suits your purposes and then start doing, and ask when you're blocked somewhere
07:37
<prem_>
hmm okey., alkisg i will try these options., but booting directly from wifi will not work out ?
07:38
<alkisg>
It's very slow
07:38
So when it works you will decide that you don't want it
07:39
<prem_>
actually i have connected my router to eth1 and configured the ltsp dhcp server to run on that eth1
07:39
alkisg, now for testing , how can i make the dhcp service to run on wlan0
07:39
this is bit comfusing to me
07:39
<alkisg>
It will take you about 1 week to understand how to boot a client over wireless
07:40
When you do it, you won't like it, and switch to something else
07:40
<prem_>
alkisg, am already lookign it for more than 10 days
07:40
<alkisg>
I didn't see you ask here :)
07:40
But anyways, as I said, the result won't be acceptable for 10 clients
07:41
To have an acceptable result, you need (1) ltsp fat clients, with (2) the squashfs ltsp image locally on some media, usb, internal disk, whatever
07:44
vagrantc: Trusty has initramfs-tools 0.103ubuntu3 and all the recent kernels I find need >= 0.110~... should I start looking in squeeze instead of jessie/wheezy?
07:46* alkisg tries http://packages.debian.org/wheezy/linux-image-3.2.0-4-486 ...
07:54
<vagrantc>
alkisg: ouch.
07:54
alkisg: that would be a bad idea
07:54
alkisg: initramfs-tools should be fairly safe to backport...
07:54
<alkisg>
I'm halfway into ltsp-update-image, let's see how that goes...
07:55
I think Ubuntu keeps a silly diff for initramfs-tools that's why they neglect syncing it...
07:58
<vagrantc>
alkisg: yeah, the version string isn't very meaningful... many of the features in newer debian initramfs-tools were first implemented in "old" ubuntu versions.
07:58
<alkisg>
I don't know if other debian maintainers would allow it, but the X-Ubuntu-Depends/recommends etc tricks are working fine for ltsp :)
07:58
<vagrantc>
tried to propose it for qemu, but they didn't like it... did something i think looks messier
07:59
<alkisg>
Ubuntu has thousands of 0ubuntuX packages, most of them for no real good reasons
07:59
Meh I forgot to run update-kernels again
08:00staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu)
08:01
<alkisg>
Yey, I got all the correct symlinks :)
08:04
Hehe, got a few glitches, but it made it up to the LDM login screen
08:07staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 245 seconds)
08:12
<alkisg>
Ah the main glitch was just the "fake dual head" intel issue, it just needs an xrandr calll...
08:14
<vagrantc>
wow.
08:15
that's good to know.
08:15
<alkisg>
Note that that laptop isn't only missing the -pae flag, it's also missing the cmov instruction :)
08:16
<vagrantc>
right
08:22
<alkisg>
OK first try, hdparm /dev/nbd0 on 3 NBD clients (then AOE test will follow)
08:23
7.02, 15.93 (gigabit), 6.64
08:26
Hmm no the aoe script failed, the link was down at the time of modprobe aoe
08:28
echo 3 > /proc/sys/vm/drop_caches; hdparm -t /dev/etherd/e0.0
08:28
11.27 MB/sec, 15.12 MB/sec, 11.26 MB/sec
08:29
Yeah that does sound better... let me also test with a really stupid atheros gigabit nic I have...
08:32work_alkisg has joined IRC (work_alkisg!~alkisg@194.63.234.224)
08:33alkisg has left IRC (alkisg!~alkisg@plinet.ioa.sch.gr, Read error: Operation timed out)
08:34work_alkisg is now known as alkisg
08:35
<alkisg>
Ouch no the atheros NIC suffers from the flow control issue even with aoe
08:35
2.49 MB/sec, 3.15 MB/sec, 3.29 MB/sec
08:42khildin has joined IRC (khildin!~khildin@ip-213-49-87-104.dsl.scarlet.be)
08:46
<alkisg>
Testing with intel...
