IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 21 November 2011 (all times are UTC)
| 00:01 | Parker955_Away is now known as Parker955 | |
| 00:04 | monteslu has joined IRC (monteslu!~monteslu@ip68-109-174-213.ph.ph.cox.net) | |
| 00:19 | sndwrx has left IRC (sndwrx!~sndwrx@host-64-188-228-94.windjammercable.net, Ping timeout: 260 seconds) | |
| 00:54 | fosser_josh has left IRC (fosser_josh!~prathames@02799a2d.bb.sky.com, Quit: Leaving.) | |
| 01:16 | sbalneav has joined IRC (sbalneav!~sbalneav@S0106586d8f030f4c.wp.shawcable.net) | |
| 01:30 | Yet_another_Bill has joined IRC (Yet_another_Bill!billy@nat/redhat/x-yvkqelamxntugkkx) | |
| 01:49 | dberkholz has joined IRC (dberkholz!UNKNOWN@gentoo/developer/dberkholz) | |
| 03:44 | leio_ has joined IRC (leio_!~leio@gentoo/developer/leio) | |
| 03:45 | leio has left IRC (leio!~leio@gentoo/developer/leio, Ping timeout: 244 seconds) | |
| 04:07 | Parker955 is now known as Parker955_Away | |
| 04:54 | loather-work has joined IRC (loather-work!~khudson@wsip-98-175-250-115.sd.sd.cox.net) | |
| 04:56 | loather-work has left IRC (loather-work!~khudson@wsip-98-175-250-115.sd.sd.cox.net, Client Quit) | |
| 05:09 | staffencasa has left IRC (staffencasa!~staffenca@128-193-149-224.oregonstate.edu, Ping timeout: 260 seconds) | |
| 05:30 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 05:40 | sbalneav has left IRC (sbalneav!~sbalneav@S0106586d8f030f4c.wp.shawcable.net, Quit: leaving) | |
| 07:09 | lifeboyy has joined IRC (lifeboyy!~roland@196-215-40-134.dynamic.isadsl.co.za) | |
| 07:11 | lifeboy has left IRC (lifeboy!~roland@41.183.40.96, Ping timeout: 245 seconds) | |
| 07:57 | map7 has joined IRC (map7!~map7@ppp118-209-114-57.lns20.mel4.internode.on.net) | |
| 07:57 | khildin has joined IRC (khildin!~khildin@ip-80-236-227-210.dsl.scarlet.be) | |
| 08:13 | vagrantc has left IRC (vagrantc!~vagrant@c-76-105-252-50.hsd1.or.comcast.net, Ping timeout: 252 seconds) | |
| 08:13 | leio_ has left IRC (leio_!~leio@gentoo/developer/leio, Ping timeout: 260 seconds) | |
| 08:14 | leio has joined IRC (leio!~leio@gentoo/developer/leio) | |
| 08:16 | loather has left IRC (loather!~khudson@wsip-98-175-250-115.sd.sd.cox.net, Quit: This computer has gone to sleep) | |
| 08:21 | lifeboyy is now known as lifeboy | |
| 08:24 | <muppis> Google Earth doesn't pop up as local app in menu of 10.04.
| |
| 08:25 | Tried even with full path to app as in .desktop file.
| |
| 09:10 | sndwrx has joined IRC (sndwrx!~sndwrx@host-64-188-228-94.windjammercable.net) | |
| 09:23 | dobber has joined IRC (dobber!~dobber@213.169.45.222) | |
| 09:41 | <alkisg> stgraber: epoptes is now in Debian unstable, you can pull it whenever you want: http://packages.debian.org/sid/epoptes
| |
| 10:26 | vmlintu has left IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi, Ping timeout: 244 seconds) | |
| 10:28 | Steve_the_Pirate has joined IRC (Steve_the_Pirate!~Gary@host109-155-255-130.range109-155.btcentralplus.com) | |
| 10:31 | cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Remote host closed the connection) | |
| 10:41 | cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg) | |
| 10:48 | <lifeboy> Is there a reason why this
| |
| 10:48 | root@Ashton:/# ltsp-update-image --arch=i386
| |
| 10:48 | Parallel mksquashfs: Using 12 processors
| |
| 10:48 | Creating 4.0 filesystem on /opt/ltsp/images/i386.img.tmp, block size 131072.
| |
| 10:48 | [======================================================================================\ ] 170981/246896 69%
| |
| 10:48 | comes to a standstill on a powerful server when one switches the focus away from the window in an X-server environment?
| |
| 10:59 | artista_frustrad has left IRC (artista_frustrad!~artista_f@187.7.191.18, Ping timeout: 252 seconds) | |
| 11:05 | <alkisg> ltsp-update-image runs with `nice` so that your users are not affected by the compression load
| |
| 11:09 | <lifeboy> alkisg: I checked again, yes, it actually does continue, it just slows down to a trickle
| |
| 11:10 | <khildin> great piece of software ... ltsp huh?... :)
| |
| 11:11 | <lifeboy> btw: alkisg and all the others that helped with my kernel compile problem: I asked in #ubuntu-kernel for help and smb and some others there assisted in getting the problem sorted and a kernel compiled with the modules I need.
| |
| 11:12 | It seems that the "Local version - append to kernel release" string I added could have caused the problem.
| |
| 11:12 | but thanks for all of you who helped and listened! :-)
| |
| 11:12 | artista_frustrad has joined IRC (artista_frustrad!~artista_f@187.55.101.27) | |
| 11:19 | artista_frustrad has left IRC (artista_frustrad!~artista_f@187.55.101.27, Ping timeout: 240 seconds) | |
| 11:29 | <alkisg> Nice :)
| |
| 11:32 | artista_frustrad has joined IRC (artista_frustrad!~artista_f@187.55.101.27) | |
| 11:37 | <lifeboy> The great thing is that I now have a client that uses less than 2 watts of power and with an LED 18.5" monitor @ less than 20W, I run the whole lot over PoE from the server's central UPS, all on the same cable (client and monitor both work on 14volts)
| |
| 11:38 | and it's all driven by LTSP!
| |
| 11:42 | <muppis> Sounds perfect.
| |
| 11:43 | I did other way around. Booted HP DL380 as client.
