[02:23] <vipaca> Hello all is ther a seperate channe; forsparc ?
[02:35] <ScottK> faulkes-: Thanks.
[03:26] <vipaca> Hello anybody out there dealing with sparc
[03:27] <dendrobates> vipaca: I have a couple sparc boxes
[03:27] <vipaca> really
[03:28] <vipaca> Im really a sun hardware noob
[03:28] <vipaca> but I inheireted two v210s
[03:28] <vipaca> I just ubuntu on one of them
[03:29] <vipaca> installed fine I was suspect of network connectivty
[03:29] <vipaca> now that the box is up the nics are acting falkey
[03:29] <vipaca> dendrobates: what kinda boxes do you have?
[03:31] <dendrobates> vipaca: t2000, t5200
[03:32] <dendrobates> vipaca: or something like that I am unsure of the model of the second one.  But they are both newer.
[03:33] <vipaca> I see these are newerthen v210?
[03:33] <vipaca> there definatley beefer
[03:34] <vipaca> what kinda of nics you got?
[03:37] <dendrobates> I believe they are marketed as Sun gigaswift, or something like that.
[03:38] <dendrobates> but they are intel GE.
[03:38] <vipaca> when you do lspci
[03:39] <vipaca> Did you run into any nic problems when you installed ubuntu
[03:40] <dendrobates> I did not.  What version are you installing?
[03:40] <vipaca> 7.1
[03:40] <vipaca> LATEST
[03:42] <dendrobates> are you seeing anything in dmesg?
[03:43] <vipaca> let me check
[03:44] <helfire> hey, is it easy to change from php to suphp?
[03:45] <vipaca> I hope this doesn
[03:46] <vipaca> get me hacked
[03:46] <vipaca> http://pastie.caboo.se/151347
[03:46] <vipaca> dendrobates:Shouldn't be anything dmesg that could :/
[03:47] <dendrobates> vipaca:  no, it;s pretty safe.
[03:47] <vipaca> dendrobates:Im not sure how to read these things
[03:48] <vipaca> [   32.575597] tg3: eth%d: Cannot get nvarm lock, tg3_nvram_init failed.
[03:49] <dendrobates> vipaca: yeah, I haven't seen that before.
[03:50] <vipaca> hey you pointed me to dmesg
[03:50] <vipaca> very good
[03:50] <vipaca> now I at least no that the driver is not happy
[03:50] <vipaca> or kernel module or what ever tg3 is
[03:51] <vipaca> thanks
[03:51] <dendrobates> the driver seems to be finding your interfaces
[03:51] <dendrobates> I think that message is not relevant.
[03:52] <vipaca> maybe and maybe not
[03:52] <dendrobates> what is the actual problem you are seeing?
[03:52] <vipaca> well something like this I can only get light when connacting cable to 3 outof 4 nics at a time
[03:53] <vipaca> and depending it changes whiich one
[03:53] <vipaca> once Iget a cable connected andlights on that port no connectivity after static ip conf
[03:54] <dendrobates> I see the kernel module finding all four cards.
[03:55] <vipaca> sure and ifconfig shows them all but no lights
[03:55] <vipaca> no connectivity
[03:57] <dendrobates> You should dig deeper into the logs.  See if you can find any errors in /var/log/syslog
[03:58] <dendrobates> does ifconfig look ok?
[03:58] <vipaca> it does to me
[03:59] <dendrobates> have any of the ports worked at all?
[04:00] <vipaca> not on ubuntu
[04:00] <vipaca> and I couldn
[04:01] <dendrobates> what does your /etc/network/interfaces look like?
[04:02] <dendrobates> what are you connecting these ports to?
[04:31] <vipaca> dendrobates:sorry daddy duty called
[04:31] <vipaca> a switch
[04:32] <vipaca> interfaces is empty now but when its configure no good
[04:32] <vipaca> Ive bee using just temp ip setupwith ifconfig
[05:19] <kgoetz> anyone here tried exim4-daemon-light? i have it half installed in a chroot, and its breaking my packaging system :(
[05:19] <kgoetz> wondering if its a bug in the package or because its chrooted
[05:49] <kraut> moin
[05:49] <kgoetz> hey mate
[05:49] <kgoetz> fwiw, i think its a chroot issue
[09:20] <roote> hi!
