IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 7 February 2013 (all times are UTC)
| 00:03 | dorphalsig has joined IRC (dorphalsig!be93997b@gateway/web/freenode/ip.190.147.153.123) | |
| 00:05 | <dorphalsig> Hello, I'm having a bit of trouble with local printing
| |
| 00:05 | (it just wont work)
| |
| 00:05 | if anyody lives in here, maybe you could help me?
| |
| 00:10 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 00:18 | anunnaki has left IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net, Read error: Connection reset by peer) | |
| 00:22 | <vagrantc> dorphalsig: try to describe your problem, and wait a while, and people will respond if they can
| |
| 00:22 | explain your setup, what OS is installed, version of ltsp...
| |
| 00:35 | <dorphalsig> vagrantc: I'm on debian 6 (squeeze), running the stable ltsp build, 5.2.4
| |
| 00:36 | trying to connet a USB HP LJ 1300 printer
| |
| 00:36 | and I can't print
| |
| 00:36 | I mean I've added it to CUPS
| |
| 00:36 | I know the client recognizes it (because I have /deb/usb/lp0)
| |
| 00:36 | but I cant print
| |
| 01:04 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 01:18 | adrianorg__ has left IRC (adrianorg__!~adrianorg@177.156.225.34, Ping timeout: 260 seconds) | |
| 01:36 | vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving) | |
| 01:42 | Enslaver has joined IRC (Enslaver!~Enslaver@74.7.196.242) | |
| 01:52 | dorphalsig has left IRC (dorphalsig!be93997b@gateway/web/freenode/ip.190.147.153.123, Ping timeout: 245 seconds) | |
| 02:13 | Enslaver has left IRC (Enslaver!~Enslaver@74.7.196.242, Quit: This computer has gone to sleep) | |
| 02:37 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 248 seconds) | |
| 02:53 | simpoir has left IRC (simpoir!~simpoir@209.141.57.61) | |
| 03:10 | anunnaki has joined IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net) | |
| 03:11 | Parker955_Away is now known as Parker955 | |
| 03:34 | Enslaver has joined IRC (Enslaver!~Enslaver@c-98-196-42-169.hsd1.tx.comcast.net) | |
| 03:47 | Parker955 is now known as Parker955_Away | |
| 04:35 | vnc786 has joined IRC (vnc786!~chatzilla@49.248.129.178) | |
| 04:38 | ajith has joined IRC (ajith!1bfb78c2@gateway/web/freenode/ip.27.251.120.194) | |
| 04:38 | <ajith> JI
| |
| 04:38 | HI
| |
| 04:39 | Chan..How are you doing
| |
| 04:39 | I am having problem with client login in ubuntu clusters
| |
| 04:39 | can you tell ,me how can i resolve it
| |
| 04:56 | sha has joined IRC (sha!~sha@g231163205.adsl.alicedsl.de) | |
| 04:57 | anunnaki has left IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net, Remote host closed the connection) | |
| 04:59 | sha_ has left IRC (sha_!~sha@d148237.adsl.hansenet.de, Ping timeout: 255 seconds) | |
| 05:04 | <ajith> I am having problem with client login in ubuntu clusters
| |
| 05:04 | can you tell ,me how can i resolve it
| |
| 05:05 | HI
| |
| 05:21 | ajith has left IRC (ajith!1bfb78c2@gateway/web/freenode/ip.27.251.120.194, Quit: Page closed) | |
| 05:22 | anunnaki has joined IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net) | |
| 06:36 | cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 256 seconds) | |
| 06:43 | Enslaver has left IRC (Enslaver!~Enslaver@c-98-196-42-169.hsd1.tx.comcast.net, Ping timeout: 245 seconds) | |
| 06:48 | Enslaver has joined IRC (Enslaver!~Enslaver@74.7.196.242) | |
| 06:54 | khildin has joined IRC (khildin!~khildin@ip-80-236-222-28.dsl.scarlet.be) | |
| 07:29 | Enslaver has left IRC (Enslaver!~Enslaver@74.7.196.242, Ping timeout: 260 seconds) | |
| 07:49 | Enslaver has joined IRC (Enslaver!~Enslaver@c-98-196-42-169.hsd1.tx.comcast.net) | |
| 08:02 | dievel has joined IRC (dievel!~dievel@2-229-104-66.ip196.fastwebnet.it) | |
| 08:06 | mealstrom has joined IRC (mealstrom!~mealstrom@46.63.71.254) | |
| 08:39 | Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com) | |
| 08:40 | dobber has joined IRC (dobber!~dobber@89.190.199.210) | |
| 09:43 | vnc786 has left IRC (vnc786!~chatzilla@49.248.129.178, Quit: ChatZilla 0.9.89 [Firefox 11.0/20120410121533]) | |
| 09:53 | work_alkisg is now known as alkisg | |
| 10:00 | ajith has joined IRC (ajith!b7522f02@gateway/web/freenode/ip.183.82.47.2) | |
| 10:01 | teacher has joined IRC (teacher!c23fe7c6@gateway/web/freenode/ip.194.63.231.198) | |
| 10:01 | <teacher> Αλκη;
| |
| 10:01 | <ajith> hey I faciung problem with nbd and error is unregistered device at major 43
| |
| 10:01 | <alkisg> teacher: hello
| |
| 10:01 | <ajith> please help me me out in this
| |
| 10:01 | <alkisg> teacher: γράψε: /j #ts.sch.gr
| |
| 10:01 | Για να Îρθεις στο ελληνικÏŒ κανάλι υποστήριξης
| |
| 10:01 | (εδÏŽ είναι το αγγλικÏŒ)
| |
| 10:02 | <ajith> I am trying to connect from my thin client agter looking in the the log nbd from server it is givng me this mesage
| |
| 10:02 | teacher is now known as dorita | |
| 10:09 | <Enslaver> wow i think my unionfs overlay inside my initramfs i wrote into dracut actually might work
| |
| 10:09 | its just slow
| |
| 10:09 | <ajith> hello Enslaver
| |
| 10:09 | <Enslaver> howdy
| |
| 10:09 | <ajith> r u facing proble with dracut
| |
| 10:09 | error
| |
| 10:10 | it is nothing but restart ur nfs
| |
| 10:10 | and see it will work for sure
| |
| 10:10 | <Enslaver> I haven't gotten an error yet
| |
| 10:11 | I'm a little concerned about how the cow file system writes/rewrites its inodes
| |
| 10:11 | <ajith> okay
| |
| 10:11 | what is cow i also noticed that in edubuntu
| |
| 10:11 | <Enslaver> copy on write
| |
| 10:11 | its also a mammal that makes a moo noise.