| |
| 11:45 | vmlintu has joined IRC (vmlintu!~vmlintu@dsl-hkibrasgw2-ff78c300-68.dhcp.inet.fi) | |
| 11:49 | monteslu has left IRC (monteslu!~monteslu@ip68-109-174-213.ph.ph.cox.net, Read error: Operation timed out) | |
| 11:50 | artista_frustrad has left IRC (artista_frustrad!~artista_f@187.55.101.27, Ping timeout: 258 seconds) | |
| 12:02 | artista_frustrad has joined IRC (artista_frustrad!~artista_f@187.55.101.27) | |
| 12:02 | monteslu has joined IRC (monteslu!~monteslu@ip68-109-174-213.ph.ph.cox.net) | |
| 12:15 | alkisg1 has joined IRC (alkisg1!~alkisg@ppp089210050147.dsl.hol.gr) | |
| 12:15 | alkisg1 has left IRC (alkisg1!~alkisg@ppp089210050147.dsl.hol.gr, Changing host) | |
| 12:15 | alkisg1 has joined IRC (alkisg1!~alkisg@ubuntu/member/alkisg) | |
| 12:16 | toscalix has joined IRC (toscalix!~toscalix@p54BD2F3D.dip.t-dialin.net) | |
| 12:16 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 248 seconds) | |
| 12:17 | alkisg1 is now known as alkisg | |
| 12:33 | artista_frustrad has left IRC (artista_frustrad!~artista_f@187.55.101.27, Ping timeout: 240 seconds) | |
| 12:33 | toscalix has left IRC (toscalix!~toscalix@p54BD2F3D.dip.t-dialin.net, Remote host closed the connection) | |
| 12:46 | artista_frustrad has joined IRC (artista_frustrad!~artista_f@187.55.101.27) | |
| 12:47 | Steve_the_Pirate has left IRC (Steve_the_Pirate!~Gary@host109-155-255-130.range109-155.btcentralplus.com, Ping timeout: 240 seconds) | |
| 12:49 | Steve_the_Pirate has joined IRC (Steve_the_Pirate!~Gary@host109-155-255-130.range109-155.btcentralplus.com) | |
| 12:56 | jammcq has left IRC (jammcq!~jam@70-91-230-209-BusName-Michigan.hfc.comcastbusiness.net, Quit: leaving) | |
| 12:59 | gentgeen__ has left IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net, Remote host closed the connection) | |
| 13:04 | <muppis> I'd like to have client hostnames to be resolved by server.
| |
| 13:05 | artista_frustrad has left IRC (artista_frustrad!~artista_f@187.55.101.27, Ping timeout: 252 seconds) | |
| 13:09 | brunolambert has joined IRC (brunolambert!~bruno@nat/revolutionlinux/x-gdfsinsvemwhggol) | |
| 13:12 | <Hyperbyte> muppis, ddns. :)
| |
| 13:13 | <muppis> Hyperbyte, sorry, I forgot the question. Any suggestion how to do that?
| |
| 13:13 | <Hyperbyte> ddns.
| |
| 13:13 | Dynamic dns updates.
| |
| 13:14 | Your local DHCP server can contact your local DNS server whenever it hands out a lease, and ask the server to create the approriate DNS record.
| |
| 13:14 | <muppis> No software to suggest?
| |
| 13:14 | <Hyperbyte> http://blogger.ziesemer.com/2008/10/linux-lan-dchp-dynamic-dns.html
| |
| 13:14 | Just your regular isc dhcpd server (default Ubuntu dhcpd) and bind nameserver (default Ubuntu dns server)
| |
| 13:15 | I wouldn't go through the trouble though, it's not that easy to set up (speaking from experience).
| |
| 13:15 | <muppis> As being lazy, I replaced all those with dnsmasq. :)
| |
| 13:15 | * alkisg too | |
| 13:16 | <alkisg> But dnsmasq supports scripts on dhcp leases, maybe that could be used instead
| |
| 13:16 | <Hyperbyte> You'd need to set up bind to allow updates from localhost, with the approriate rndc key
| |
| 13:17 | And then you'd need to configure your dhcpd server (whatever one you use) to authorize to bind with the correct rndc key and send the update
| |
| 13:17 | <muppis> Hmm.. Hold a second.
| |
| 13:18 | * Hyperbyte holds | |
| 13:18 | <Hyperbyte> No actually
| |
| 13:18 | * Hyperbyte lunches | |
| 13:18 | <Hyperbyte> Back in a bit.
| |
| 13:19 | artista_frustrad has joined IRC (artista_frustrad!~artista_f@187.55.101.27) | |
| 13:19 | <muppis> I just remembered that dnsmasq makes a record for a client if it receives the name in request.
| |
| 13:20 | <alkisg> muppis: in the dhcp request? Hmmm we could use the lts.conf HOSTNAME directive there
| |
| 13:20 | Let me try if that works
| |
| 13:21 | <muppis> I'll try that too.
| |
| 13:21 | <alkisg> Of course dnsmasq is also a dhcp server, so you can put the hostnames in dnsmasq
| |
| 13:21 | But it'd be cool to have even the ltsp123 automatic names resolvable
| |
| 13:21 | <muppis> That's my goal.
| |
| 13:27 | <alkisg> Haha it just works, nice :D
| |
| 13:27 | The problem is that we don't have lts.conf available before the dhcp request
| |
| 13:27 | So we don't know the client hostname
| |
| 13:27 | So, we'd have to do 2 dhcp requests
| |
| 13:28 | One to be able to fetch lts.conf, and another to notify dnsmasq about our hostname, which kinda sucks
| |
| 13:28 | <muppis> Yeah.
| |
| 13:28 | <alkisg> What we *could* do, is create a hostname based on the mac address
| |
| 13:29 | That way the clients would even have a static hostname, even if one doesn't use static IPs
| |
| 13:29 | muppis: could you file a bug report asking that? I.e. to change the default ltsp hostnames to mac-based hostnames?
| |
| 13:30 | E.g. ltsp0090f5805d24
| |
| 13:30 | <muppis> Isn't that in option already?
| |
| 13:30 | <alkisg> It's an option in lts.conf
| |
| 13:30 | We don't have access to lts.conf when we ask for a dhcp lease
| |
| 13:31 | So it would have to be the default option for this to work
| |
| 13:31 | <muppis> I can file a bug for it.
| |
| 13:32 | <alkisg> I wonder if that would be considered a security issue, if so we could use a hash function
| |
| 13:32 | E.g. md5sum of the mac address, or crc32
| |
| 13:32 | That would make the hostnames smaller too
| |
| 13:32 | artista_frustrad has left IRC (artista_frustrad!~artista_f@187.55.101.27, Ping timeout: 245 seconds) | |
| 13:33 | <Hyperbyte> alkisg, why would that be a security issue?
| |
| 13:33 | <muppis> Which makes you think that as security issue?
| |
| 13:33 | <Hyperbyte> :)
| |
| 13:33 | <alkisg> Exposing the mac address in the hostname
| |
| 13:33 | <muppis> Isn't that done anyway?
| |
| 13:33 | <alkisg> I don't know if that would be an issue, I'm just wondering
| |
| 13:33 | <Hyperbyte> Mac address is exposed as soon as a machine sends one packet onto the network
| |
| 13:34 | <alkisg> Yes, but it's not exposed to the internet
| |
| 13:34 | matrix3000 has left IRC (matrix3000!~Eric@rrcs-70-61-255-227.central.biz.rr.com, Ping timeout: 252 seconds) | |
| 13:34 | <alkisg> But e.g. the browser may send your hostname? Or something?
| |
| 13:34 | <Hyperbyte> Mac addresses can't be used to communicate over the internet, or via browsers.
| |
| 13:34 | *via routers
| |
| 13:35 | <alkisg> It tells someone information about your hardware
| |
| 13:35 | <Hyperbyte> Mac addresses are only usuable on one network segment (i.e, your switch and all machines connected to it)
| |
| 13:35 | <muppis> In local net it can finded anyway, if wanted.
| |
| 13:35 | <alkisg> (03:33:58 μμ) Hyperbyte: Mac address is exposed as soon as a machine sends one packet onto the network => all the machines know all machines? My arp cache table is usually not complete..
| |
| 13:36 | <Hyperbyte> alkisg, that's because usually machines don't sniff out all network traffic. ;-)
| |
| 13:36 | <alkisg> But anyway, *I* am not concerned, I hope others won't be too :D
| |
| 13:36 | <Hyperbyte> And only pick up traffic meant for them.