[09:21] <roote> anyone knowledgeable about linux software RAID? Shouldn't a RAID1 be faster than a single disk?
[09:21] <roote> i'm getting the same speeds testing a single harddisk VS RAID1
[09:22] <roote> i thought it was supposed to give me almost double the speed
[09:36] <soren> For reads, it should be significantly faster than a single disk. For writes, it's the same or slower.
[09:36] <soren> It depends a bit on the hardware involved. If they're on the same IDE channel, you won't get much improvement (if any).
[09:48] <_ruben> on same channel i'd expect it to be much slower even
[09:54] <soren> _ruben: Depending on how I/O bound it is compared to seek bound.. Dunno. I'm just guessing.
[09:57] <_ruben> same here
[10:22] <c1|freaky> is there an RSS news feed somewhere for the server people?
[10:34] <soren> What sort of news are you looking for?
[10:38] <c1|freaky> news about ubuntu releases, security updates and other server-related stuff
[10:52] <soren> c1|freaky: There are mailing lists for all of that. There's security-announce, ubuntu-server, and ubuntu-announce.
[10:55] <c1|freaky> where can i subscribe to ubuntu-srever?
[10:55] <c1|freaky> *server
[11:00] <_ruben> c1|freaky: https://lists.ubuntu.com/
[11:01] <stiv_> hey everybody ... i have multiple ubuntu-servers w. feisty that i update using my self-signed mirror w. some own packages (works fine!). These machines are not in a LAN, so updating via https would be a good idea, but i fell over this bug: https://bugs.launchpad.net/ubuntu/+source/apt/+bug/109294
[11:01] <ubotu> Launchpad bug 109294 in apt "HTTPS sources fail to update" [Medium,Fix released]
[11:02] <stiv_> it's only fixed in gutsy and i can't find a way to adapt the apt source package (i could build it myself) to fix this issue ... is there a patch?
[11:05] <stiv_> "These machines are not in a LAN," --> yes they are :-) but not all in the same the update server is in. "are not in a LAN" sounds weired
[11:09] <c1|freaky> thank you :)
[12:28] <roote> soren and _ruben : the disks of the RAID1 are on different SATA channels. I tested speed with hdparm -tT /dev/sda, hdparm -tT /dev/sdb and hdparm -tT /dev/md0 (and 1 and 2, since i have 3 partitions on each disk, and corresponding 3 raid arrays).
[12:29] <roote> soren and _ruben : they all give me roughly the same speeds (around 60MB/s)
[13:00] <_ruben> roote: i never put much faith in hdparm as a benchmark
[13:00] <_ruben> if only i could recall the name of that benchmark tool ... its a girl's name ..
[13:00] <_ruben> bonnie++
[13:01] <_ruben> http://www.acnc.com/benchmarks.html for several other options
[13:06] <Gargoyle> is there a meta package to install a LAMP stack as if you had chosen it during installation?
[13:11] <_ruben> Gargoyle: i think that'd be sudo apt-get install lamp-server^
[13:11] <Gargoyle> not showing up under apt-cache search.
[13:12] <_ruben> correct (i think), tried that recently as well .. the install part does work tho ;)
[13:13] <Gargoyle> with the ^ ?
[13:13] <_ruben> yes
[13:13] <_ruben> the ^ indicates a task
[13:13] <_ruben> or you could use tasksel
[13:14] <Gargoyle> ahh well seems I got everything with apache2 mysql-server php5 and php5-mysql
[13:16] <_ruben> grr .. cant connect to mit's pgp server .. lets see if i can find an alternative
[14:30] <sommer> jdstrand: hey, I've updated the firewall documentation section... have time for a review?
[14:30] <jdstrand> sure! email?
