| === Taj is now known as manakfula | ||
| manakfula | Hello everyone, I have a couple quick questions about edubuntu | 02:47 |
|---|---|---|
| manakfula | I'm familiar with ubuntu so it's nothing really technical | 02:47 |
| manakfula | and I apologize if I get disconnected, my wireless sucks...I'll be back if that happens | 02:48 |
| manakfula | does edubuntu come with a spanish distribution or spanish language option at the time of installation? | 02:49 |
| manakfula | (im not sure if that whole message went through or not I only see part of it) | 02:49 |
| manakfula | I'll retype it incase it didn't...does it come with a spanish disto or installation option? | 02:50 |
| manakfula | or just the spanish learning/tutor program? | 02:51 |
| === Taj is now known as Manakfula | ||
| Manakfula | sorry about that told you my wireless sucks | 02:55 |
| Manakfula | so is anyone here or is the room just idle? | 02:55 |
| davew | Am trying to use the profile editor in 9.04 but there is a bug in it. Actually this is one feature that I need as I am using LTSP with 30 clients. Anyone got any ideas | 11:20 |
| alkisg | davew: sbalneav is trying to fix the bugs and hopes it'll be ready for 9.10. There's a new version at his PPA that works for some people. | 11:21 |
| davew | Have tried the sbalneav fix and it didn't work for me - are you saying it doesn't work for everybody ? | 11:22 |
| alkisg | Yup. I don't know the last ubuntu version that sabayon worked in, maybe it was back in gutsy or something... | 11:23 |
| alkisg | davew: what exactly do you want to lock with sabayon? Maybe there are ways to do it with gconf... | 11:23 |
| davew | Oh dear I'm in trouble. Without this I don't know what I can do !!! | 11:24 |
| davew | Do you know when 9.10 which I presume is the next version is out ? | 11:25 |
| alkisg | 9 = 2009, 10 = 10th month = october | 11:26 |
| davew | Understood - I am in trouble ! I understand that Sabayon has not worked for about 12 months, is that true ? | 11:28 |
| alkisg | Well, I bet it's working right now for people that have older ubuntu versions... | 11:29 |
| alkisg | But yeah, I don't think it has worked in the recent ubuntu releases | 11:30 |
| davew | O.K. Thanks for that I will have to strip out 9.04 and put in something older or another distro which has similar features. Thanks for your help I have been trying to get some information for several days. Cheers . | 11:32 |
| alkisg | davew: but again, it may be possible to lock down your desktop for your clients with gconf-editor | 11:32 |
| davew | Will this allow me to do it globally for say a particular group ? | 11:33 |
| alkisg | Hmmm kind of. Global locking is global, not per group. But you could copy the settings on all members of a group. | 11:34 |
| davew | O.K. I'll have a look - thanks again . | 11:35 |
| alkisg | Bye | 11:35 |
| mss | hello ... I have been facing the hardy release DHCP issue seen by others such as : | 16:54 |
| mss | https://lists.ubuntu.com/archives/edubuntu-users/2008-December/004911.html | 16:54 |
| mss | above | 16:54 |
| mss | Today I captured some logs from my system, turning on one client at a time ...and logs show how slow things get .. | 16:54 |
| mss | where can I upload so experts can take a look ? | 16:55 |
| mss | I have captured daemon and kern logs among others | 16:56 |
| mss | hi alkisg | 18:05 |
| alkisg | Hello mss, hi all | 18:06 |
| mss | we spoke a few days ago about the hardy DHCP problem : | 18:06 |
| mss | https://lists.ubuntu.com/archives/edubuntu-users/2008-December/004911.html | 18:06 |
| mss | I have daemon, kern and other logs showing the problem | 18:06 |
| mss | I booted clients one at a time ....the first two boot fast .. the next few start slowing down ... | 18:07 |
| alkisg | mss, it seems like your case is a dhcp server problem, but then again, I'm not sure, e.g. "slowing down"? Aquiring a dhcp lease should only take some msec... | 18:08 |
| mss | yes .. I understand that ....but it has now been seen by several folks that this DHCP "slowing down" problem exists ... | 18:09 |
| mss | in some cases .... | 18:09 |
| alkisg | Any links for bug reports? | 18:10 |
| alkisg | mss: an easy thing to try would be another dhcp server, e.g. dnsmasq | 18:11 |
| alkisg | This way you would be sure if it is or not a dhcp3-server - related problem | 18:11 |
| mss | I could try that ..but am not clear on exactly how to set up a different dhcp server ... other folks who have moved to a different dhcp server, fould it works fine . | 18:13 |
| alkisg | mss: also, try to post in #ltsp-discuss, you may have additional chances for an answer there... | 18:13 |