| |
| 13:36 | <muppis> Local arp cache knows only connected addresses' and in ltsp server..
| |
| 13:36 | <alkisg> I thought switches were intelligent enough to only send packets to interested clients
| |
| 13:37 | <Hyperbyte> alkisg, broadcasts (such as dhcp requests) are excluded from that logic
| |
| 13:37 | <alkisg> Switching to mac-address-based hostnames makes more sense than IPs anyway, because they're static
| |
| 13:37 | Hyperbyte, indeed, but the hostname is there even after the dhcp request
| |
| 13:38 | So e.g. you can find a thin client's mac address even days after its dhcp request
| |
| 13:38 | Anyway, I guess noone would worry about that
| |
| 13:38 | <Hyperbyte> If I can ping it, I can find it anyways.
| |
| 13:38 | And I can't ping it, I can't connect with it.
| |
| 13:39 | Only thing would be that remote websites and such, could possibly figure out what kind of network card the thin client has
| |
| 13:39 | <muppis> alkisg, so, against which packet I should file the bug?
| |
| 13:39 | <alkisg> !ltsp-bug
| |
| 13:39 | <ltsp> alkisg: ltsp-bug: To file a bug report for upstream LTSP, go to https://bugs.launchpad.net/ltsp
| |
| 13:40 | <alkisg> So... ltsp0090f5805d24 is too long, isn't it? How about getting a "modulo 10000" of that? Wouldn't that be rare enough for most installations?
| |
| 13:40 | Or maybe omit the vendor bits from the mac...
| |
| 13:41 | <muppis> In very rare cases that can cause duplicates.
| |
| 13:41 | <alkisg> What would dnsmasq do in that case?
| |
| 13:42 | Or should we just keep the full mac there?
| |
| 13:42 | <muppis> I'm not sure about the linux version, but in dd-wrt refuses work at all.
| |
| 13:42 | ogra_ has left IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de, Excess Flood) | |
| 13:42 | <muppis> Full mac or some sorter hash.
| |
| 13:43 | <alkisg> Hmm I had labs of windows machines with the same hostnames, I didn't have dhcp problems.
| |
| 13:43 | And also previously all ltsp clients had hostname=="ltsp"
| |
| 13:43 | ogra_ has joined IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de) | |
| 13:44 | <muppis> Of course. As I think what it would, it just gives the IP and put the name for it.
| |
| 13:44 | <alkisg> Wikipedia says "3 bytes UOI, and 3 bytes NIC specific"
| |
| 13:45 | Maybe we could keep the nic-specific bits. Or we could rewrite the whole mac in base-64
| |
| 13:46 | <muppis> Doesn't Base64 be longer than mac?
| |
| 13:47 | <alkisg> No, hex digits as letters express 16 bits
| |
| 13:47 | **16 different values, 4 bits
| |
| 13:47 | Base64 == 64 different values, 6 bits
| |
| 13:47 | So it would be like ltspa1b2c3d4
| |
| 13:48 | while keeping the 3 last bytes would be like ltsp1a2b3c, even smaller, but with possible collissions?
| |
| 13:49 | <muppis> Sounds acceptable.
| |
| 13:49 | <Hyperbyte> Possibility of collision is pretty small I'd say.
| |
| 13:50 | <muppis> But now I need to hit the road. Gotta fecth some food for kids before picking them from daycare.
| |
| 13:50 | <alkisg> I think so too, maybe big organizations like those would even use the same vendor, making collissions even more rare
| |
| 13:50 | bb muppis
| |
| 13:50 | <Hyperbyte> I'm not good at chance calculation, but I believe possibility of collision is one in 33 million.
| |
| 13:50 | <muppis> bye for today.
| |
| 13:51 | <Hyperbyte> Bye muppis!
| |
| 13:51 | <alkisg> Hyperbyte: 3 bytes = 256*256*256 = 16 million different values
| |
| 13:51 | <Hyperbyte> alkisg, yes, but then the matching one
| |
| 13:51 | <alkisg> The first doesn't collide
| |
| 13:51 | <Hyperbyte> I'd say it's 2x256x256x256
| |
| 13:51 | <alkisg> The second one has 1/16 milllions to collide
| |
| 13:51 | <Hyperbyte> mhmm
| |
| 13:51 | <alkisg> The third one, 2/16m to collide, etc
| |
| 13:52 | <Hyperbyte> See, told you I wasn't good at this.
| |
| 13:52 | They'd also have to be on the same network segment, by the way.
| |
| 13:52 | <alkisg> So if you have 10000 computers, there's a good change for collisions, if the numbers are random - which they aren't
| |
| 13:52 | Because the same vendor ships different numbers by design
| |
| 13:53 | <Hyperbyte> If you have 10.000 LTSP clients on network segment, you'd need fiber to handle all the bandwith.
| |
| 13:53 | *on one
| |
| 13:53 | :P
| |
| 13:53 | <alkisg> Yeah I think we can settle on keeping the last 6 digits (3 bytes) of the mac, if the other devs agree
| |
| 13:53 | And even make that the default, because using the IP as the default doesn't really help anywhere
| |
| 13:54 | While with the mac we can have ddns updates
| |
| 13:54 | <Hyperbyte> By the way
| |
| 13:54 | <alkisg> udhcpc -h calculated-hostname, ready
| |
| 13:54 | <Hyperbyte> Using only last number of IP as default, you have a lot more collision possibilities. :P
| |
| 13:54 | <alkisg> No, because it checks the netmask
| |
| 13:54 | So if you have 255.0.0.0 => 16m different hostnames there
| |
| 13:54 | <Hyperbyte> It does?
| |
| 13:54 | <alkisg> ltsp123476473
| |
| 13:55 | Yup, I wrote that part :D
| |
| 13:55 | <Hyperbyte> Ah.
| |
| 13:55 | Smart. :)
| |
| 14:04 | <stgraber> alkisg: done
| |
| 14:04 | <alkisg> stgraber: thanks! Btw how do you feel about changing the default ltsp client hostnames to mac-based?
| |
| 14:04 | We can have dynamic dns updates with udhcpc and ipconfig this way
| |
| 14:04 | Like, ltspa1b2c3
| |
| 14:05 | Implementation == udhcpc -h hostname, done
| |
| 14:07 | <stgraber> alkisg: I don't have a problem with that. I think it's going to be more difficult for users to give on the phone for tech support, but at the same time it'll be more accurate than what we have at the moment
| |
| 14:08 | <alkisg> Nice, I'll try it for some time and probably commit the change after a couple of weeks
| |
| 14:22 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 14:24 | Gadi has joined IRC (Gadi!~romm@ool-4571ca04.dyn.optonline.net) | |
| 14:24 | matrix3000 has joined IRC (matrix3000!~Eric@rrcs-70-61-255-227.central.biz.rr.com) | |
| 14:30 | fosser_josh has joined IRC (fosser_josh!~prathames@02799a2d.bb.sky.com) | |
| 14:30 | fosser_josh has left IRC (fosser_josh!~prathames@02799a2d.bb.sky.com) | |
| 14:36 | Yoe has joined IRC (Yoe!~wouter@samba.grep.be) | |
| 14:37 | <Yoe> anyone here who wants to do a talk about LTSP in the FOSDEM cross-distro devroom? (4-5 february, brussels)
| |
| 14:37 | * Yoe is desperately looking for an 'interesting' schedule... | |
| 14:45 | <Hyperbyte> Yoe, you should ask nubae probably.