| |
| 10:14 | btw your error with the NBD has to due to the NBD kernel module not being loaded, 43 is NBD's major device number
| |
| 10:15 | /dev/nbd0 is major 43 which is a block device created by the kernel or kernel module
| |
| 10:16 | try running modprobe nbd
| |
| 10:17 | woot, my sys root is now rw :)
| |
| 10:21 | dorita has left IRC (dorita!c23fe7c6@gateway/web/freenode/ip.194.63.231.198, Quit: Page closed) | |
| 10:30 | teacher has joined IRC (teacher!c23fe7c5@gateway/web/freenode/ip.194.63.231.197) | |
| 10:30 | teacher is now known as dorita | |
| 10:30 | <alkisg> dorita: /j #ts.sch.gr
| |
| 10:46 | adrianorg__ has joined IRC (adrianorg__!~adrianorg@187.113.246.15) | |
| 11:07 | <knipwim> Enslaver: i see you wrote a dracut unionfs module :)
| |
| 11:08 | be happy to test that this weekend
| |
| 11:08 | dorita has left IRC (dorita!c23fe7c5@gateway/web/freenode/ip.194.63.231.197, Quit: Page closed) | |
| 11:29 | <Enslaver> well, its in the process, it works when called from rdshell
| |
| 11:30 | for some reason it won't when ran all the way through
| |
| 11:31 | nm i think i just figured it out
| |
| 11:31 | <knipwim> Enslaver: i'm doing a commit for the NBD mount this weekend, or a commit request
| |
| 11:32 | so the named nbd exports are more robust
| |
| 11:32 | also, i remember i had some questions about the spec filew you pastebinned
| |
| 11:32 | but when i looked again, it was already gone
| |
| 11:35 | <Enslaver> i believe its still on my launched trunk
| |
| 11:35 | launchpad*
| |
| 11:36 | i'd like to see the debian package config so i can double check mine with it
| |
| 11:36 | knipwim: you any good with dracut?
| |
| 11:37 | This is my first time trying to code for it and I'm hitting a snag.
| |
| 11:37 | <knipwim> Enslaver: i've looked at it for the nbd mount
| |
| 11:38 | and tried an aufs plugin
| |
| 11:38 | some time ago
| |
| 11:39 | <dievel> Hi everybody! I'm using Edubuntu 12.04LTS with ltsp-server 5.3.7-0ubuntu2.3. Days ago Alksig pointend mi out that it's a bit an old version of LTSP. Does exist now a way to upgrade via PPA/backports?
| |
| 11:41 | <Enslaver> so i have my unionfs module directory, 3 scripts, install, check and unionfsroot which does the work
| |
| 11:42 | i install unionfsroot to /sbin in the initramfs but how do i tell it to run?
| |
| 11:42 | insthook?
| |
| 11:44 | inst "$moddir/unionfsroot" "/sbin/unionfsroot"
| |
| 11:44 | inst_hook cmdline 90 "$moddir/unionfsroot"
| |
| 11:44 | something like that?
| |
| 11:48 | really it just needs to run late in the pre-pivot stage
| |
| 11:59 | omg unionfs is a cpu hog
| |
| 12:00 | <alkisg> dievel: is that the case with ubuntu server and debian chroot?
| |
| 12:25 | <ogra_> Enslaver, unionfs as the kernel hack or unionfs as the fuse filesystem ?
| |
| 12:26 | <Enslaver> fuse
| |
| 12:26 | <ogra_> the latter is unusably slow (we did some tests in ubuntu before going to aufs)
| |
| 12:26 | <Enslaver> well the nice thing about unionfs-fuse is that i can run it in userspace and not rely on a custom kernel, although its probably what i'm gonna have to do anyway
| |
| 12:26 | * ogra_ would suggest some other overlay method | |
| 12:27 | <ogra_> well, the prob with userspace is that it is another extra layer that makes everything slow
| |
| 12:27 | <Enslaver> like what? dmsquash-live?