| mss | https://lists.ubuntu.com/archives/edubuntu-users/2008-December/004914.html | 18:13 |
| mss | looks like nobody has filed a bug yet ... | 18:14 |
| alkisg | "We figured that as the box got busy serving up the tftp kernels the dhcp would start to fail/timeout." I haven't heard of anyone having this problem, though... | 18:14 |
| alkisg | You may try dnsmasq like this: | 18:15 |
| alkisg | sudo invoke-rc.d dhcp3-server stop | 18:15 |
| alkisg | sudo apt-get install dnsmasq | 18:15 |
| alkisg | create an /etc/dnsmasq.d/ltsp.conf file with: | 18:15 |
| alkisg | port=0 # this makes dnsmasq *not* function as a dns server | 18:16 |
| alkisg | dhcp-range=10.160.31.100,10.160.31.120,8h # change with your specific range | 18:16 |
| alkisg | sudo invoke-rc.d dnsmasq restart | 18:16 |
| mss | The logs I captured show exactly htis problem . | 18:16 |
| alkisg | ===> after that, try to boot the clients. | 18:16 |
| mss | thanks .. I will try dnsmasq like you sugested .. | 18:17 |
| alkisg | mss: do you have a small log part that shows that the server gets a dhcp_discover, and doesn't answer with a dhcp_offer? | 18:17 |
| alkisg | If so, you should definately file a bug against dhcp3-server... | 18:18 |
| mss | yes ...I have the logs ... it answers .. but very slowly .. let me pul it up for you ... | 18:18 |
| mss | here's the first client . | 18:20 |
| mss | Aug 9 12:48:54 coco dhcpd: DHCPDISCOVER from 00:1e:0b:39:fa:d1 via eth1 | 18:20 |
| mss | Aug 9 12:48:55 coco dhcpd: DHCPOFFER on 192.168.0.245 to 00:1e:0b:39:fa:d1 via eth1 | 18:20 |
| mss | Aug 9 12:48:57 coco dhcpd: DHCPREQUEST for 192.168.0.245 (192.168.0.254) from 00:1e:0b:39:fa:d1 via eth1 | 18:20 |
| mss | Aug 9 12:48:57 coco dhcpd: DHCPACK on 192.168.0.245 to 00:1e:0b:39:fa:d1 via eth1 | 18:20 |
| alkisg | The client needed 2 seconds to process the offer and request it? !!! | 18:21 |
| mss | here's the next client by which time things start to slow down | 18:21 |
| mss | Aug 9 13:02:44 coco dhcpd: DHCPDISCOVER from 00:1e:0b:39:fa:ef via eth1 | 18:21 |
| mss | Aug 9 13:03:09 coco dhcpd: DHCPOFFER on 192.168.0.247 to 00:1e:0b:39:fa:ef via eth1 | 18:21 |
| mss | Aug 9 13:03:34 coco dhcpd: DHCPREQUEST for 192.168.0.247 (192.168.0.254) from 00:1e:0b:39:fa:ef via eth1 | 18:21 |
| mss | Aug 9 13:03:59 coco dhcpd: DHCPACK on 192.168.0.247 to 00:1e:0b:39:fa:ef via eth1 | 18:21 |
| alkisg | Yeah that *doesn't* look like a dhcp problem to me, but a network problem | 18:22 |
| alkisg | If it was a server problem, the client wouldn't need (34-09) seconds to process the offer and send a request | 18:22 |
| mss | the strange thing is ....even before the DHCPDISCOVER shows up in the log ... I can see the client has already acquired the address and on its way .. | 18:23 |
| mss | but the DHCPACK shows up much later in the log .. | 18:23 |
| alkisg | There are 2 dhcp requests for each client, did you know that? | 18:24 |
| alkisg | One during the pxe boot, and one later *after* the kernel boots | 18:24 |
| mss | yes .... I showed only the first of each client | 18:24 |
| alkisg | Ah... well then I still don't see this as a dhcp server problem | 18:25 |
| alkisg | If a client is booted and you try to operate it when other clients boot, does it function correctly, or is it very very slow? | 18:25 |
| alkisg | E.g. having both full duplex and half duplex clients on the same switch can bring down the network speed to a crawl... | 18:25 |
| mss | well....the first three- four clients boot.... | 18:26 |
| mss | then the rest do not get any response in time before the client times out .. | 18:26 |
| mss | all clients are identical .. hp5135s | 18:27 |
| alkisg | ok, but during this "clients do not get any response" phase, if you try to sit down on a client a work on it, does it work fine, or is it slow? If it were slow, this would suggest a network problem... | 18:27 |
| mss | the clients that are up seem to work fine .... | 18:28 |
| alkisg | Can you upload a full log file somewhere? | 18:28 |
| mss | yes ..have all the logs | 18:28 |
| mss | can I email ... only 400k | 18:29 |
| alkisg | OK... my nickname at gmail dot com | 18:29 |
| alkisg | also, what NICs do the clients have? | 18:30 |
| mss | clients ? no idea, but I know other have been using these clients fine .... | 18:35 |
| alkisg | With the same distro? | 18:36 |
| alkisg | (and version)? | 18:36 |
| mss | sent email - it is a zip file of all logs | 18:36 |
| mss | yes ... | 18:36 |
| mss | I believe Asmo Koskinen asmo.koskinen at arkki.info is using them .. | 18:37 |
| mss | saw him mention in one of his posts | 18:37 |