| |
| 14:45 | Or knipwim.
| |
| 14:47 | dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174) | |
| 14:47 | vmlintu has left IRC (vmlintu!~vmlintu@dsl-hkibrasgw2-ff78c300-68.dhcp.inet.fi, Ping timeout: 240 seconds) | |
| 14:48 | jammcq has joined IRC (jammcq!~jam@70-91-230-209-BusName-Michigan.hfc.comcastbusiness.net) | |
| 14:48 | <Hyperbyte> Hey jammcq :)
| |
| 14:48 | <jammcq> hey Hyperbyte
| |
| 14:48 | how's it going?
| |
| 14:49 | <Hyperbyte> Busybusybusy. :)
| |
| 14:49 | But good.
| |
| 14:49 | Wednesday they're finally installing phone lines in our new office.
| |
| 14:49 | 100 lines total.
| |
| 14:50 | I only need two internet, since we use VoIP.
| |
| 14:50 | But for this type of building (it's in industrial area) they don't install less than 100.
| |
| 14:50 | heh
| |
| 14:50 | Only have to pay installation for 100 though, monthly costs just for two.
| |
| 14:50 | <jammcq> wow
| |
| 14:51 | <Hyperbyte> They start at 8 in the morning too
| |
| 14:51 | Which means I have to start at 8 in the morning as well, probably earlier. :(
| |
| 15:20 | vmlintu has joined IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi) | |
| 15:35 | mgariepy has left IRC (mgariepy!~mgariepy@ubuntu/member/mgariepy, Quit: Leaving) | |
| 15:54 | primeministerp has joined IRC (primeministerp!~ppouliot@nat/novell/x-btapkgzpqbcsdfok) | |
| 15:57 | mgariepy has joined IRC (mgariepy!~mgariepy@ubuntu/member/mgariepy) | |
| 15:57 | vongrippen has joined IRC (vongrippen!~vongrippe@75-149-119-43-Memphis.hfc.comcastbusiness.net) | |
| 16:02 | artista_frustrad has joined IRC (artista_frustrad!~artista_f@187.55.100.62) | |
| 16:16 | staffencasa has joined IRC (staffencasa!~staffenca@128-193-144-154.oregonstate.edu) | |
| 16:21 | vagrantc has joined IRC (vagrantc!~vagrant@c-76-105-252-50.hsd1.or.comcast.net) | |
| 16:22 | sbalneav has joined IRC (sbalneav!~sbalneav@205-200-22-198.static.mtsallstream.net) | |
| 16:22 | <sbalneav> Morning all
| |
| 16:23 | <mgariepy> morning scotty ;)
| |
| 16:24 | gentgeen__ has joined IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net) | |
| 16:25 | * vagrantc waves to sbalneav | |
| 16:26 | <ogra_> !s
| |
| 16:26 | <ltsp> ogra_: s: Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
| |
| 16:26 | gentgeen__ has left IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net, Remote host closed the connection) | |
| 16:26 | <ogra_> :)
| |
| 16:26 | <vagrantc> sbalneav: so, i goes to test this new libnss-sshsock, and i couldn't reproduce the problem with the old version...
| |
| 16:26 | sbalneav: uploaded anyways ... heh.
| |
| 16:30 | komunista has joined IRC (komunista!~slavko@adsl-195-098-014-097.dynamic.nextra.sk) | |
| 16:30 | <jammcq> sbalneav: Scotty !!!!!!!!!!!!!!!!!
| |
| 16:32 | gentgeen__ has joined IRC (gentgeen__!~kevin@c-98-236-71-64.hsd1.pa.comcast.net) | |
| 16:32 | * vagrantc waves to jammcq | |
| 16:32 | <jammcq> hey vagrantc
| |
| 16:34 | dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection) | |
| 16:35 | artista_frustrad has left IRC (artista_frustrad!~artista_f@187.55.100.62, Ping timeout: 244 seconds) | |
| 16:39 | <sbalneav> vagrantc: I fixed the problem. So, that's good :)
| |
| 16:39 | jammcq: hey ho
| |
| 16:41 | <vagrantc> sbalneav: and i uploaded it :)
| |
| 16:41 | <sbalneav> vagrantc: During December, stgraber and I are going to give a shot at trying to get the pam stuff going with a regular DM. You in?
| |
| 16:41 | hey ogra_
| |
| 16:41 | mgariepy: Bonjour, mes ami
| |
| 16:41 | <ogra_> yo
| |
| 16:41 | <lifeboy> Hey all, should I post a howto of my kernel compile for a Vortex86 processor based thin client for LTSP somewhere in the LTSP wiki or would it be more appropriate in the Ubuntu/Kernel wiki?
| |
| 16:42 | <vagrantc> sbalneav: sure :)
| |
| 16:42 | <ogra_> did you build an ubuntu package using the ubuntu source packages ?
| |
| 16:42 | if so, put it on the ubuntu wiki
| |
| 16:43 | <lifeboy> ogra_: yes
| |
| 16:43 | <ogra_> did you build a zImage/vmlinuz from upstream git trees, use the ltsp wiki
| |
| 16:44 | <lifeboy> The reason for the build is exclusively to provide LTSP needed modules, and I basically used the debian/rules scripts
| |
| 16:44 | <ogra_> uuh
| |
| 16:45 | you should have used fakeroot debian/rules editconfigs
| |
| 16:45 | <lifeboy> yes, I did, I just didn't say it all ;-)
| |
| 16:45 | <ogra_> ah, k :)
| |
| 16:45 | yeah, thats very ubuntu specific, shouldnt go into the upstream wiki
| |
| 16:46 | <vagrantc> lifeboy: so you were successful in the end?
| |
| 16:47 | <lifeboy> Yes. I was here while the US was still sleeping ;-)
| |
| 16:47 | <lifeboy> btw: alkisg and all the others that helped with my kernel compile problem: I asked in #ubuntu-kernel for help and smb and some others there assisted in getting the problem sorted and a kernel compiled with the modules I need.
| |
| 16:47 | <lifeboy> It seems that the "Local version - append to kernel release" string I added could have caused the problem.
| |
| 16:47 | <lifeboy> but thanks for all of you who helped and listened! :-)
| |
| 16:47 | artista_frustrad has joined IRC (artista_frustrad!~artista_f@187.55.100.62) | |
| 16:56 | khildin has left IRC (khildin!~khildin@ip-80-236-227-210.dsl.scarlet.be, Ping timeout: 245 seconds) | |
| 16:57 | Damianos has joined IRC (Damianos!~Damianos@adsl-070-145-074-043.sip.cha.bellsouth.net) | |
| 16:58 | lifeboy has left IRC (lifeboy!~roland@196-215-40-134.dynamic.isadsl.co.za, Quit: Time to go...) | |
| 16:58 | lifeboy has joined IRC (lifeboy!~roland@196-215-40-134.dynamic.isadsl.co.za) | |
| 17:01 | lifeboyy has joined IRC (lifeboyy!~roland@196-215-40-134.dynamic.isadsl.co.za) | |
| 17:02 | lifeboy has left IRC (lifeboy!~roland@196-215-40-134.dynamic.isadsl.co.za) | |
| 17:02 | lifeboyy has left IRC (lifeboyy!~roland@196-215-40-134.dynamic.isadsl.co.za, Client Quit) | |
| 17:02 | bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at) | |
| 17:02 | lifeboyy has joined IRC (lifeboyy!~roland@196-215-40-134.dynamic.isadsl.co.za) | |
| 17:03 | lifeboyy is now known as lifeboy | |
| 17:03 | lifeboy has left IRC (lifeboy!~roland@196-215-40-134.dynamic.isadsl.co.za, Client Quit) | |
| 17:04 | lifeboy has joined IRC (lifeboy!~roland@196-215-40-134.dynamic.isadsl.co.za) | |
| 17:04 | lifeboy has joined IRC (lifeboy!~roland@196-215-40-134.dynamic.isadsl.co.za) | |
| 17:08 | khildin has joined IRC (khildin!~khildin@ip166-155-209-87.adsl2.static.versatel.nl) | |
| 17:11 | <sbalneav> Gadi: Did you need any help from me on testing/integrating the nss_sshsock stuff?