| |
| 12:27 | <ogra_> no way around that, if you want it at a usable spped you have to use kernel implementations (unionfs, aufs overlayfs)
| |
| 12:27 | <alkisg> ogra_: does that work on chunks of files, or on complete files? I.e. if I want to modify 1 byte of a .dvd file (4 Gb), will that need 4 Gb tmpfs?
| |
| 12:28 | <ogra_> i think so
| |
| 12:28 | <Enslaver> now if unionfs worked at the block level
| |
| 12:28 | <ogra_> i'm not the one who actively played with the code, but i helped testing etc
| |
| 12:28 | <alkisg> If it worked at file + chunk level, it would be much more efficient...
| |
| 12:28 | <ogra_> npone of the fuse variants out there are usable for a rootfs
| |
| 12:29 | <alkisg> Isn't clickfs fuse based?
| |
| 12:29 | <Enslaver> What about using ZFS like sun's VDI?
| |
| 12:29 | <ogra_> dunno, does that support union ? (and also is there any sane kernel patch)
| |
| 12:30 | <Enslaver> It supports union if its run at the block level i believe
| |
| 12:30 | which pretty much eliminates union all together
| |
| 12:30 | playing with this unionfs it might not be so bad
| |
| 12:31 | It's very customizable
| |
| 12:31 | <ogra_> try it, but in any case you will need kernel patches
| |
| 12:31 | ZFS isnt in any distro kernel i know
| |
| 12:31 | <Enslaver> yah i have my kernel compile running now :(
| |
| 12:31 | i thought it was in 2.6.28+ ?
| |
| 12:32 | I might be thinking of fuse
| |
| 12:33 | <alkisg> http://zfs-fuse.net/
| |
| 12:33 | <Enslaver> I've been reading up on zfsonlinux, looks promising, think i'll make my next kernel include aufs, zfs, nbd and some other fun stuff
| |
| 12:34 | <alkisg> Is there any advantage to zfs over e.g. btrfs?
| |
| 12:34 | * alkisg can't wait for btrfs to be stable, its snapshots will make ltsp-pnp instant | |
| 12:35 | <Enslaver> doesn't suse use btrfs?
| |
| 12:35 | <alkisg> I don't think any distro ships it as the default
| |
| 12:35 | <ogra_> that would be brave
| |
| 12:35 | its far from being done
| |
| 12:35 | <Enslaver> fc17?
| |
| 12:35 | <alkisg> (...and file-based compression will make squashfs unneeded)
| |
| 12:35 | I don't think it will be in fc19 either
| |
| 12:36 | <Enslaver> /usr/share/dracut/modules.d/90btrfs
| |
| 12:36 | looks like dracut is ready for it
| |
| 12:36 | <alkisg> Most distros support it, but the filesystem itself isn't ready
| |
| 12:36 | <Enslaver> oic
| |
| 12:38 | someone join #btrfs and tell them i said to hurry
| |
| 12:38 | <alkisg> Hehe
| |
| 12:38 | They'll probably reply "send money and we'll hurry" :)
| |
| 12:49 | <Enslaver> i'll remind them I'm not obama
| |
| 12:58 | finally, /root/rpmbuild/RPMS/x86_64/kernel-2.6.32279.el6.thinclient.x86_64-2.x86_64.rpm
| |
| 12:59 | <alkisg> !ltsp-bug
| |
| 12:59 | <ltsp> ltsp-bug: To file a bug report for upstream LTSP, go to https://bugs.launchpad.net/ltsp
| |
| 13:14 | jammcq has left IRC (jammcq!~jam@c-69-245-75-255.hsd1.mi.comcast.net, Quit: leaving) | |
| 13:14 | alkisg is now known as work_alkisg | |
| 13:20 | biomorph has joined IRC (biomorph!~biomorph@91.85.204.16) | |
| 13:20 | telex has left IRC (telex!~telex@freeshell.de, Remote host closed the connection) | |
| 13:22 | telex has joined IRC (telex!~telex@freeshell.de) | |
| 13:32 | Gremble has left IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com, Quit: I Leave) | |
| 14:03 | dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174) | |
| 14:18 | <dievel> alkisg: sorry, I was away from computer. Nope, I have a normal edubuntu installation with a i386 chroot environment, coming from the installation. I had some issue with epoptes, because in the image was not presente the epoptes client.
| |
| 14:19 | Helping me, you noticed that the LTSP version in edubuntu 12.04LTS, from repos, was quite old and missing some nice feature.
| |
| 14:19 | So I was guessing If i could upgrade the packages without messing up the whole server.
| |
| 14:32 | <warren> fedora 18 sort of uses btrfs by default
| |
| 14:33 | Enslaver: we can't ship kernel-based zfs in fedora or epel
| |
| 14:33 | Enslaver: incompatible license
| |
| 14:34 | Enslaver: you could potentially patch unionfs or aufs into your thinclient-only kernel, but that will be a maintenance burden entirely on you.