| mss | 8.04.10 .. last updated 12 days ago with all current (at that time ) updates | 18:37 |
| mss | client has 128 MB ram ... shoud that be an issue ? | 18:38 |
| alkisg | No | 18:39 |
| alkisg | ...but even with the first client, there are problems: http://alkisg.pastebin.com/d2b2bbb97 | 18:40 |
| alkisg | Even though the client gets a dhcp_ack, it requests the same IP again | 18:40 |
| mss | I thought that was the kernel | 18:40 |
| alkisg | There are 3 acks tehre | 18:40 |
| alkisg | 1 = pxe, 2 = kernel, 3 = problem | 18:41 |
| mss | every client does this twice .... | 18:41 |
| mss | but there are only 2 DISC/OFFER s | 18:49 |
| mss | you are saying the last pair of lines REQUEST/ACK is a problem ? | 18:49 |
| alkisg | I'm saying that it seems to ignore the "12:49:16 coco dhcpd: DHCPACK" | 18:50 |
| alkisg | So, if it doesn't get the first ack, it seems like a network problem exists even with 1 client | 18:50 |
| mss | or if the dhcp server is slow and it gets two reqs one after the other ... is it possible it is responding to both ? | 18:52 |
| alkisg | Hmmmm though I had send a patch last year for ipconfig that had to do with DHCPNAKs, maybe it didn't make it for 8.04 and that bug is still there | 18:53 |
| alkisg | http://www.zytor.com/pipermail/klibc/2008-June/002311.html | 18:54 |
| mss | it says your patch was applied :-) | 19:02 |
| alkisg | Yeah I know, but I don't know if it make it for hardy | 19:02 |
| alkisg | I think it's only fixed in 8.10 and 9.04... | 19:02 |
| mss | thought it should have at least made it to 8.04.10 | 19:02 |
| alkisg | But anyway mss I don't think that ipconfig is your main problem here | 19:03 |
| alkisg | I mean, in the second client dhcp takes dozens of seconds!!! | 19:04 |
| mss | I will try the dnsmasq and see how it goes ... I have the ltsp setup at a remote location ( no internet) .. so I can take the dnsmasq.deb file and install .. correct ? | 19:04 |
| mss | yes ... and then it gets wors .... | 19:04 |
| alkisg | I think you'll need 2 debs, dnsmasq.deb and dnsmasq-base.deb | 19:04 |
| alkisg | But do try this: | 19:04 |
| alkisg | 1. Boot 4-5 clients (as many as they boot) | 19:05 |
| alkisg | 2. Try to work on one of them. Does it work ok? | 19:05 |
| mss | but when I left the system alone for a few hours and came back and booted a client .. the first one again came up fast and fine ... then the second one slowed .. and the process repeated ... | 19:05 |
| alkisg | Ah | 19:05 |
| mss | it is in the logs ... | 19:05 |
| alkisg | 3. After a while, open another client and see if it boots fast <== you just answered that | 19:06 |
| mss | after 14:xx . I came back at 17:xx | 19:06 |
| alkisg | You don't even have dhcpd.conf in the logs, though... | 19:06 |
| mss | in the logs | 19:06 |
| alkisg | Nor /etc/network/interfaceds | 19:06 |
| mss | dhcpd conf standard as edubuntu requires .. done during installation .. nothing special | 19:07 |
| mss | attached ifconfig .. if it helps | 19:07 |
| alkisg | I wonder if it's a tftp problem, i.e. the whole network slows down when the clients get the kernel via tftp | 19:07 |
| mss | possible ... could you take a loo at kern logs ... it just keeps going on for a client log after the login screen is presented on the clinet ..is that a problem ? | 19:08 |
| mss | but each client should take no more than 10 secs to get the image | 19:09 |
| mss | on a FE LAN | 19:09 |
| alkisg | Yeah, unless there's a problem :) (which is the only thing certain here :)) | 19:10 |
| alkisg | I've seen tftp needing half an hour to server a kernel in problematic cases... | 19:10 |
| alkisg | *serve | 19:11 |
| mss | :-) | 19:11 |
| mss | if the PHY is messed up and retransmissions happen .. I guess it can happen | 19:11 |
| alkisg | How difficult would it be to try out 9.04? E.g. in a different partition? | 19:11 |
| alkisg | The installation takes < 30 mins... | 19:12 |
| mss | I am trying to set up a different system in a couple of days .. I can try 9.04 on that | 19:12 |
| mss | is it a LTS release ? | 19:12 |
| alkisg | No, but it was released a year later than 8.04, so they'll be supported for about the same time (only 6 months difference) | 19:13 |
| mss | so .. I will try both the dnsmasq and the 9.04 and let you know in a day or two ... | 19:14 |
| mss | thanks for your help | 19:15 |
| alkisg | mss, ps: | 19:15 |
| alkisg | /usr/share/doc/ltsp-server/examples/dhcpd-dnsmasq <== example file for dnsmasq config | 19:15 |
| * alkisg gotta go... bye all | 19:15 | |
| mss | thanks | 19:16 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!