| |
| 17:13 | <lifeboy> I posted a howto response to my original post on the XCore86 forums at http://www.deviceonchip.com/index.php?option=com_kunena&func=view&catid=5&id=631&limit=6&limitstart=12&Itemid=85#656 that describes how I eventually did it.
| |
| 17:17 | mistik1 has left IRC (mistik1!mistik1@unaffiliated/mistik1, Ping timeout: 244 seconds) | |
| 17:18 | artista_frustrad has left IRC (artista_frustrad!~artista_f@187.55.100.62, Ping timeout: 244 seconds) | |
| 17:22 | <Gadi> sbalneav: hey. Sorry - was on phone. No, I think it tested just fine for me. Don't think I have even pulled the latest/greatest as yet
| |
| 17:24 | <sbalneav> ah, cool
| |
| 17:25 | the new one doesn't go off into an infinite loop if it's placed before the "compat" setting.
| |
| 17:25 | <Gadi> I suppose that's a feature
| |
| 17:25 | :)
| |
| 17:26 | * vagrantc was unable to reproduce it with the old version before testing the new version | |
| 17:26 | <vagrantc> just like ltspfs
| |
| 17:31 | artista_frustrad has joined IRC (artista_frustrad!~artista_f@187.55.100.62) | |
| 17:39 | artista_frustrad has left IRC (artista_frustrad!~artista_f@187.55.100.62, Ping timeout: 260 seconds) | |
| 17:56 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 17:57 | vongrippen has left IRC (vongrippen!~vongrippe@75-149-119-43-Memphis.hfc.comcastbusiness.net, Quit: vongrippen) | |
| 17:58 | usuario has joined IRC (usuario!~usuario@189.169.105.217) | |
| 18:02 | artista_frustrad has joined IRC (artista_frustrad!~artista_f@187.6.155.105) | |
| 18:03 | hughessd has joined IRC (hughessd!~hughessd@173-164-117-109-Oregon.hfc.comcastbusiness.net) | |
| 18:06 | hughessd has left IRC (hughessd!~hughessd@173-164-117-109-Oregon.hfc.comcastbusiness.net, Remote host closed the connection) | |
| 18:07 | hughessd has joined IRC (hughessd!~hughessd@173-164-117-109-Oregon.hfc.comcastbusiness.net) | |
| 18:10 | loather has joined IRC (loather!~khudson@wsip-98-175-250-115.sd.sd.cox.net) | |
| 18:12 | lifeboy has left IRC (lifeboy!~roland@196-215-40-134.dynamic.isadsl.co.za, Quit: Time to go...) | |
| 18:13 | artista_frustrad has left IRC (artista_frustrad!~artista_f@187.6.155.105, Ping timeout: 258 seconds) | |
| 18:18 | loather has left IRC (loather!~khudson@wsip-98-175-250-115.sd.sd.cox.net, Quit: This computer has gone to sleep) | |
| 18:22 | usuario has left IRC (usuario!~usuario@189.169.105.217, Ping timeout: 276 seconds) | |
| 18:25 | artista_frustrad has joined IRC (artista_frustrad!~artista_f@187.6.155.105) | |
| 18:28 | <alkisg> vagrantc: if we switched to mac-based hostnames, e.g. ltspa1b2c3d4e5f6 (or just ltspa1b2c3 if we keep the 3-byte NIC part and skip the 3-byte vendor part), then we can run `udhcpc -h hostname` in the initramfs, and have dynamic DNS updates ==> static hostnames+DNS entries for all clients with no static entries in DHCP. Wouldn't that be a good idea?
| |
| 18:28 | klibc ipconfig supports a :::hostname: syntax, but I'm not sure if it sends it in the dhcp request
| |
| 18:29 | It should be very easy to send a patch for it, though
| |
| 18:31 | <mgariepy> alkisg, you simply change the HOSTNAME_EXTRA default ?
| |
| 18:31 | <vagrantc> alkisg: i don't want to have to dynamically generate usernames
| |
| 18:32 | based on mac addresses
| |
| 18:32 | <alkisg> mgariepy: no, unfortunately at that point, before the dhcp lease, we don't have access to lts.conf
| |
| 18:32 | But we can "calculate" the hostname we send to the dhcp request based on the mac address
| |
| 18:34 | vagrantc: now you create 255 users, right?
| |
| 18:35 | So if one has a 255.0.0.0 subnet, he'd get in trouble with that method. Maybe we can try some dynamic method?
| |
| 18:36 | E.g. we can have just as many user accounts as we have PCs, and dynamically map clients to user accounts
| |
| 18:36 | <vagrantc> alkisg: for each network, yes.
| |
| 18:37 | alkisg: one one network, the accounts are ltsp100-255, on another diskless100-255, on another build100-255
| |
| 18:37 | alkisg: but the number of clients may change
| |
| 18:38 | <alkisg> vagrantc: how about a kernel option that would enable mac-based hostnames? pxelinux.cfg/default => ltsp-hostnames=mac or something?
| |
| 18:38 | <vagrantc> alkisg: mac based hostnames could be the default, as long as it's easy to switch back to the current behavior
| |
| 18:38 | <alkisg> vagrantc: are those guest accounts, with their settings/files deleted after each logout?
| |
| 18:39 | <vagrantc> alkisg: basically (deleted daily)
| |
| 18:40 | <alkisg> OK let me think if some of the pam modules could be used to dynamically create user accounts, as needed...
| |
| 18:40 | <Gadi> uhh... that would be counter the stack, no?
| |
| 18:41 | I mean, tough to auth and create at the same time
| |
| 18:41 | <alkisg> Btw, we can try to have ddns updates even with ip-based hostnames. udhcpd -h ltsp123
| |
| 18:42 | Gadi: well, lightdm does have guest session support
| |
| 18:42 | I don't know how it does it, but it should be doable...
| |
| 18:42 | * Gadi thinks the user needs to exist before PAM | |
| 18:43 | <Gadi> if lightdm runs as root, then no doubt it can create the user as needed
| |
| 18:43 | before PAM
| |
| 18:43 | otherwise, maybe it runs as a dedicated user
| |
| 18:43 | on a PC, there is only one guest at a time, no?