08:47work_alkisg has joined IRC (work_alkisg!~alkisg@clnt-dide.ioa.sch.gr)
08:48alkisg has left IRC (alkisg!~alkisg@194.63.234.224, Read error: Connection reset by peer)
08:48work_alkisg1 has joined IRC (work_alkisg1!~alkisg@194.63.234.224)
08:50work_alkisg1 has left IRC (work_alkisg1!~alkisg@194.63.234.224, Client Quit)
08:50work_alkisg1 has joined IRC (work_alkisg1!~alkisg@194.63.234.224)
08:52work_alkisg has left IRC (work_alkisg!~alkisg@clnt-dide.ioa.sch.gr, Ping timeout: 260 seconds)
08:52work_alkisg1 is now known as alkisg
08:52
<alkisg>
Meh, realtek performs a _lot_ better than intel since 12.04... :-/
08:52
Intel: 2.59 MB/sec, 17.97 MB/sec, 1.03 MB/sec
09:00bennabiy has left IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com, Read error: Connection reset by peer)
09:01vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving)
09:04bennabiy has joined IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com)
09:08gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com)
09:10christophe_y2k has joined IRC (christophe_y2k!~christoph@man06-3-78-237-22-85.fbx.proxad.net)
09:13alkisg has left IRC (alkisg!~alkisg@194.63.234.224, Remote host closed the connection)
10:09vmlintu has joined IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi)
10:23PhoenixSTF has joined IRC (PhoenixSTF!~rudi@78.29.154.124)
10:29yopla has left IRC (yopla!~Adium@136.47.102.84.rev.sfr.net)
10:36vmlintu has left IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi, Ping timeout: 272 seconds)
10:37gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 252 seconds)
10:51vmlintu has joined IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi)
10:55Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
10:58alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
11:09Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 264 seconds)
11:10Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
11:13christophe_y2k has left IRC (christophe_y2k!~christoph@man06-3-78-237-22-85.fbx.proxad.net, Quit: Leaving.)
11:24vmlintu has left IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi, Ping timeout: 272 seconds)
11:36vmlintu has joined IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi)
11:53willianmazzardo has joined IRC (willianmazzardo!~textual@187.4.15.116)
12:01fjck has joined IRC (fjck!iamparadox@c-71-206-130-134.hsd1.va.comcast.net)
12:01gvy has left IRC (gvy!~mike@altlinux/developer/mike, Read error: Connection reset by peer)
12:02gvy has joined IRC (gvy!~mike@altlinux/developer/mike)
12:05highvolt1ge has joined IRC (highvolt1ge!~highvolta@abathur.jonathancarter.org)
12:06NeonLich1 has joined IRC (NeonLich1!~NeonLicht@darwin.ugr.es)
12:06vmlintu_ has joined IRC (vmlintu_!~vmlintu@cs181214103.pp.htv.fi)
12:06ogra_` has left IRC (ogra_`!~ogra_@p5098ed03.dip0.t-ipconnect.de, Ping timeout: 272 seconds)
12:06alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 272 seconds)
12:06vmlintu has left IRC (vmlintu!~vmlintu@cs181214103.pp.htv.fi, Ping timeout: 272 seconds)
12:06highvoltage has left IRC (highvoltage!~highvolta@ubuntu/member/highvoltage, Ping timeout: 272 seconds)
12:06|Paradox| has left IRC (|Paradox|!iamparadox@c-71-206-130-134.hsd1.va.comcast.net, Ping timeout: 272 seconds)
12:06NeonLicht has left IRC (NeonLicht!~NeonLicht@darwin.ugr.es, Ping timeout: 272 seconds)
12:06fjck is now known as |Paradox|
12:08alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
12:10ogra_ has joined IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de)
12:15prem_ has left IRC (prem_!~prem@218.248.24.19, Remote host closed the connection)
13:01Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Quit: Leaving.)
13:10brunolambert has joined IRC (brunolambert!brunolambe@nat/revolutionlinux/x-rmepiswqaekiphtc)
13:58bennabiy has left IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com, Remote host closed the connection)
13:59bennabiy has joined IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com)
14:27alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Remote host closed the connection)
14:29alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47)
14:38dsugar100 has joined IRC (dsugar100!~dsugar@columbia.tresys.com)
14:53nocturn has joined IRC (nocturn!~nocturn@unaffiliated/nocturn)
14:59alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection)
15:10vmlintu_ has left IRC (vmlintu_!~vmlintu@cs181214103.pp.htv.fi, Ping timeout: 245 seconds)
15:31staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu)
16:09alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
16:22brunolambert has left IRC (brunolambert!brunolambe@nat/revolutionlinux/x-rmepiswqaekiphtc, Quit: Leaving.)
16:46staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 272 seconds)
17:16bennabiy has left IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com, Ping timeout: 240 seconds)
17:17staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu)
17:19bennabiy has joined IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com)
17:30staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 248 seconds)
17:47gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com)
17:47staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu)
18:01vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
18:09staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 240 seconds)
18:11staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu)
18:15corviv has joined IRC (corviv!~corviv@91.210.103.190)
18:18alkisg_jessie has joined IRC (alkisg_jessie!59d2a008@gateway/web/freenode/ip.89.210.160.8)
18:19corviv has left IRC (corviv!~corviv@91.210.103.190, Remote host closed the connection)
18:22bennabiy has left IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com, Ping timeout: 240 seconds)
18:22bennabiy has joined IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com)
18:25ekenina has joined IRC (ekenina!~ekenina@91.210.102.154)
18:27ekenina has left IRC (ekenina!~ekenina@91.210.102.154, Remote host closed the connection)
18:29
<alkisg_jessie>
Heh, vmlinuz and initrd in debian have the exact half size of the ubuntu ones...
18:31* ogra_ whispers "plymouth fonts"
18:35
<vagrantc>
which really matters for network booting ppc64
18:36
<ogra_>
totally
18:36
<alkisg_jessie>
I think the size is due to many many modules being either compiled in vmlinuz or included in initrd... and I think we don't care much about all of them, they're not used until the network root is mounted...
18:37
<ogra_>
i dont think debian and ubuntu differ much in what modules they include
18:37
<alkisg_jessie>
I even get a 5 second lag on ubuntu's vmlinuz _before_ it loads the initrd... at least since 12.04
18:38
...which isn't there on debian or mainline kernels
18:38
Somebody forgot a "debug sleep 5" there :P
18:41
<ogra_>
you will find that most bloat on the ubuntu side is related to plymouth, raid or encryption
18:41
(and libdrm-(radeon|nouveau|foo)
18:41
<alkisg_jessie>
The main difference is: 8,2M firmware
18:41
<ogra_>
(oh, and lvm)
18:42
<alkisg_jessie>
I.e. the initrd size is due to the driver's firmware not included in debian by default as it's non-free
18:42
In debian that's only: 44K firmware
18:44
<ogra_>
oh, yeah
19:23alkisg_jessie has left IRC (alkisg_jessie!59d2a008@gateway/web/freenode/ip.89.210.160.8, Ping timeout: 272 seconds)
19:26vmlintu has joined IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi)
19:55
<alkisg>
http://www.syslinux.org/wiki/index.php/PXELINUX#Fetching_images_via_HTTP.2FFTP
19:55gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Remote host closed the connection)
19:56gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com)
19:57
<alkisg>
...with that we could easily tell ltspd to dynamically construct/send pxelinux.cfg/* files to the clients, by looking at lts.conf KERNEL_PARAMETERS directives :D
19:59* vagrantc nod
19:59
<vagrantc>
s
20:00gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 245 seconds)
20:02
<vmlintu>
alkisg: puavo-tftpd can act as a http-tftpd bridge: https://github.com/opinsys/puavo-tftp
20:18bennabiy has left IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com, Ping timeout: 240 seconds)
20:19bennabiy has joined IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com)
20:51
<alkisg>
vmlintu: yeah that's a good idea too, since tftp is a necessary service for PXE booting, it could also act as an ltspd server. Before starting to implement ltspd, we should review all possible directions - of course taking stability, performance and easy of maintainance into account too.
20:54gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com)
20:57
<vmlintu>
alkisg: we've been now using that approach for a year
20:58
<alkisg>
vmlintu: how much slower than http is it?
20:58
<vmlintu>
our pxelinux.cfg configurations are coming from a http server
20:59
alkisg: do you mean tftp compared to http?
20:59
<alkisg>
Yup
21:00
<vmlintu>
I really haven't done much benchmarks as we really cannot use http in pxe as quite a few nics seem to have broken firmware..
21:00
<alkisg>
So lpxelinux.0 doesn't work in some nics?
21:02gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Remote host closed the connection)
21:02gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com)
21:06
<vmlintu>
is the code different from gpxe?
21:07
I just remember that with gpxe/ipxe pretty much all realtek nics that we had failed to work: http://etherboot.org/wiki/troubleshooting
21:08
I do not remember hearing about lpxelinux.0 before
21:10
<alkisg>
I think it would use the PXE stack of the NIC instead of the iPXE stack
21:12
<vmlintu>
Some NICs have built-in http support?
21:14
I have to admit that I didn't bother digging too deep into this as tftp was fast enough and puavo-tftpd made it possible to dynamically create the tftp responses
21:17
We use it now so that when puavo-tftpd get a pxelinux.cfg request, it makes an http request to a http server that generates the response that puavo-tftpd then sends back over tftp: https://github.com/opinsys/puavo-users/blob/master/rest/resources/hosts.rb
21:19
Here's another tftpd server writter in Go that does the same thing: https://github.com/epeli/hooktftp
21:19
<alkisg>
No, the PXE stack allows a nic to do IP networking. pxelinux.0 does tftp over the stack, and lpxelinux.0 does http over that stack, afaik
21:19
Since you're already using http, maybe you don't even need the tftp part...
21:20
<vmlintu>
I need to try out the lpxelinux.0 to see if it works better than gpxe/ipxe
21:21
It would make life easier
21:24
alkisg: is lpxelinux.0 packaged in debian/ubuntu?
21:24
<alkisg>
vmlintu: no idea, only today I saw the ipxelinux.0 name for the first time
21:24
*lpxelinux
21:25
<vmlintu>
ok
22:05alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection)
22:12gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Remote host closed the connection)
22:22gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com)
22:33khildin has left IRC (khildin!~khildin@ip-213-49-87-104.dsl.scarlet.be, Ping timeout: 246 seconds)
22:46khildin has joined IRC (khildin!~khildin@ip-213-49-83-40.dsl.scarlet.be)
23:22gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Remote host closed the connection)