[14:30] <sommer> on the way, thanks
[14:33] <sommer> also I went ahead and committed the changes so if you have the repo checked out you can see it that way :)
[14:36] <jdstrand> sommer: thanks reading
[14:37] <jdstrand> sommer: heh-- comma needed after 'configuration' in 2nd sentence of 3.2
[14:38] <jdstrand> actually, I'll send you a diff
[14:38] <sommer> jdstrand: okay
[14:38] <sommer> just for the one comma or.... ?
[14:45] <jdstrand> sommer: no, couple of other teeny things
[14:46] <sommer> jdstrand: ah, I appreciate the feed back
[15:23] <fuzed> hello?
[15:24] <fuzed> I'm wondering about X setup on server, is there a specific set of packages needed, does one need to run xorgconfig(which I can't seem to find) or what
[15:25] <jdstrand> sommer: sent diff
[15:26] <jdstrand> sommer: thanks again!
[15:26] <sommer> jdstrand: awesome, thank you
[15:26] <sommer> fuzed: it's not really recommended, but you can find a guide here: https://help.ubuntu.com/community/ServerGUI
[15:26] <soren> fuzed: Ask in #ubuntu
[15:27] <fuzed> thnks
[15:36] <sommer> jdstrand: so you only need one commit in the rules file?
[15:36] <jdstrand> sommer: just at the end
[15:36] <jdstrand> sommer: let me double check that
[15:37] <sommer> gotcha, is there a site, man page, or something that documents iptables-restore files?
[15:37] <sommer> I got most of that info from forum and mailing list examples
[15:37] <_ruben> info on iptables-restore/save is *very* limited
[15:38] <sommer> I wanted to add a link in the resources for further rules file details, but couldn't really find anything official
[15:38] <_ruben> even netfilter/iptables developers dont know the 'rules' ;)
[15:38] <jdstrand> sommer: for the most part, it is the same as an iptables rule, with 'iptables'
[15:38] <jdstrand> the parts that are different are the default policys and COMMIT
[15:39] <sommer> I see, I guess I was wondering about these lines ":PREROUTING ACCEPT [0:0]", what does the [0:0] mean?
[15:39] <_ruben> 0 bytes / 0 packets
[15:39] <jdstrand> resets the counters
[15:39] <sommer> ah, gotcha
[15:40] <sommer> I guess once you understand that part of it, it's really just iptables commands... like you said
[15:40] <sommer> _ruben: thanks
[15:41] <_ruben> correct .. just slightly different formatted ;)
[15:41] <sommer> cool, I think we're good for the resources section then
[15:42] <sommer> jdstrand: also, to double check even if you set /etc/default/ufw to forward you still need to enable forwarding in /etc/sysctl.conf?
[15:43] <sommer> I think I missed a command to reload sysctl once you edit the file :-)
[15:43] <jdstrand> sommer: yes. I should probably change that
[15:43] <jdstrand> sommer: and you are right about COMMIT
[15:43] <sommer> so it is needed for both tables?
[15:43] <jdstrand> one COMMIT per table
[15:44] <jdstrand> so the *filter table needs a COMMIT and the *nat table needs a COMMIT
[15:44] <sommer> gotcha, I'll adjust that and add the sysctl command
[15:45] <jdstrand> so all POSTROUTING part should be entirely *before* the *filter section
[15:47] <sommer> cool, does the new file cover the major aspects of ufw?
[15:47] <jdstrand> eg http://paste.ubuntu-nl.org/55851/
[15:47] <jdstrand> sommer: yes
[15:48] <jdstrand> sommer: obviously that paste includes some of my rules that shouldn't be in there ;)
[15:49] <jdstrand> sommer: particularly leave out the OUTPUT and PREROUTING lines
[15:50] <jdstrand> sommer: I am going to file a bug on the sysctl part
[15:51] <sommer> jdstrand: with regard to the documentation?