| |
| 18:44 | <alkisg> No idea, I haven't tried to login as guest a second time simultaneously
| |
| 18:44 | * vagrantc also uses ssh keys for authenticating the guest accounts | |
| 18:44 | Steve_the_Pirate has left IRC (Steve_the_Pirate!~Gary@host109-155-255-130.range109-155.btcentralplus.com, Ping timeout: 240 seconds) | |
| 18:45 | <Gadi> as another data point, ltsp-cluster talks to a server side daemon to create a user and exchange an ssh key prior to initiating the ssh socket
| |
| 18:45 | then, sets that user as LDM_USERNAME
| |
| 18:45 | <alkisg> One thought would be to have the same guest account for everyone, and just a different (and temp) home for each. I don't know if fuse or chroots can help there.
| |
| 18:46 | Any security concerns there?
| |
| 18:46 | <Gadi> alkisg: not sure how nicely things that use /tmp will play that way
| |
| 18:46 | ie, /tmp/orbit-$USER
| |
| 18:46 | etc
| |
| 18:47 | <alkisg> There's a pam module for different tmp dirs
| |
| 18:47 | I've never used it - I'm just brainstorming :)
| |
| 18:47 | security concerns => about the ltsp-cluster method
| |
| 18:47 | ?
| |
| 18:47 | Could that be used without ltsp-cluster too, then?
| |
| 18:48 | <Gadi> afaik, yes. tho, not sure how much it needs the control center piece if at all
| |
| 18:48 | <mgariepy> alkisg, it could i guess. ltsp-cluster-accountmanager creates account on the server.
| |
| 18:49 | <alkisg> If the client can ask the server to create an ltspa1b2c3 user, then only as many user accounts as there are clients would be created, and wouldn't even need to be deleted afterwards
| |
| 18:49 | <mgariepy> Gadi, it doesn't req the control-center
| |
| 18:49 | currently the code is to make the user acount echo ip|base64
| |
| 18:50 | which is kinda ugly ;) haha
| |
| 18:50 | <alkisg> mgariepy: are there any security concerns there? E.g. if I have access to your network, could I flood your server with account creation requests?
| |
| 18:50 | ...or use them to login, etc etc?
| |
| 18:52 | <mgariepy> (echo -n "$LDM_SERVER " ; echo "getkey" | nc $nc_q_param $LDM_SERVER 8001 && echo) > /root/.ssh/known_hosts
| |
| 18:52 | cat P00-ltsp-cluster
| |
| 18:55 | if you install ltsp-cluster-accountmanager on ltsp-server and bypass [ -f /etc/ltsp/getltscfg-cluster.conf ] in /usr/share/ldm/rc.d/P00-ltsp-cluster it should work
| |
| 18:55 | accountmanager create a base64 named account base on incomming connexion
| |
| 18:55 | <alkisg> Gotcha, yeah that doesn't sound hard
| |
| 18:56 | <mgariepy> it's not hard, the code is there and already working ;)
| |
| 18:57 | Steve_the_Pirate has joined IRC (Steve_the_Pirate!~Gary@84.45.15.228) | |
| 19:01 | Damianos has left IRC (Damianos!~Damianos@adsl-070-145-074-043.sip.cha.bellsouth.net, Quit: Damianos) | |
| 19:07 | <alkisg> We have the same situation in the edubuntu live dvd too; it'd be nice if we were able to dynamically create accounts
| |
| 19:13 | <vagrantc> alkisg: i don't like the idea of shared accounts, even if the HOMEDIR mangling works properly
| |
| 19:14 | alkisg: users with the same UID would be able to read each other's homedirs, and probably even connect to their X sessions, etc.
| |
| 19:15 | libpam-tmpdir is largely broken because many applications ignore the TMPDIR environment variable (and other variables) for ... security reasons.
| |
| 19:15 | it would be interesting to dynamically create user accounts on client boot
| |
| 19:15 | though i typically use many login servers independent of one another
| |
| 19:18 | <muppis> alkisg, it is hard to patch that default hostname generation?
| |
| 19:18 | <alkisg> vagrantc: if it's possible to create fuse filesystems that only processes on the same group can see, maybe it's possible to isolate both /home/username and /tmp then... no idea if it's technically possible. But ok dynamic account creation on boot is surely doable, and if it's based on mac then the same client will always get the same user account too
| |
| 19:18 | <muppis> Because I'd like to have it working tomorrow. ;)
| |
| 19:18 | <alkisg> muppis: how do you need to change it?
| |
| 19:19 | You mean to have ddns updates by tomorrow?
| |
| 19:19 | <muppis> Yes.
| |
| 19:19 | <alkisg> Just put the -h parameter to the udhcpc call in the initramfs
| |
| 19:19 | You don't need to modify the hostname, use the default ltsp123 one
| |
| 19:19 | We have code for that already in the initramfs somewhere
| |
| 19:20 | <muppis> Nice.
| |
| 19:20 | <alkisg> muppis: you're using udhcpc, right?
| |
| 19:20 | Not ipconfig or ipappend 3...
| |
| 19:20 | <muppis> Sure thing.
| |
| 19:22 | <alkisg> Errr sorry that can't be done, it needs to be mac-based
| |
| 19:23 | The code for mac-based hostnames is in nfs-bottom/ltsp
| |
| 19:23 | <muppis> Then I have to figure out something.
| |
| 19:23 | <alkisg> case "$HOSTNAME_EXTRA" in
| |
| 19:23 | mac)
| |
| 19:23 | HOSTNAME_EXTRA=$(ip link show $DEVICE | awk '/ether/{print $2}' | tr ':' '-')
| |
| 19:24 | That's what you need to calculate the hostname, and then put the result to udhcpc -h $hostname
| |
| 19:24 | mistik1 has joined IRC (mistik1!mistik1@unaffiliated/mistik1) | |
| 19:24 | <alkisg> And also put HOSTNAME_EXTRA=mac in lts.conf
| |
| 19:26 | muppis: btw, another way is to put HOSTNAME_EXTRA=mac, and just put all your clients in your dns, statically
| |
| 19:26 | In /etc/hosts or in your dnsmasq configuration
| |
| 19:26 | No need to touch any code at all, this way. No ddns updates, but the dns resolution works.
| |
| 19:27 | Steve_the_Pirate has left IRC (Steve_the_Pirate!~Gary@84.45.15.228, Ping timeout: 240 seconds) | |
| 19:27 | <alkisg> Meh again I'm saying stupid things, you don't have their IPs
| |
| 19:28 | <muppis> That's what I'm trying to avoid, static information in dnsmasq.
| |
| 19:29 | Because dnsmasq is going to be service that can be thrown from server to other by Pacemaker.
| |
| 19:34 | artista_frustrad has left IRC (artista_frustrad!~artista_f@187.6.155.105, Ping timeout: 276 seconds) | |
| 19:43 | <alkisg> muppis: yeah you'd need that mac code I pasted above then. Put it in the udhcpc script in the initramfs.
| |
| 19:44 | <muppis> I try remember that at tomorrow in work. :)
| |
| 19:46 | But doesn't that cause problems if clietn got multiple nics and all are connected?
| |
| 19:47 | <vagrantc> oh fun.
| |
| 19:48 | curiously, my syslinux documents only IPAPPEND 1, 2 and 4
| |
| 19:49 | * vagrantc wonders what happened to 3 | |
| 19:49 | <alkisg> muppis: no, it shouldn't matter, any of those nics would be unique too
| |
| 19:49 | vagrantc: those can be binary OR'ed, afaik
| |
| 19:49 | What is ipappend 4?
| |
| 19:49 | <muppis> alkisg, ok then.