| |
| 14:34 | Enslaver: for years Valerie Aurora has been talking about an upstream union solution
| |
| 14:34 | I wonder what happened to that
| |
| 14:34 | jedeye has joined IRC (jedeye!~masalie@uwcfw.uwc.ac.za) | |
| 14:38 | <warren> Enslaver: sorry to hear that fuse unionfs is too slo
| |
| 14:38 | slow
| |
| 14:40 | Enslaver: https://lwn.net/Articles/447650/
| |
| 14:41 | Enslaver: latest news I can fid
| |
| 14:41 | find
| |
| 14:41 | designbybeck has joined IRC (designbybeck!~quassel@x176y051.angelo.edu) | |
| 14:42 | <warren> Enslaver: err, is it measurably slower?
| |
| 14:42 | how much?
| |
| 14:46 | Enslaver: https://lwn.net/Articles/482779/ Feb 2012 on union mount
| |
| 14:46 | mkuzu has joined IRC (mkuzu!~quassel@178.250.90.62) | |
| 14:47 | ajith has left IRC (ajith!b7522f02@gateway/web/freenode/ip.183.82.47.2, Quit: Page closed) | |
| 14:51 | Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com) | |
| 14:58 | <warren> Enslaver:
| |
| 14:58 | Enslaver: http://code.google.com/p/sfs/
| |
| 14:59 | Enslaver: whatever this is claims to be a FAST fuse based unionfs
| |
| 14:59 | Enslaver: might be worth comparing
| |
| 15:02 | <sbalneav> Morning all
| |
| 15:02 | jammcq has joined IRC (jammcq!~jam@c-69-245-75-255.hsd1.mi.comcast.net) | |
| 15:02 | <sbalneav> Morning jammcq
| |
| 15:05 | <jammcq> sbalneav: Scotty !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
| |
| 15:05 | meamy has joined IRC (meamy!~hannes@pd95cdee4.dip0.t-ipconnect.de) | |
| 15:11 | Enslaver has left IRC (Enslaver!~Enslaver@c-98-196-42-169.hsd1.tx.comcast.net, Quit: This computer has gone to sleep) | |
| 15:19 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 15:25 | biomorph has left IRC (biomorph!~biomorph@91.85.204.16, Remote host closed the connection) | |
| 15:33 | staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu) | |
| 15:37 | yalu has left IRC (yalu!~yalu@234.105-64-87.adsl-dyn.isp.belgacom.be, Read error: Connection reset by peer) | |
| 15:37 | yalu has joined IRC (yalu!~yalu@197.185-64-87.adsl-dyn.isp.belgacom.be) | |
| 15:44 | jedeye has left IRC (jedeye!~masalie@uwcfw.uwc.ac.za, Ping timeout: 248 seconds) | |
| 15:44 | Gremble has left IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com, Quit: I Leave) | |
| 15:44 | F-GT has left IRC (F-GT!~phantom@ppp59-167-136-109.static.internode.on.net, Ping timeout: 260 seconds) | |
| 15:45 | Enslaver has joined IRC (Enslaver!~Enslaver@adsl-70-241-84-25.dsl.hstntx.swbell.net) | |
| 15:46 | <Enslaver> I'm back, was warren here?
| |
| 15:46 | F-GT has joined IRC (F-GT!~phantom@ppp59-167-136-109.static.internode.on.net) | |
| 15:56 | mkuzu has left IRC (mkuzu!~quassel@178.250.90.62, Remote host closed the connection) | |
| 16:00 | <Hyperbyte> !seen warren
| |
| 16:00 | <ltsp> warren was last seen in #ltsp 1 hour, 1 minute, and 13 seconds ago: <warren> Enslaver: might be worth comparing
| |
| 16:01 | <Hyperbyte> Enslaver, http://irclogs.ltsp.org/ ;-)
| |
| 16:02 | komunista has joined IRC (komunista!~slavko@adsl-195-168-244-224.dynamic.nextra.sk) | |
| 16:02 | <Enslaver> !seen my brain
| |
| 16:02 | <ltsp> (seen [<channel>] <nick>) -- Returns the last time <nick> was seen and what <nick> was last seen saying. <channel> is only necessary if the message isn't sent on the channel itself. <nick> may contain * as a wildcard.
| |
| 16:02 | <Enslaver> !seen waldo
| |
| 16:02 | <ltsp> I have not seen waldo.
| |
| 16:02 | <Enslaver> Ha, me either
| |
| 16:03 | <jammcq> keep in mind, it's like 5am in hawaii where Warren is
| |
| 16:03 | <Enslaver> Linux people don't sleep
| |
| 16:04 | BTW jim, i could use your help with this dracut module
| |
| 16:05 | <jammcq> i've never seen dracut
| |
| 16:05 | <Enslaver> i've been looking at some code you did back a long time ago
| |
| 16:05 | <jammcq> so I dunno how much help i'd be
| |
| 16:05 | <Enslaver> # disklessrc file for initrd (Initial Ram Disk)
| |
| 16:05 | # Taken in part from James A. McQuillan <jam@McQuil.Com> linuxrc for LTSP
| |
| 16:05 | <jammcq> heh
| |
| 16:06 | <Enslaver> You might be some help :)
| |
| 16:06 | <jammcq> I might be of some help, but my stuff has been so changed over the years, I prolly won't recognize it
| |
| 16:07 | <Enslaver> Well i've been working on the dracut module to incorporate unionfs into the initramfs, problem is i've been having some major performance issues with it
| |
| 16:07 | You might know a way to better optimize the overlay
| |
| 16:08 | <jammcq> when I was doing that type of development, we weren't using any overlay filesystems. I'd just create a ramfs for /tmp and mount that.
| |
| 16:08 | but I really like the idea of making all of / writeable
| |
| 16:08 | that solves a number of problems that I ran into
| |
| 16:09 | I didn't even have bind mounts back then
| |
| 16:09 | <Enslaver> well thats interesting, looking at my machine again the performance is fine.
| |
| 16:10 | The way i'm doing the overlay is i make 1 directory a tmpfs filesystem (COW) and the other is the chroot RO nfs (chroot-ro)
| |
| 16:12 | then i'm doing a 'union' of those 2 directories into a 3rd one (/sysroot) and then dracut takes care of the root swap
| |
| 16:13 | so all changes write out to cow, chroot-ro stays ro and /sysroot is (rw), issue before was the server cpu was going bonkers, do you know of any way to help the per formance for that situation?
| |
| 16:13 | <jammcq> I saw ogra's comments about using fuse modules
| |
| 16:14 | are you doing this in user space or kernel space?
| |
| 16:14 | <Enslaver> user space
| |
| 16:15 | my next option is to go to aufs with a custom kernel
| |
| 16:15 | or sfs as warren mentioned
| |
| 16:16 | <jammcq> dunno what sfs is
| |
| 16:16 | <Enslaver> but i think thats fuse based also
| |
| 16:16 | sfs = SwitchFS
| |
| 16:16 | anunnaki has left IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net, Ping timeout: 256 seconds) | |
| 16:16 | <jammcq> the problem with userspace is there's too much context switching
| |
| 16:17 | so the cpu is busy doing that all the time
| |
| 16:17 | the vfs layer is kernel space, then you throw a big layer in there that's userspace, it's switching back and forth too much
| |
| 16:19 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 16:20 | <Enslaver> I really am hoping for a solution that is both portable across distro's and in kernel space with the flexibility of something like zfs or btrfs.
| |
| 16:21 | I wonder if the new 3.6 kernel offers anything worthwhile
| |
| 16:22 | viggy_prabhu has joined IRC (viggy_prabhu!~viggy@117.192.96.228) | |
| 16:26 | anunnaki has joined IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net) | |
| 16:28 | meamy has left IRC (meamy!~hannes@pd95cdee4.dip0.t-ipconnect.de, Remote host closed the connection) | |
| 16:32 | <alkisg> One portable solution is to use nbd-server -c, copy on write :) But it's not very flexible...
| |
| 16:37 | khildin has left IRC (khildin!~khildin@ip-80-236-222-28.dsl.scarlet.be, Quit: I'm gone, bye bye) | |
| 16:41 | dobber has left IRC (dobber!~dobber@89.190.199.210, Remote host closed the connection) | |
| 16:47 | <ogra_> or just default to nfs ;)
| |
| 16:47 | its not that there wouldnt be code for that already
| |
| 16:55 | <alkisg> NFS copy on write?
| |
| 16:55 | Or do you mean overlayfs? Sure, once that's portable across distros...
| |
| 16:56 | Maybe with the "read only root to /usr" move, it'll be easy to support netbooting without writeable /
| |
| 16:57 | Parker955_Away is now known as Parker955 | |
| 16:58 | <ogra_> yeah, i actually meant nfs + overlay
| |
| 17:00 | theoretically thopugh you should be able to just mount various dirs in tmpfses and have some copying and bind mounting mechanics to get around the need for any union at all
| |
| 17:00 | <alkisg> The problem with nfs is that it's much slower than e.g. squashfs + nbd (as you already know)
| |
| 17:00 | And even with compressed btrfs, compression doesn't help, as it's uncompressed before sent over the network
| |
| 17:00 | <ogra_> your /proc/mounts would become a mess indeed :)
| |
| 17:01 | nfs flies compared to any of the fuse implementations
| |
| 17:01 | <alkisg> bindfs is many many times faster than nfs
| |
| 17:01 | So I don't think it's a problem with fuse
| |
| 17:02 | But a problem with each implementation for itself
| |
| 17:02 | <ogra_> the initial LTSP5 implementation didnt use any unionfs stuff at all
| |
| 17:02 | (the prototype mdz wrote, on which i based all my work)
| |
| 17:02 | <alkisg> It used bind mounts?
| |
| 17:02 | <ogra_> it is possible but boots rather slow due to all the setup it has to do
| |
| 17:03 | it used a tmpfs, bind mounts and links, yeah
| |
| 17:03 | <alkisg> We still supported that until about a year ago, although untested
| |
| 17:03 | * jammcq remembers the slow boot | |
| 17:03 | Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com) | |
| 17:03 | <alkisg> But it had a significant overhead on which files/dirs would be writeable
| |
| 17:03 | (writing them all down in the code, I mean)
| |
| 17:03 | <ogra_> i still admire the beauty and simpleness of that initial design
| |
| 17:04 | <alkisg> On a typical boot, more than 100 files get modified
| |
| 17:04 | <ogra_> we only booted an Xserver and DM ...
| |
| 17:04 | it had no features at all
| |
| 17:04 | <alkisg> So that bind-mounts method caused numerous errors, even if booting continued after that
| |
| 17:04 | <ogra_> with that it was like 50 files or so it had to move around
| |
| 17:04 | <alkisg> And I'm betting some of the slow down was due to the errors, not the bind-mounts overhead
| |
| 17:05 | And now with fat clients that number arises a lot
| |
| 17:05 | <ogra_> it had no bugs back then :)
| |
| 17:05 | <jammcq> heh, it just wasn't finished :)
| |
| 17:05 | * alkisg tested bind-mounts about 2 years ago | |
| 17:05 | <ogra_> the environmental code changed a lot over the years
| |
| 17:05 | and yeah, it was never really finished
| |
| 17:05 | <alkisg> Many files were missing from the writeable list
| |
| 17:05 | <ogra_> yeah, nobody cared for it
| |
| 17:06 | still it was a beautiful design
| |
| 17:06 | * alkisg prefers the simplicity of /proc/mounts with a writeable cow / .... | |
| 17:06 | <ogra_> sure, but you depend on someone maintaining the kernel patch for it
| |
| 17:07 | <alkisg> I'm just depending on _any_ copy on write method being available
| |
| 17:07 | More than 5 are available for ubuntu now...
| |
| 17:07 | <ogra_> you are just lucky that ubuntu livefses use the same setup and thus you can rely on union mounting to always be supported
| |
| 17:07 | well, usable cow methods ..
| |
| 17:07 | i dont count the fuse ones here
| |
| 17:07 | and in the kernel there is exactly one
| |
| 17:08 | which causes the kernel team a lot of grief every release
| |
| 17:08 | <alkisg> Aufs is still available: /lib/modules/3.2.0-37-generic-pae/kernel/ubuntu/aufs/aufs.ko
| |
| 17:08 | overlayfs too, nbd-server -c too...
| |
| 17:08 | <ogra_> oh, i wasnt aware we kept overlay, i thought it was dropped
| |
| 17:09 | <alkisg> That's the new one that replaced aufs
| |
| 17:09 | <ogra_> then the other way round :)
| |
| 17:10 | i'm surprised the kernel team still maintains both, is that in raring ?
| |
| 17:10 | <alkisg> precise
| |
| 17:11 | <ogra_> ah, old crap :P
| |
| 17:11 | who uses that anyway
| |
| 17:11 | * alkisg thought that ubuntu will only ship lts releases from 14.04 on :D | |
| 17:11 | <ogra_> we'll see :)
| |
| 17:12 | the infrastructure isnt there yet ...
| |
| 17:12 | but it is being worked on
| |
| 17:17 | <Enslaver> aufs looks doable
| |
| 17:33 | sbalneav has left IRC (sbalneav!~sbalneav@mail.legalaid.mb.ca, Remote host closed the connection) | |
| 17:41 | mikkel has joined IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk) | |
| 17:46 | <Enslaver> ugh this is driving me nuts, dracut won't run my script. I can run it from rdshell fine, it gets put in the correct mount directory, permissions are good, the hook is there. but won't run, bla
| |
| 17:53 | yalu has left IRC (yalu!~yalu@197.185-64-87.adsl-dyn.isp.belgacom.be, Ping timeout: 246 seconds) | |
| 17:55 | Parker955 is now known as Parker955_Away | |
| 17:56 | yalu has joined IRC (yalu!~yalu@91.176.210.82) | |
| 18:03 | yalu has left IRC (yalu!~yalu@91.176.210.82, Ping timeout: 260 seconds) | |
| 18:04 | yalu has joined IRC (yalu!~yalu@204.111-64-87.adsl-dyn.isp.belgacom.be) | |
| 18:13 | anunnaki has left IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net, Ping timeout: 248 seconds) | |
| 18:15 | anunnaki has joined IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net) | |
| 18:23 | <jammcq> Enslaver: i've never played with dracut, but I'm looking at the wiki page and I'd check that the script ends with '.sh' and starts with a 2-digit number
| |
| 18:24 | viggy_prabhu has left IRC (viggy_prabhu!~viggy@117.192.96.228, Ping timeout: 246 seconds) | |
| 18:27 | sbalneav has joined IRC (sbalneav!~sbalneav@mail.legalaid.mb.ca) | |
| 18:29 | <Enslaver> the script ends in .sh but no number in front, and looking at the others they don't have numbers either
| |
| 18:30 | I'm wondering if there isn't some kind of argument that needs to be passed to call it
| |
| 18:35 | yalu has left IRC (yalu!~yalu@204.111-64-87.adsl-dyn.isp.belgacom.be, Ping timeout: 252 seconds) | |
| 18:36 | yalu has joined IRC (yalu!~yalu@109.134.134.159) | |
| 18:37 | viggy_prabhu has joined IRC (viggy_prabhu!~viggy@117.192.96.228) | |
| 18:50 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 18:55 | JuJuBee has joined IRC (JuJuBee!~mike_knic@24.148.112.102) | |
| 18:58 | <JuJuBee> Could I install virtualbox and a windows vm into an image to use with fat clients? or is that too much?
| |
| 18:59 | <alkisg> It runs fine, just needs 256 extra ram
| |
| 18:59 | And 10% extra cpu
| |
| 18:59 | So e.g. with 768 client RAM you can run an XP VM fine
| |
| 18:59 | <Enslaver> Check out winPE
| |
| 19:00 | <alkisg> You can also use the same VM from many clients
| |
| 19:01 | Either with nbd-server -c, or with nfs/sshfs + snapshot
| |
| 19:03 | <JuJuBee> alkisg: so the other option would be to have a vm for each user?
| |
| 19:03 | <alkisg> That works too, we've tested up to 11 VMs
| |
| 19:03 | <JuJuBee> Which is the simpler implementation?
| |
| 19:05 | Disc space is not an issue on the server
| |
| 19:05 | <alkisg> 1 VM is much simpler for maintainance
| |
| 19:07 | <JuJuBee> So the 1 vm would be located in the image?
| |
| 19:08 | <alkisg> No, it's user data, it belongs somewhere under /home
| |
| 19:08 | And that way you can also modify it without re-running ltsp-update-image
| |
| 19:08 | An alternative is to publish it with nbd-server
| |
| 19:09 | That way it's much faster than sshfs
| |
| 19:09 | And then tell vbox to use /dev/nbd2 on the client
| |
| 19:12 | <JuJuBee> OK. I don't know much about nbd-server so I will have to do some reading.
| |
| 19:14 | As for the nfs/sshfs + snapshots are you saying have a single vm file in /home/VMs (or whatever) and tell the users to create a snapshot and save snapshots in ~/vbox (or wherever)?
| |
| 19:15 | <alkisg> Yes, but you should use an automated script for that, not "tell the users"...
| |
| 19:15 | vbox also supports shared folders + a gina authentication method
| |
| 19:15 | So you can integrate the linux accounts + homes with the windows accounts + homes
| |
| 19:16 | Or you could use samba 4 as a pdc, whatever you like
| |
| 19:16 | <JuJuBee> Only have 1 app that requires windows and only once in a while, so don't even need user accounts per se on win.
| |
| 19:17 | <alkisg> And that app doesn't run under wine?
| |
| 19:17 | <JuJuBee> Haven't tried in a few years, but it didn't
| |
| 19:17 | <alkisg> It would probably be easier to make it run on wine, if it's only 1 app
| |
| 19:18 | <JuJuBee> Maybe I will give it another try
| |
| 19:23 | viggy_prabhu has left IRC (viggy_prabhu!~viggy@117.192.96.228, Quit: Ex-Chat) | |
| 19:38 | <JuJuBee> Just tried installing under wine and it crashes on launch... :(
| |
| 19:45 | <||cw> wine can be tweaked a lot
| |
| 19:45 | an often needs to be
| |
| 19:53 | Gremble has left IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com, Quit: I Leave) | |
| 20:02 | <warren> Enslaver: please try to avoid solutions that require you to maintain a totally custom kernel
| |
| 20:02 | Enslaver: that shoots down one of our arguments of getting the custom kernel into EPEL, as I used to be able to argue that the source is a "mere copy" of RHEL6 with a different config file.
| |
| 20:02 | Enslaver: EPEL6 is already a security maintenance burden
| |
| 20:03 | Enslaver: yes, sfs appears to be yet another fuse-based overlay, but it claims to be really fast, so it might be worth try.
| |
| 20:03 | Enslaver: I suspect unionfs-fuse is slow not primarily because of fuse, but it just isn't optimized.
| |
| 20:05 | Enslaver: and while going to the latest upstream kernel may make it easier to maintain, it also is likely to break compatibility with EL-6 userspace that remains static.
| |
| 20:05 | <Enslaver> but to be able to support nbd we need a custom config regardless
| |
| 20:06 | <warren> Enslaver: custom config isn't a maintenance burden
| |
| 20:06 | shipping non-upstream kernel drivers is
| |
| 20:06 | <Enslaver> then our options are wait for a kernel based overlay fs or use unionfs
| |
| 20:07 | which union isn't turning out that badly, its actually better with this kernel
| |
| 20:07 | <warren> tried sfs?
| |
| 20:08 | i'm curious how much faster it is
| |
| 20:08 | <Enslaver> not yet, its on my list as soon as i figure out this koji thing
| |
| 20:08 | my kernel compiles keep getting canceled
| |
| 20:09 | <warren> give me a koji URL to the build ID
| |
| 20:11 | <Enslaver> 4936712
| |
| 20:12 | might be something I'm doing wrong, I'm still reading docs on how to use
| |
| 20:12 | koji build --scratch dist-6E-epel kernel-2.6.32-279.el6.thinclient.src.rpm is what i'm running
| |
| 20:13 | <warren> hmm, I dunno
| |
| 20:13 | I can't find anything by 4936712. can you give me the URL it spits out?
| |
| 20:14 | <Enslaver> I didn't see a url from the output
| |
| 20:15 | here's what i pull up under my name in the system:
| |
| 20:15 | https://koji.fedoraproject.org/koji/taskinfo?taskID=4936683
| |
| 20:15 | ah crap i didn't get the patch removed
| |
| 20:16 | <warren> http://kojipkgs.fedoraproject.org//work/tasks/6684/4936684/build.log
| |
| 20:17 | <Enslaver> yah lemme get the right src up and i'll paste ya the url
| |
| 20:18 | <warren> Enslaver: so i was correct, it wasn't the readonly filesystem that caused the x86_64 X failure
| |
| 20:18 | it was a kernel or X issue
| |
| 20:18 | ?
| |
| 20:19 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 20:19 | <Enslaver> The X issue isn't fixed
| |
| 20:20 | JuJuBee has left IRC (JuJuBee!~mike_knic@24.148.112.102) | |
| 20:20 | <Enslaver> yeah i didn't figure it was the ro filesystem, i had the nfs mounted rw for testing and it was occurring then
| |
| 20:22 | ldm segfaults with no descriptive error, i've traced it down to what i thought was a prefdm issue
| |
| 20:23 | i've gone so far as to completely wipe my x86_64 chroot and start over thinking i missed something
| |
| 20:23 | <warren> works on i686 but not x86_64
| |
| 20:23 | <Enslaver> correct
| |
| 20:23 | <warren> suggests a path issue, since other distros lack our multilib layout
| |
| 20:24 | <Enslaver> I have set -x on almost all the scripts currently, logging their output, nothing really sticks out.
| |
| 20:25 | I think it might be related to upstart
| |
| 20:27 | ltspuser_71 has joined IRC (ltspuser_71!40866104@gateway/web/freenode/ip.64.134.97.4) | |
| 20:27 | ltspuser_71 is now known as bantarr | |
| 20:28 | <bantarr> Hello, I was checking with this IRC channel to ask if anyone has successfully integrated their LTSP server with a samba4 domain controller to authenticate user logins on a thin client
| |
| 20:29 | I got my server to join, but cannot login with a domain user. My samba4 is set to use internal DNS
| |
| 20:32 | komunista has left IRC (komunista!~slavko@adsl-195-168-244-224.dynamic.nextra.sk, Quit: Leaving.) | |
| 20:38 | <bantarr> Also, I can use kinit administrator@domain and that works
| |
| 20:42 | but when I try to ssh to it (even ssh to localhost) using a domain user like: ssh 'domain/username'@localhost it keeps rejecting password
| |
| 20:53 | <warren> Enslaver: oops
| |
| 20:53 | Enslaver: high performance is an "upcoming" feature of sfs
| |
| 20:53 | Enslaver: meaning it might not be possible =)
| |
| 20:53 | <Enslaver> :(
| |
| 20:53 | You would think they would design the fs with that in mind
| |
| 20:54 | i'm curious to test unionfs on nbd, the hangup might be due to nfs
| |
| 21:21 | bantarr has left IRC (bantarr!40866104@gateway/web/freenode/ip.64.134.97.4, Ping timeout: 245 seconds) | |
| 21:27 | anunnaki has left IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net, Ping timeout: 246 seconds) | |
| 21:29 | komunista has joined IRC (komunista!~slavko@adsl-195-168-244-224.dynamic.nextra.sk) | |
| 21:30 | anunnaki has joined IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net) | |
| 21:34 | <warren> Enslaver: software plans are aspirational
| |
| 21:34 | Enslaver: I have planned features on k12linux.org that are 4 years late now
| |
| 21:39 | Enslaver: how slow exactly did unionfs-fuse make it?
| |
| 21:39 | <Enslaver> the only major slowdown was in the boot process
| |
| 21:39 | mealstrom has left IRC (mealstrom!~mealstrom@46.63.71.254, Ping timeout: 256 seconds) | |
| 21:43 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 21:46 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 21:59 | Parker955_Away is now known as Parker955 | |
| 22:22 | spectra has left IRC (spectra!~spectra@debian/developer/spectra, Ping timeout: 264 seconds) | |
| 22:27 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 22:28 | mikkel has left IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk, Quit: Leaving) | |
| 22:28 | spectra has joined IRC (spectra!~spectra@debian/developer/spectra) | |
| 22:30 | komunista has left IRC (komunista!~slavko@adsl-195-168-244-224.dynamic.nextra.sk, Quit: Leaving.) | |
| 22:33 | highvoltage has left IRC (highvoltage!~highvolta@ubuntu/member/highvoltage, Read error: Operation timed out) | |
| 22:35 | anunnaki has left IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net, Remote host closed the connection) | |
| 22:39 | anunnaki has joined IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net) | |
| 22:40 | dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Read error: Operation timed out) | |
| 22:46 | anunnaki has left IRC (anunnaki!~anunnaki@c-174-54-115-236.hsd1.pa.comcast.net, Remote host closed the connection) | |
| 22:53 | JuJuBee has joined IRC (JuJuBee!~mike_knic@24.105.231.111) | |
| 23:12 | JuJuBee has left IRC (JuJuBee!~mike_knic@24.105.231.111) | |
| 23:44 | Parker955 is now known as Parker955_Away | |
| 23:50 | lmds_ has left IRC (lmds_!~lmds@tui.pi-et-ro.net, Ping timeout: 255 seconds) | |
| 23:52 | xet7 has joined IRC (xet7!~xet7@a91-156-247-230.elisa-laajakaista.fi) | |