[15:51] <sommer> the bug I mean
[15:51] <jdstrand> sommer: you shouldn't have to adjust sysctl.conf in ufw when you adjust /etc/default/ufw
[15:51] <jdstrand> no, in ufw
[15:51] <jdstrand> so leave the sysctl.conf part out of the ufw section
[15:52] <sommer> ah, will do
[15:55] <jdstrand> sommer: fyi bug #191596
[15:55] <ubotu> Launchpad bug 191596 in ufw "ufw: should not have to enable ip_forward in sysctl.conf" [Undecided,Confirmed] https://launchpad.net/bugs/191596
[15:57] <sommer> jdstrand: here's my new before.rules: http://paste.ubuntu-nl.org/55852/
[15:58] <jdstrand> sommer: looks good
[15:59] <sommer> cool, I'll move the sysctl step to the iptables masquerade section and apply the changes
[15:59] <jdstrand> sommer: can you send me your final changes-- I am going to walk through this section on my system
[15:59] <sommer> sure will do
[15:59] <sommer> before applying ;-)
[15:59] <jdstrand> sommer: won't happen til later in the day though
[15:59] <jdstrand> feel free to apply-- I think they are right-- I am just going to double check it all
[16:00] <sommer> jdstrand: no problem, minor changes won't be a big deal until SF, I just wanted to get the major section update in before FF
[16:00] <jdstrand> yes-- would only be minor
[17:23] <mathiaz> zul: I'm currently merging mysql-dfsg-5.0 from debian. I think that the changes you made for man conflict need to be extended to mysql-client-5.0 and libmysql-dev-5.0.
[17:23] <mathiaz> zul: these two packages also shipp conflicting man page.
[17:36] <vipaca> Where do I open a bug against sparc server
[17:37] <mathiaz> vipaca: is a kernel problem ? a problem with a specific package ?
[17:46] <vipaca> tg3
[17:46] <vipaca> perhaps
[17:48] <vipaca> What im seeing is that eth2 is tied to port 0 of 4
[17:48] <mathiaz> vipaca: so if it's a kernel bug, file a bug against the kernel package.
[17:56] <zul> mathiaz: ok im just having lunch
[18:03] <mathiaz> zul: np. I'm doing the merge.
[18:59] <fuzed> ok, does it say anywhere that you have to run mandb initially (-c) and rerun occaisonally,  why don't the package scripts or the pkg manager do that automatically?
[19:35] <peterdv> fuzed: You should have the files "/etc/cron.daily/man-db" and "/etc/cron.weekly/man-db"  doing just that
[19:38] <vipaca> can you give me a link for fileing a bug against the kernel
[19:42] <mathiaz> vipaca: https://bugs.launchpad.net/ubuntu/+source/linux/+filebug
[20:12] <frame45> does anyone listen to Linux Reality Podcast
[20:13] <foo> Anyone located in southern california?
[20:15] <frame45> FTP Server? should I use Ubuntu Server edition (command line) or the regular distro (gui) and then run like VFTP? (is VFTP right)
[20:16]  * foo likes proftpd
[20:17]  * ScottK dislikes FTP period due to plain text password transmission.
[20:17] <foo> actually, good point. sftp++
[20:18] <frame45> I work at a Video Production co. we need to setup an SFTP server so that our client can download about 60Mb-100Mb video files from it.
[20:19] <foo> frame45: Hm, is the client only downloading?
[20:19] <frame45> yes
[20:20] <foo> frame45: Do you plan on giving them ftp/sftp access?
[20:20] <foo> Are the files already on the server? Or does someone need to upload to the server so people can download?
[20:21] <ScottK> For an Ubuntu Server if you install openssh-server, sftp will just work.
[20:21] <foo> ScottK: Right, but that may not be the best solution here
[20:21] <frame45> yes.   Most of our client use Mac I want them to be able to log-in via Connect to server or with an FTP client software
[20:21] <zul> vsftp
[20:22] <foo> frame45: That's too complicated for a client. Why not setup HTTP authentication with directory listing enabled so they can download files via HTTP?
[20:22] <frame45> I will be creating and administering (uploading to the FTP)
[20:22] <foo> frame45: Ok, so someone will need to upload files, then client downloads. You can upload via sftp, and client could download via HTTP. I think that's your best bet.
[20:22] <frame45> ok, that would be great I just don't know enough about all this yet i guess.
[20:25] <CarlFK> how do I setup /etc/network/interfaces to use a wifi card?
[20:25] <CarlFK> pretty sure I did it that way before...
[20:26] <infinity> CarlFK: man 5 interfaces
[20:26] <infinity> CarlFK: It's well documented.
[20:26] <CarlFK> ah, I stopped at man interfaces
[20:26] <CarlFK> thanks
[20:43] <moos3> anyone awake in here
[20:44] <ScottK> Yes
[20:44] <moos3> I need some server help
[20:44] <foo> moos3: ask away
[20:44] <moos3> http://ubuntuforums.org/showthread.php?t=695341
[20:44] <moos3> is a thread that I started last night when I upgraded from 7.04 to 7.10
[20:45] <moos3> and lost my server because of some unknown reason
[20:46] <moos3> it worked fine in 7.04 but for some reason when I upgraded to 7.10 it hangs and I can't figure out why
[20:46] <foo> moos3: is all of that the last 8 lines before you get into a shell?
[20:47] <moos3> yeah
[20:47] <moos3> its when its trying to load the kernel
[20:47] <moos3> what gets me kernel 2.6.20-16 boots until it gets to kernel log daemon where it hangs
[20:48] <moos3> and kernel 2.6.22-14 fails when trying to load the kernel
[20:48] <foo> moos3: Ask in #ubuntuforums too
[20:51] <foo> moos3: Wish I could help, but not sure, sure. Let me know when you find out
[20:51] <moos3> ok
[20:55] <mathiaz> Ubuntu Server meeting in #ubuntu-meeting in 5 minutes
[20:55] <zul> nothing has been added to the agenda recently has it?
[20:55] <ScottK> It's a good warning though, so I can do my action from that last meeting.
[20:57] <mathiaz> zul: well - it depends how you define recently
[20:57] <zul> heh i just saw it on the wiki
[20:58] <zul> wiki esp
[21:05] <XiXaQ> didn't someone tell me that vmware tools is now in hardy repositories?
[21:07] <pwnguin> is there anything close to a remote installer for ubuntu/debian?
[21:07] <XiXaQ> remote installer?
[21:08] <pwnguin> specifically: something to control the install via a network; i can place a cd in the drive and boot it, but we'd rather not drag a mouse monitor etc for the duration
[21:08] <zul> like kickstart?
[21:08] <pwnguin> i think kickstart is a boot from network deal
[21:09] <pwnguin> but this is all new to me, so perhaps
[21:09] <faulkes-_> look at PXE booting for install
[21:09] <faulkes-_> not sure what exists for an automatic though
[21:10] <pwnguin> what about say a serial console?
[21:12] <XiXaQ> that's interesting. Indeed, it would be nice if you could connect to the installer using ssh or vnc.
[21:12] <soren> you can.
[21:13] <pwnguin> do you need to install sshd first?
[21:13] <soren> No.
[21:13] <soren> Hang on.
[21:15] <soren> http://wiki.debian.org/DebianInstaller/NetworkConsole
[21:17] <pwnguin> i gotta say, that's a terrible wiki page
[21:19] <mathiaz> XiXaQ: you're probably looking for open-vm-tools.
[21:19] <pwnguin> i mean the first paragraph basically says the topic's pointless, without ever explaining what it is :(
[21:20] <pwnguin> i dont need to preseed d-i, but d-i over ssh would be handy
[21:20] <XiXaQ> mathiaz, ah, thanks. :)
[21:23] <pwnguin> it kinda sounds like something already exists
[21:24] <moos3> foo: ok I did what you said now I'm getting the follow
[21:27] <soren> XiXaQ: A new and fresh version will hit the archive within 24 hours. (of open-vm-tools)
[21:27] <XiXaQ> great :)
[21:27] <moos3> foo: http://pastebin.org/19460
[21:28] <XiXaQ> hmm. It doesn't seem to have any configuration options, like vmware-tools does. Like screen resolution, etc.
[21:30] <soren> open-vm-tools-gui
[22:03] <Goosemoose> hi guys
[22:03] <Goosemoose> cjwatson_, you around?
[22:03] <pwnguin> soren: were you suggesting a preseeded install?
[22:04] <Goosemoose> anyone else know the status of the utility to connect to active directory for authentication?
[22:04] <pwnguin> launchpad says needs code review
[22:05] <pwnguin> https://blueprints.edge.launchpad.net/ubuntu/+spec/windows-authentication-integration
[22:05] <cjwatson_> Goosemoose: I'm around, but know nothing about Active Directory
[22:05] <Goosemoose> hey cjwatson_
[22:05] <Goosemoose> hmm, thought it was you that was working on the tool in hardy for authentication
[22:05] <Goosemoose> darn, who was it!
[22:06] <soren> Goosemoose: No, it's dendrobates.
[22:06] <Goosemoose> ahh ok
[22:06] <soren> pwnguin: That's an option. You can also just specify it on the kernel command line.
[22:06] <cjwatson_> pwnguin: section 6.3.8.3 of the installation guide (package installation-guide-i386) is probably more useful.
[22:06] <cjwatson_> and the rest of the installation guide in general
[22:06] <cjwatson> the option soren refers to is anna/choose_modules=network-console
[22:07] <dendrobates> Goosemoose: https://launchpad.net/~dendrobates/+archive
[22:07] <pwnguin> of course, under 6.3.8. Miscellaneous
[22:07] <cjwatson> pwnguin: Kickstart doesn't intrinsically involve booting from the network
[22:08] <dendrobates> Goosemoose: a new version has just been uploaded.  You will need to wait for it to build.
[22:08] <cjwatson> it's usually invoked that way just because its main use is installing a zillion systems at once and obviously you don't want to go around inserting a CD into each one
[22:08] <cjwatson> but Kickstart is just another automation method, not an interactive remote control method
[22:10] <pwnguin> ok, so does the server CD allow me to access the menu via console, or do i still need to haul in a monitor / keyboard?
[22:11] <pwnguin> we've got a laptop in there with a serial port for serial console hacking ;)
[22:11] <cjwatson> server CD >= feisty should let you do it
[22:12] <pwnguin> ok; thanks for the help and pointers!
[22:21] <ivoks> so, when we eliminate dbconfig, how are we supposed to create both database and user during installation?
[22:24] <Goosemoose> dendrobates, will it work with the current edubuntu or do i have to wait for hardy?
[22:25] <dendrobates> Goosemoose: I am just packaging it for hardy, but I believe Likewise is packaging it for gutsy, check http://www.likewisesoftware.com/products/likewise_open/
[22:26] <Goosemoose> ok, any idea if I can get it to work with a preseed cfg file?
[22:37] <jronnblom> soren: you still around?
[22:39] <Gargoyle> Can anyone help me out with apache warning me that namevirtualhost *:0 has no virtual hosts?
[22:39] <Gargoyle> I have to files in sites-available and they both look OK to me.
[22:40] <antdedyet> Gargoyle: you probably don't want apache listening on port 0.
[22:40] <Gargoyle> I havent specified that anywhere
[22:40] <soren> jronnblom: Always.
[22:40] <antdedyet> Gargoyle: You might rather have it on a standard or non-standart port, such as 80 or 8080, respectively
[22:41] <Gargoyle> I have namevirtualhost * should I change that to namevirtualhost *:80 ?
[22:41] <jronnblom> soren: In the /etc/init.d/open-vm-tools I think there might be a problem with the loading of the vmxnet driver
[22:41] <Gargoyle> Or should it get picked up from the main config?
[22:41] <jronnblom> 		if vmxnet_needed
[22:41] <jronnblom> 		then
[22:41] <jronnblom> 			rmmod pcnet32
[22:42] <jronnblom> even if vmxnet_needed returns 1 its never evaluated by the if
[22:42] <antdedyet> Gargoyle: You will only need "NameVirtualHost *" once in your httpd.conf
[22:43] <Gargoyle> ahh
[22:43] <Gargoyle> ok
[22:43] <soren> jronnblom: Why?
[22:43] <jronnblom> soren:                 vmxnet_needed
[22:43] <jronnblom> 		if [ $? -eq 1 ]
[22:43] <jronnblom> 		then
[22:43] <jronnblom> 			rmmod pcnet32
[22:43] <soren> That's the exact same thing.
[22:43] <Gargoyle> makes the sites-available a bit more awkward, but not a major problem.. thanks antdedyet
[22:43] <soren> Except opposite, but ykwim
[22:45] <soren> jronnblom: Ok, are you actually experiencing a problem, or are you just looking at the code and think there's something wrong?
[22:45] <jronnblom> im running hardy alpha4 on vmware ESX I by default the pcnet32 driver is loaded
[22:46] <antdedyet> Gargoyle: np -- you can always create a namevhost-local.conf in conf.d and write "NameVirtualHost *" in there
[22:46] <jronnblom> it also loads the vmxnet driver
[22:46] <antdedyet> That will allow you less conflicts with distribution default config updates
[22:46] <jronnblom> soren: at first I thougth the vmxnet_needed purpose was to remove the pcnet32 and replace it with vmxnet
[22:47] <jronnblom> soren: now Im not sure anymore
[22:48] <antdedyet> Of course, you could also just set proper dpkg up to not ever overwrite conflicting conffiles
[22:54] <jronnblom> soren: should both the pcnet32 and the vmxnet be loaded in ESX ?
[23:04] <Centaur5> Is there a way to make a dhcp server only give out IP addresses to PXE boot requests?
[23:04]  * antdedyet shrugs as he sees someone's gutsy box hard lock after trying the likewise-open built package from their site after failing to join the domain
[23:06] <antdedyet> The hardy pkg worked fine for me, except joining the domain failed.
[23:06] <antdedyet> (my error tho, I don't know the network well)
[23:08] <antdedyet> Centaur5: My method would involve manually toggling the MACs in and out of the PXE boot config section located in dhcp... probably not what you're looking for.
[23:08] <antdedyet> Centaur5: what dhcp server are you using? isc?
[23:15] <Centaur5> antdedyet: Right, I do that to have one server do thin clients but the other server I want to network install (hopefully when I can figure that out).
[23:16] <Centaur5> When I have both servers running though they both give out IP addresses and I only want one of them to give addresses to regular workstations.
[23:17] <faulkes-> Centaur5: you can tell dhcp to ignore or handle specific mac addresses
[23:18] <Centaur5> faulkes-: great suggestion but I have guest machines coming and going that are constantly connecting.
[23:22] <Centaur5> I'm probably going to unfortunately have to physically separate with a new hub and cables the machines that need installations or just deal with 2 servers giving out IP addresses?
[23:22] <antdedyet> Centaur5: that's almost exactly what I was about to suggest ... network segregration is a good thing when multiple dhcp servers are involved
[23:23] <antdedyet> Centaur5: but the new servers to be installed are going to eventually join the existing infrastructure?
[23:26] <faulkes-> iirc, I believe some switches support forwarding tftpboot / pxe / dhcp requests to specific hosts on a per port basis
[23:26] <Centaur5> We sell and repair computers and I'm trying to figure out the best way to load Ubuntu on new machines and network installs would be excellent.  New machines are built and installed on any available bench so we would have to dedicate an area for new installs which would be mainly wasted.
[23:26] <faulkes-> however, those would typically be high end switches/routers ;)
[23:28] <faulkes-> why don't you have just one dhcp server do both?
[23:28] <antdedyet> Centaur5: why would the space be wasted? it's saving you a ton of time; making your work more effective.
[23:28] <faulkes-> iirc it is entirely possible and shouldn't affect regular client machines
[23:29] <faulkes-> pxe calls dhcp to get an address and then connects to tftpboot to get config files and other data
[23:29] <faulkes-> so a regular machine booting, unless pxe boot is enabled in the bios, will only ever get a dhcp address
[23:29] <Centaur5> antdedyet: Well we would have to find a decent amount of necessary bench space. We have 9 machines to install now but next week we may have 0.
[23:31] <antdedyet> Centaur5: run two patch cables to each dedicate machine space on the bench and label them appropriately as "new network install access" and "regular ole internet" or whatever...
[23:31] <faulkes-> again, I have to question the requirement of two dhcp servers
[23:31] <Centaur5> antdedyet: Yeah, I thought about that as a possibility.
[23:32] <faulkes-> one dhcp server is quite capable of handling both pxe booted machines and regular dhcp requests
[23:32] <Centaur5> faulkes-: Only mainly cause I try to keep all hard file transferring on a different server than the one with our applications.
[23:33] <faulkes-> as long as you have a good switch, it shouldn't matter
[23:33] <faulkes-> if you are using a hub, different story
[23:33] <faulkes-> that's just MO though
[23:33] <Centaur5> faulkes-: It's mainly to prevent bogging down the 2 thin clients we use off that primary server as well.  We have a gigabit 24 port switch.
[23:34] <Centaur5> It's not managed though, we should have done managed.
[23:34] <faulkes-> if you have gigabit ethernet available, bandwidth shouldn't be an issue
[23:34] <faulkes-> managed or not really doesn't matter
[23:34] <faulkes-> switching is handled at the hardware layer
[23:35] <faulkes-> once the switch learns the mac destination, it won't flood other ports
[23:35] <faulkes-> so unless your thin clients are attempting to access the pxe booted machines during install
[23:35] <Centaur5> faulkes-: Okay, we currently have a raid 10 on the primary server that runs the thin clients and all web apps.  When I did a 8 gig data backup it made the thin clients unusable.  Do you think a network install will cripple them?
[23:36] <faulkes-> it shouldn't
[23:36] <faulkes-> it's raid, you are basicly doing one huge read
[23:36] <faulkes-> how was the backup done?
[23:36] <Centaur5> Oh right, I was doing the backup to the server so I was writing.
[23:37] <faulkes-> again, I'm not saying you won't have issues, but I think it's at least worth an investigation before you start rewiring stuff and creating a workbench area
[23:37] <Centaur5> It's also a hardware raid controller so I don't imagine we can get much faster without getting very expensive.
[23:37] <soren> Gah..
[23:38] <faulkes-> if the backup software runs on the server itself, you also have to deal with the fact the it likely stole a good portion of the cpu in the server
[23:38] <faulkes-> not to mention disk i/o
[23:38] <faulkes-> whereas reading from a raid controller to a gige network should all happen at the hardware level and the ip stack
[23:39] <Centaur5> faulkes-: It was actually just a 8 gig transfer from a different machine.  Maybe I'll test the network install on that server then and hopefully solve that issue.
[23:39] <faulkes-> well, that is my suggestion, if it works, you would save alot of time and energy
[23:39] <Centaur5> Now my question is has anybody successfully done a network install server on Gutsy?
[23:40] <faulkes-> iirc Loye has
[23:40] <Centaur5> I did one on Edgy 8 months ago but Gutsy appears to be different.
[23:41] <faulkes-> http://ubuntu-tutorials.com/2007/10/11/how-to-configure-pxe-network-booting-on-ubuntu-for-network-based-installations/
[23:41] <faulkes-> which specically references 7.10
[23:42] <Centaur5> Yeah, I tried that one.
[23:43] <Centaur5> The server gets this error in syslog when a workstation attempts to boot. in.tftpd[14220]: received address was not AF_INET, please check your inetd config
[23:48] <faulkes-> suggested work around includes using xinetd instead of inetd
[23:48] <faulkes-> https://bugs.launchpad.net/ubuntu/+source/inetutils/+bug/32330
[23:48] <ubotu> Launchpad bug 32330 in inetutils "Amanda fails because client thinks server is IPv6 when it's IPv4." [High,Confirmed]
[23:49] <faulkes-> although that references amanda, it is the same issue you are seeing
[23:49] <Centaur5> ahh...new package