| |
| 19:49 | <vagrantc> alkisg: appends SYSUUID
| |
| 19:50 | whatever that is
| |
| 19:50 | <alkisg> Um, /me googles...
| |
| 19:50 | <vagrantc> we should default to ipappend 2
| |
| 19:53 | <alkisg> Ah it's the long UUID right next to the MAC, http://sirlagz.net/?p=299
| |
| 19:54 | vagrantc: yes - and I think there was work in pxelinux recently to pass the lease info to the initrams too, maybe that way we could even completely avoid the second dhcp request
| |
| 19:55 | We'd set a "static" IP with the info we get from the pxe stack, and finally run dhclient in the real file system to manage the lease
| |
| 19:56 | komunista has left IRC (komunista!~slavko@adsl-195-098-014-097.dynamic.nextra.sk, Quit: Leaving.) | |
| 19:59 | <alkisg> vagrantc: btw, we could also default to IPAPPEND 3, skip the dhcp request in the initramfs completely, and run dhclient in the real file system afterwards and request the same IP from there
| |
| 19:59 | Is there any reason why the dhcp server would refuse that request?
| |
| 20:01 | <muppis> Too strict configuration.
| |
| 20:01 | <alkisg> ...and if it is, then it would be possible to refuse it on lease renewals too, thus crashing the client
| |
| 20:01 | <muppis> But why in world any dhcpd refuses renewals?
| |
| 20:03 | <alkisg> Hmm the down side is that we lose the info provided by dhcp at initramfs, we only have next-server, gateway etc there, so we'd have to rely on next-server, lts.conf etc for the whole configuration
| |
| 20:09 | <vagrantc> alkisg: one problem is that the dhcp server might hand you a new ip address, and that doesn't typically go over well.
| |
| 20:10 | <alkisg> vagrantc: yes, but we have the same problem later on too, on lease renewals, we can't really do anything from the client side for this, it needs to be cared of on the server side
| |
| 20:10 | <muppis> vagrantc, doesn't that usually mean too short lease time?
| |
| 20:10 | Or response from wrong dhcpd.
| |
| 20:10 | <vagrantc> could be either
| |
| 20:11 | but short lease times is a good way to prevent DoS attacks on your DHCP server.
| |
| 20:11 | <alkisg> For lease renewals typically the server that issued the lease answers
| |
| 20:11 | <vagrantc> well, not a good way, but a way ...
| |
| 20:12 | i've seen the occasional broken card which triggers extra DHCP requests till there are no ip addresses left
| |
| 20:13 | <alkisg> Doesn't the dhcp server send the same IP in this case? (same mac, same clientid, lease not yet expired?)
| |
| 20:13 | <vagrantc> depends on how broken the card is
| |
| 20:13 | <alkisg> Mhm
| |
| 20:13 | <vagrantc> when dealing with things not working as they should, all bets are off :)
| |
| 20:14 | thankfully, that's been rare
| |
| 20:14 | <alkisg> What dhcp info would we need in the initramfs, that we couldn't get from dhclient later on on the real file system?
| |
| 20:14 | (I mean, if we always used IPAPPEND 3)
| |
| 20:14 | (so we only had next-server, BOOTIF, and gateway)
| |
| 20:14 | * vagrantc wonders if the absence in documentation suggests ipappend 3 is deprecated | |
| 20:15 | <alkisg> I've read an answer by hpa about the OR'ing bit
| |
| 20:15 | So as long as IPAPPEND 1 and 2 exist, 3 will exist too
| |
| 20:16 | But I think I've also read him say that "using IPAPPEND 1 means that your network configuration is broken", I don't know if he still thinks that way or if he thinks about deprecating IPAPPEND 1...
| |
| 20:17 | mistik1_ has joined IRC (mistik1_!mistik1@unaffiliated/mistik1) | |
| 20:17 | mistik1 has left IRC (mistik1!mistik1@unaffiliated/mistik1, Read error: No route to host) | |
| 20:17 | <alkisg> On the other hand, the new code that notifies the initramfs about the pxe lease suggests they still care about that part
| |
| 20:17 | mistik1_ is now known as mistik1 | |
| 20:17 | Parker955_Away is now known as Parker955 | |
| 20:17 | khildin has left IRC (khildin!~khildin@ip166-155-209-87.adsl2.static.versatel.nl, Ping timeout: 248 seconds) | |
| 20:26 | _UsUrPeR_ has left IRC (_UsUrPeR_!~jsass@fw.acurrus.com, Quit: Ex-Chat) | |
| 20:27 | _UsUrPeR_ has joined IRC (_UsUrPeR_!~jsass@fw.acurrus.com) | |
| 20:29 | jammcq has left IRC (jammcq!~jam@70-91-230-209-BusName-Michigan.hfc.comcastbusiness.net, Quit: leaving) | |
| 20:30 | dberkholz has left IRC (dberkholz!UNKNOWN@gentoo/developer/dberkholz, Read error: Operation timed out) | |
| 20:48 | <shawnp0wers> For those users managing large groups of people (schools, etc) -- what backend are you using for user authentication?
| |
| 20:48 | <muppis> LDAP
| |
| 20:48 | <shawnp0wers> do you interface with non-Linux OSes?
| |
| 20:48 | I mean, do those accounts work for Windows clients, OSX clients, etc?
| |
| 20:48 | Or do you strictly use Linux?
| |
| 20:49 | I'm currently using OSX server for my user authentication, but I'm running into serious limits when it comes to authenticating with services (web filtering, etc)
| |
| 20:49 | <muppis> They can be managed to work with Windows, but haven't done that.
| |
| 20:50 | <shawnp0wers> I see. One solution for me would be to get an Active Directory server to handle user authentication -- but that is expensive, and quite frankly not attractive
| |
| 20:50 | Largely because I'm clueless when it comes to Windows administration
| |
| 20:51 | It seems like every appliance in the world now supports Active Directory, for obvious reasons, but I'd like to avoid that if possible.
| |
| 20:51 | So, I'm curious what most folks here use, as I use LTSP for better than half my workstations.
| |
| 20:52 | <muppis> Google OpenLDAP for Windows.
| |
| 20:53 | <shawnp0wers> Oh, I'd MUCH rather run OpenLDAP under Linux -- my concern is compatibility with Windows/OSX clients.
| |
| 20:54 | I'd like to avoid a Windows server if at all possible. :D
| |
| 20:54 | <muppis> That's a client for Windows, not server.
| |
| 20:55 | <shawnp0wers> oh, ok, gotcha
| |
| 20:55 | my googling found a server package. :)
| |
| 20:55 | I misunderstood. :)
| |
| 20:55 | <muppis> Funny. I founded a client. :)
| |
| 20:57 | map7 has left IRC (map7!~map7@ppp118-209-114-57.lns20.mel4.internode.on.net, Ping timeout: 260 seconds) | |
| 21:01 | brunolambert has left IRC (brunolambert!~bruno@nat/revolutionlinux/x-gdfsinsvemwhggol, Quit: brunolambert) | |
| 21:02 | brunolambert has joined IRC (brunolambert!~bruno@nat/revolutionlinux/x-wqxemyeqwzfewjdr) | |
| 21:28 | primeministerp has left IRC (primeministerp!~ppouliot@nat/novell/x-btapkgzpqbcsdfok, Remote host closed the connection) | |
| 21:57 | brunolambert has left IRC (brunolambert!~bruno@nat/revolutionlinux/x-wqxemyeqwzfewjdr, Quit: brunolambert) | |
| 22:02 | dberkholz has joined IRC (dberkholz!UNKNOWN@smtp.gentoo.org) | |
| 22:02 | dberkholz has joined IRC (dberkholz!UNKNOWN@gentoo/developer/dberkholz) | |
| 22:22 | sbalneav has left IRC (sbalneav!~sbalneav@205-200-22-198.static.mtsallstream.net, Quit: leaving) | |
| 22:42 | <matrix3000> google-fu ahahaha
| |
| 22:42 | learned a new slang term today
| |
| 22:53 | <shogunx> hey anyone... squeeze ltsp related. how to enable local devices?
| |
| 22:54 | <vagrantc> should just be enabled by default
| |
| 22:54 | shogunx: still have the add the users to fuse group on the server
| |
| 23:01 | <shogunx> okie. thanx vagrantc. i should get a chance to hack on the hp arm client soon, but I really wanted to plant the raspberrypi as a ltsp client in your head on that end;)
| |
| 23:02 | <vagrantc> i guess the hp t5325 was discontinued in favor of a newer model?
| |
| 23:02 | which maybe means it's possible to get on the cheap...
| |
| 23:03 | <shogunx> thats a good thought too.
| |
| 23:04 | <vagrantc> i suspect hardware into my hands would be more useful than in my head :)
| |
| 23:04 | <shogunx> if i win the buy first release lottery (i click fast enough after they announce they are for sale) i will send you one.
| |
| 23:05 | <vagrantc> heh
| |
| 23:07 | <alkisg> Thin client hats... not a bad idea :P
| |
| 23:07 | <vagrantc> what's the raspberrypi like?
| |
| 23:07 | <shogunx> its a little $25 arm board that does hd video, apparently.
| |
| 23:08 | <alkisg> But no lan by default, right?
| |
| 23:08 | <shogunx> if it does pxe, or we can encourage it to, it makes for a very compelling solution.
| |
| 23:08 | no, theres lan i think.
| |
| 23:08 | <vagrantc> indeed!
| |
| 23:09 | <alkisg> The Model B version of the device includes 10/100 wired Ethernet. There is no Ethernet on the Model A version (which we expect to be taken up mostly by the education market), but Wi-Fi will be available via a standard USB dongle.
| |
| 23:09 | So that would be the $35 model
| |
| 23:09 | <shogunx> Does the device support networking? Is there Wi-Fi?
| |
| 23:09 | The Model B version of the device includes 10/100 wired Ethernet. There is no Ethernet on the Model A version (which we expect to be taken up mostly by the education market), but Wi-Fi will be available via a standard USB dongle.
| |
| 23:09 | heh
| |
| 23:09 | beat me to it.
| |
| 23:09 | <alkisg> :)
| |
| 23:09 | <shogunx> umm, ok. 35.
| |
| 23:09 | still cheap like nickel beer night.
| |
| 23:10 | <alkisg> I had a lot of time since I last looked in that page, I think I'd only seen the A model
| |
| 23:10 | Yeah... you'd need to buy the power adapter separately, right?
| |
| 23:11 | PSU or however else it's called
| |
| 23:11 | <shogunx> $35 for a thin client board that can be hacked into the back of and adequately drive a 23" display sounds pretty good to me.
| |
| 23:11 | well, i tackle that a little different.
| |
| 23:11 | all the ltsp systems i put out there run on solar power.
| |
| 23:12 | so i regulate to 12 and 5 when necessary coming off my @12V battery bank anyway.
| |
| 23:12 | no inverters... no wall warts.
| |
| 23:12 | low voltage DC from source to load.
| |
| 23:13 | <alkisg> Sounds nice, as long as they don't want to work with them at nights :D
| |
| 23:13 | <shogunx> thats what the battery bank in fot
| |
| 23:13 | for ^^
| |
| 23:13 | <vagrantc> how big of a solar array for how much battery runtime for a lab of how many clients?
| |
| 23:14 | <shogunx> and the panel array and battery bank can keep it going 24/7 if necessary, tho most close down their clients and displays at 8 or 9.
| |
| 23:14 | vagrantc, i oversize the array and bank a bit to ensure 5 9's on the power system.
| |
| 23:14 | <vagrantc> sure
| |
| 23:15 | <shogunx> the last one i put in i used 6x 135 Kyocera panels.
| |
| 23:15 | there were 3 strings of 400Ah 6V batteries, 2 batteries in series per string
| |
| 23:16 | so 1200Ah @ 20h rate @ 12VDC.
| |
| 23:17 | <vagrantc> impressive
| |
| 23:17 | <shogunx> this powers a dual core turion server w 4gb of memory, a 24 port10/100 2 port 10/100/1000 cisco hub, 5 geode based thin clients with 23" led backlit displays.
| |
| 23:17 | and some basic UBNT wifi gear.
| |
| 23:18 | if the panels blow off the roof, it will keep the whole network running for a week.
| |
| 23:18 | <alkisg> What are the power requirements? The device is powered by 5v micro USB. You can read more about it here.
| |
| 23:18 | ==> so if I have a usb hub in my monitor, that would be enough to power it?!
| |
| 23:18 | <shogunx> or if you turn of the terminals, it will run the server for a month.
| |
| 23:18 | yep
| |
| 23:18 | or I just use a DC-DC converter.
| |
| 23:19 | <alkisg> I can even buy usb hubs with 5€, those should do as well
| |
| 23:19 | <shogunx> i like the geode clients, but they are getting long in the tooth, and do not support reasonable resolutions.
| |
| 23:19 | they take 4-5W each.
| |
| 23:20 | <alkisg> So clients under 40€... can't wait :)
| |
| 23:20 | <shogunx> ARM should lower that and improve performance as well, particularly on the display.
| |
| 23:26 | there are some pics the most recent install going in interspersed in this vid: http://www.youtube.com/watch?v=FTpsMK61S3w
| |
| 23:35 | <alkisg> "So in theory, could I power the Raspberry Pi off one port of a powered USB hub? Then also plug that USB hub back into the Raspberry Pi and use it to power my external hard drive? So then I only have one power supply to plug into the wall?" ==> they say that's doable, so we get multiple usb ports + power at the price of a single hub, very good
| |
| 23:38 | <shogunx> :)
| |
| 23:38 | dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Leaving...) | |
| 23:38 | <shogunx> clients under 4W too!
| |
| 23:40 | and then we put the usb hub, the display, and the raspberry pi board in one chassis, with a power bus and the necessary DC-DC conversion or regulation, and booya!
| |
| 23:41 | btw vagrantc, you could go up to about 20 clients 8 hours a day on that solar power supply without a problem.
| |
| 23:44 | each panel gets a breaker, and breakers seperate the panels from the charge controller, the charge controller from the batteries, and the the batteries from the load. morningstar ts-60 charge controller.
| |
| 23:44 | battery neg is connected to earth.
| |
| 23:47 | <vagrantc> shogunx: very cool
| |
| 23:48 | monteslu has left IRC (monteslu!~monteslu@ip68-109-174-213.ph.ph.cox.net, Read error: Operation timed out) | |
| 23:49 | <shogunx> when i say open source i mean it:)
| |
| 23:55 | bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Ping timeout: 255 seconds) | |
| 23:57 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |