[00:06] i get an error when trying: sudo apt-get install subversion libapache2-svn [00:06] Setting up libapache2-svn (1.6.5dfsg-1ubuntu1) ... [00:06] Considering dependency dav for dav_svn: [00:06] ERROR: Module dav does not exist! [00:06] ERROR: Could not enable dependency dav for dav_svn, aborting [00:14] :/ [00:53] django app deployment question [00:53] can you use the alternate cd to install a no-gui install, and if so, do you need to do anything special... i cant find a server edition alternate cd (need alt for raid options) [00:53] how should i deal with multiple apps assuming CSS hosted in /media/? [00:54] ie, the django admin app uses /media/base.css [00:55] i can alias /usr/share/pyshared/django/contrib/admin/media/css in apache [00:55] but if i come across another app that uses /media/, that fails [00:58] is there some place i can go to read about the recent changes to samba? I have one shared folder that contains symlinks to another user's folders. The symlinks and the folders all have 777 permissions, but when I try to open those folders over the share, it says "You don't have permission to access this resource". What gives? [00:58] This used to work just fine [01:00] Italian_Plumber: look at http://www.ubuntu.com/usn/USN-918-1 [01:00] Italian_Plumber: you can no longer have wide links and unix extensions enabled at the same time in samba [01:01] looking into it, i can define a seperate prefix for the admin app [01:04] so if I want to share those other users folders I have to create a new share? [01:04] Italian_Plumber: if you are serving windows clients, you can simply disable unix extensions [01:05] Italian_Plumber: if you are serving unix clients, yes, you need to create a new share [01:05] ok I'll look into that, thanks. [01:06] Well that explains what to do but doesn't explain what wide links and unix extensions are. :) [01:54] Hmm [01:55] UEC supports both kvm (by default), and Xen. I'm considering actually trying it now for once, after playing with XenServer from Citrix. heh [01:56] Can you run both kvm and xen domU's on the same UEC host? === samferry is now known as sam_ === sam_ is now known as samferry [02:51] New bug: #551356 in samba (main) "samba symbol link folder access denied" [Undecided,New] https://launchpad.net/bugs/551356 === erichammond1 is now known as erichammond [03:11] I have an LVM at /dev/mapper/vg-root When I'm in the installer do I want to select the volume and create a new partition table? I thought LVM got rid of partition tables??? [03:12] Also, I'm installing from live-cd so when I boot up is it going to be able to assemble with mdadm and make the lvm2? [03:13] It's not meaningful to create a partition table within an LVM LV. [03:13] Well I created a /dev/vg/swap and /dev/vg/root [03:13] Those are LVs, not partitions. [03:13] Within the installer it shows up as /dev/mapper/vg-root [03:14] Exactly, how do I install to the LV. [03:14] In d-i, you simply assign mount points to LVs and/or partitions. [03:15] I can't comment on ubiquity, because when last I looked it didn't support md RAID nor LVM. [03:15] Nor did the live CD support server installs. [03:15] What's ubiquity - the installer [03:15] * Maletor answers own question about installer [03:16] !info ubiquity [03:16] ubiquity (source: ubiquity): Ubuntu live CD installer. In component main, is optional. Version 2.0.10 (karmic), package size 3509 kB, installed size 12428 kB [03:16] well i just want a vg for swap and a vg for root. How can I install Ubuntu now? [03:17] err. lv [03:19] Maletor: do you want a server or a desktop? [03:19] desktop twb [03:19] (and personal server) [03:19] Maletor: I won't provide support for Ubuntu desktops. [03:20] Server then. [03:24] OK, then you need to download the server install CD. [03:24] Otherwise you'll get a GUI, and various other things that shouldn't be on a server. [03:28] Ok I'm downloading alternate-amd64 [03:28] I will boot up, find my RAID, find my LVM2 and install appropriately. [03:28] Kernel should have mdadm and LVM2 [03:28] The alternate CD is a Desktop install CD. [03:29] Different kernel's though, right? [03:29] I don't know, since I don't use Ubuntu desktops. [03:29] I'm pretty sure they are. [03:29] Ubiquity can't support RAID and LVM so why should the kernel? [03:30] I need alternate or this will not be able to mount the root filesystem because it won't be able to find (access) it. [03:30] Ubiquities failings aren't any reason to cripple the kernel. [03:30] *ubiquity's [03:30] wait, what [03:31] ubiquity can do raid and lvm setup [03:31] but as twb says, for desktop stuff - #ubuntu-desktop [03:31] or even #ubuntu [03:31] It can!? :) [03:31] if you're doing a server install, grab the server CD [03:33] Hey, I have a /dev/md1 as RAID5 and a /dev/vg/root (as ext4 -- /dev/mapper/vg-root). [03:33] [10:32pm] Maletor: Can I just select my /dev/md0 and set that to boot [03:33] [10:32pm] lifeless: stay on channel please [03:33] [10:32pm] Maletor: Select dev/vg/root to my root partition and /dev/vg/swap to swap [03:33] [10:33pm] Maletor: And install. And the kernel will be able to figure out that it needs to assemble a raid and find a lvm to startup? [03:44] is there a command line to check the Upload and the Download traffic speed my eth0 ? [03:44] of my eth0 [03:47] maxagaz: push data through it [03:48] Time how long it takes to transmit a known quantity. [03:48] If you want the theoretical maximum, mii-tool or ethtool will report the negotiated line speed for 802.3 links. === erichammond1 is now known as erichammond [04:30] twb, there's no command line tool to check the amount of data transfered per second ? [04:36] maxagaz: ntop? [04:36] soren, iftop... [04:53] hi guys... i tried configuring mrtg on my ubuntu server... getting the following warnings when i run mrtg.... [04:54] http://pastebin.com/5i8bLQH2 [04:55] I dont understand why this is happeining [04:55] the graphs are shown on the browser [04:55] but they are blank... [04:55] should i run snmpwalk on my server? [04:55] snmp agent is started on 192.168.8.1 router [04:56] swift: sounds like a chown issue [04:56] swift: who owns those log files, and who does rateup run as? [04:57] New bug: #358098 in vm-builder (universe) "vmbuilder hangs or crashes when building images on ec2 instances" [Wishlist,Triaged] https://launchpad.net/bugs/358098 [04:57] how to check that?... ls -al rateup? [04:57] New bug: #371783 in vm-builder (universe) "NameError: global name 'isodir' is not defined" [Low,Fix released] https://launchpad.net/bugs/371783 [04:57] swift: also, LANG=C is wrong. You want LC_ALL=C. [04:57] New bug: #392097 in vm-builder (universe) "vmbuilder ignores mac option when using bridge network" [Low,Fix released] https://launchpad.net/bugs/392097 [04:57] swift: no, examine the process while it's running. [04:59] twb... how do i do that? [05:00] !grounding [05:00] Hmph. [05:02] ubottu should just snarf all the dpkg bot's info entries [05:02] Error: I am only a bot, please don't think I'm intelligent :) [05:02] ubottu: shut up [05:02] swift: you need to improve your basic Unix skills [05:25] twb: moo [05:26] Ha! [05:26] hows life? [05:26] Shrug. [05:26] found a workaround for the ldap problem i mentioned [05:26] nothing like some universe packages :) [05:26] Sam-I-Am: go on. [05:26] * twb steeples fingers [05:26] you still have the bug i posted? [05:27] Can't remember. I subscribed to the Debian one. [05:27] one sec.. [05:27] https://bugs.launchpad.net/bugs/423252 [05:27] Launchpad bug 423252 in sudo "NSS using LDAP on Karmic breaks 'su' and 'sudo'" [Unknown,Confirmed] [05:27] posted solution [05:28] while i was there, i also noticed another interesting bug... introduction of using sasl external to modify ldap by default caused some acls to be added in Bad Places [05:28] like... acling things like cn=schema from outside connections [05:28] so uh... nothing can read the schema [05:29] openldap developer nicely posted that one for me... his name has a bit more pull than mine does. we'll see if anyone notices. [05:29] theres a lot of sorta duplicate bugs i'm merging for that [05:30] not being able to read rootdse kinda returns a variety of strange error messages [05:41] twb: the silence is deafening :/ [05:43] Sam-I-Am: sorry, other channels and whatnot [05:44] So nslcd magically fixes it? [05:44] Yeah, OK. [05:45] I've done my share of flip-flopping between nss-ldap and nss-ldapd [05:45] Their actual code is bloody ugly. [05:48] yeah [05:48] its... not ideal [05:48] but it works :P [05:48] i think its nss-ldapd which fixes it... nslcd seems to be related, but i dont think its the core issue [05:49] in the past, nss-ldapd has sucked ass, so this was a surprise [05:50] time to hit the sack here [07:23] hi guys, I wanted to confirm on my query... mrtg is not workin as expected.. i get the following warnings [07:23] http://pastebin.com/5i8bLQH2 [07:24] i checked the permissions on 'rateup' file... the owner is 'root' [07:24] with rwx permissions [07:25] also, now when i manually type 'sudo mrtg /etc/mrtg/mrtg.cfg'.... i get no warnings.... andonly then my graph get's updated [07:25] otherwise... i see mrtg doesn't run as a process... please advise [07:30] any advise if mrtg can cause the internet line to go down? [07:30] swift_, you need to figure out what user mrtg is running as, and chown the log files to that user === swift_ is now known as swift [08:21] swift, run it again a few times. you always get some warnings the first runs [08:21] and no, mrtg isn't running as a daemon [08:21] use cron [08:21] RoyK, i want to have it in the init tab... so that it starts at reboot [08:21] how can i do that? [08:22] crontab -e [08:22] man cron [08:22] man 5 crontab [08:26] RoyK, dont want a cron job... want to have it under init [08:27] how cqan i have it as a service ... always running? [08:28] swift: That's not how mrtg is designed. [08:28] It is designed to run from cron (and installed so) [08:28] sherr, can't i have it as a service running at all times... that's kinda bad [08:28] It is not meant to run *all* the time. Reading/generating stats all the time. [08:29] i mean.. if smokeping can.. what;s with mrtg? [08:29] swift: why? It is silly - snapshot every second? [08:29] yeah... it just shows the graph.. progress [08:29] so that i can compare [08:30] sorry to intervene, but in that case your data is static every second [08:35] t3ngu.. how is it static? [08:36] that's wierd [08:36] swift, mrtg isn't a daemon [08:36] swift, use munin instead - mrtg is OLD [08:36] it can be run as a daemon [08:36] ok - didn't know that, but still, use something else [08:36] i wanted t try mrtg... munin is running [08:36] ah :) [08:37] :D [08:37] first munin, then rrdtool and then mrtg, just to trace back all the way? [08:37] I mean static, as you dont collect data every second [08:37] :D [08:37] mrtg has an option.. runasdaemon.. in the config [08:37] you usually run something every 5mins [08:37] and then plot it. [08:38] iirc 5 minutes is maximum resolution for mrtg [08:38] there's a lot of bad hardcoding in mrtg [08:39] ok, point noted... il run it using cron [08:39] thanks guys!!! [08:39] I really don't see why you shouldn't use cron :) [08:53] hi all. I'm considering setting up a cloud for internal virtualisation use, but I wonder, how can the have a redundant front-end server? [08:53] cluster? [08:53] load balanced webservers? [08:54] I thought the front-end also bridged the traffic..? [08:54] what do you mean 'breagded traffic' ? [08:54] eucalyptus needs to be aware of the redundancy [08:55] "tunneling" the ip traffic from the nodes to the clients [08:55] bridging [08:55] or routing if you like [08:56] then i think you would like linux-HA [08:56] http://www.linux-ha.org/wiki/Main_Page [08:56] not really, I would like Eucalyptus to do this. Linux HA is something else [08:57] or perhaps I can do it with DRBD/heartbeat [08:57] hm... [08:58] i really don't know eucalyptus [08:58] haven't used it but i have used drbd and load balanced in high traffic sites [09:00] RoyK: redundancy in the front-end is kind of un-researched right now. There's a bug I file on this, let me check [09:01] bug 497087 [09:01] Launchpad bug 497087 in eucalyptus "Redundancy for cloud and cluster controller (HA)" [Wishlist,Confirmed] https://launchpad.net/bugs/497087 [09:16] any clues on why this: 'sudo mount --bind /var/www/files/ /home/user/files/' didnt persist? [09:17] and q2, how to change the default user shell type, and which one should i choose? bash? [09:18] TeTeT, so do I understand it correctly that if the cluster controller dies, all VMs become unavailable? [09:18] brontosaurusrex, persist? [09:19] RoyK: after reboot this mounted folders show as empty [09:19] * RoyK didn't know --bind, though, having used -o bind, but it's probably the same [09:19] ah [09:19] :) [09:19] man fstab [09:19] RoyK: ok [09:19] RoyK: the cluster and the cloud controller are both SPOF if you go with the simple topology [09:19] brontosaurusrex, or just edit /etc/fstab [09:19] it's self-explainatory [09:19] RoyK: you can have multiple cluster controllers, but it's not straight forward to setup in 9.10 [09:20] RoyK: in 10.04 LTS 'Lucid' the multi-cluster approach has been integrated into the installer [09:21] hm, ok, so with 10.04 I can use a pair of "front-ends" and a bunch of nodes behind them? [09:21] RoyK: any other way? without editing fstab? [09:21] brontosaurusrex, that's the way [09:21] RoyK: nope, only one Cloud controller, but multiple cluster controllers for multiple availability zones [09:21] RoyK: eventually i would make a bash script of some sort : addsomenewuser tony [09:21] RoyK: are you aware of the Intel/Canonical cloud whitepaper? might give you some ideas for architecture [09:22] brontosaurusrex, you mean useradd -m? [09:22] RoyK: i mean i will do a script with multiple commands, where user will be input variable [09:22] TeTeT, do I understand it correctly that all traffic is bridghed through the cluster controller? if so, that's a single point of failure... [09:23] brontosaurusrex, sure, as you please, but mounting on startup is done in fstab. obviously you _can_ do it other places, like in the init scripts or even with cron, but it'll be ugly [09:24] RoyK: damn [09:24] RoyK: correct, if you only have one cluster - but a cloud might have multiple clusters [09:28] RoyK: ok, so view mtab after mount command, copy-paste to fstab, reboot and see if that wroks? [09:28] works* [09:30] brontosaurusrex, heh - just learn the system [09:32] brontosaurusrex, /tmp /var/tmp/tmp none bind 0 2 [09:32] will bind-mount /tmp on /var/tmp/tmp [09:32] on boot [09:40] ok [09:46] TeTeT, erm, but still, for one cluster, there would be only one controller, right? so no failover like what exists on high-end stuff like vmware? [09:50] RoyK: you're right, on a simple topology cloud with one front-end and multiple nodes, the front-end is a single point of failure [09:50] RoyK: so the front-end should be on HA hardware [10:03] TeTeT, there's another thing I don't understand - I read in the docs that the nodes cache their drive image locally. won't this make it hard for it to fail over to another node if the one running the VM dies? [10:06] RoyK: there is no fail over between nodes. Instances running on the nodes are thought to be disposable, e.g. an unreliable system [10:07] RoyK: so in general when you run an instance, the node will ask the S3 storage controller for an image [10:08] RoyK: the node checks then if the image is in the node's cache and if not, it will copy it from S3 [10:08] RoyK: the instance itself is then a copy of the image (more or less, some magic with ssh keys and creating one image out of the image, ramdisk and kernel is done too) [10:11] hi again... i have an ubuntu server which is the gateway to our network [10:12] it has two modems connected to it.. eachhavingan internet line [10:12] authentication is done via the ubuntu server... virtual ppp interfaces are present on the server.... now, one of the modems has an snmp agent which i use to capture data in mrtg [10:13] the other modem doesn't have any snmp agent... is there a way i can get modem data via the server?... maybe by using the ppp interface corresponding to that modem [10:13] please advise [10:14] just install snmpd [10:14] there will be snmp counters on linux for all interfaces [10:19] RoyK.. so, snmpd will be able to get the required data from the server... for that particular internet line? [10:20] snmpd will report the traffic on the interface that connect to that connection [10:20] whatever's happening inside the modem won't be shown, but that is probably just fine [10:21] yes... basically... I want to monitor the traffic on the line [10:21] well, the traffic going through the server and in/out of the line will be the same, no? [10:21] RoyK, i see a lot of snmp related services on the server... snmpdelta, snmpget, snmpdf.... [10:21] yes [10:21] or are there other machines connected to this modem? [10:22] no... all traffic goes out via the server [10:22] does the server listen to snmp? [10:22] only thing is, there are two internet lines [10:22] RoyK. how can i find that out? [10:22] connected to one or two interfaces on the server= [10:22] ? [10:23] if eth0 is connected to modem A and eth1 to modem B, they have different counters [10:23] there are two lines.... each connected to separate modems [10:23] if ppp0 is connected to modem A, it also has a counter [10:23] yes.. tat's the case [10:23] ok coool! [10:23] well, just cfgmaker [10:23] it'll find it [10:24] run cfgmaker? [10:24] man cfgmaker :) [10:24] RoyK... i ran cfgmaker for the modem IP which has the snmpagent [10:25] i see the graph for that modem too [10:25] infact... there are 8graphs... confusing [10:29] one for each interface cfgmaker found [10:30] but you should run cfgmaker for the ubuntu machine [10:30] since one of the modems don't have snmp [10:30] woh1!! [10:30] over the top [10:31] im new to this.. please can u make it easier? [10:32] hm.... I gotta work too :) [10:33] just remember that your server is a router. it counts every packet that goes through. possibly those counters are the ones you're looking for [10:34] hmm... point noted... ill get this done!... thanks RoyK!!! [10:43] I need someway to install Ubuntu server on 100 machines without having to do the installation procedures on all of them .. Is there someway that I can install only on 1 machine and replicate the installation over all the others ? [10:45] hello, where are the apache config files kept? [10:46] Hello all. Having problems with my mouse when I use a virtual pc via my vmware server. I think I have to add something to the xorg.conf file, but the xserver crashes when I alter anything in there. Tried a number of suggestions for altering the xorg.conf file, but didn't work. I have installed "apt-get install xserver-xorg-input-vmmouse" so in theory it should work, but clearly this package is not working on my pc. I am using kubuntu [10:47] xenaxon: /etc/apache2/ [10:50] anyone using vmware server2? [10:54] tgalal: there are multiple ways to achieve this. You can use kickstart, preseed or you can clone the install with clonezilla [10:54] TeTeT, which is the simplest ? [10:54] TeTeT, I need a hands-off installation .. i only want to touch 1 machine [10:55] tgalal: if they are identical hardware, clonezilla is most likely the easiest path, then kickstart, the preseeding [10:56] tgalal: do some testing with clonezilla.org, https://help.ubuntu.com/9.10/installation-guide/i386/automatic-install.html [10:57] tgalal: https://help.ubuntu.com/9.10/installation-guide/i386/appendix-preseed.html [10:58] any body using vmware server2? [10:59] TeTeT, thanks a lot .. that really helped [11:01] does apache start everytime I start ubuntu [11:01] kubuntu* [11:01] or do I need to start the service manually? [11:01] xenaxon: by default it is started when installed [11:02] New bug: #551544 in puppet (main) "puppet in lucid does not support upstart status " [Undecided,New] https://launchpad.net/bugs/551544 [11:02] New bug: #551545 in qemu-kvm (main) "PXE netboot not booting localboot from virtio-disk" [Undecided,New] https://launchpad.net/bugs/551545 [11:15] guys how do I start the mysql server [11:16] xenaxon: /etc/init.d/mysqld start [11:16] xenaxon, like most other services, like au said [11:16] thanks [11:16] welcome :) [11:16] or "service mysqld start" on newer versions [11:16] ubuntu has adopted that now, it seems [11:16] since when could you do that on ubuntu? [11:17] ahh [11:17] oh, sorry I'm on kubuntu [11:17] not ub [11:17] I missed that switching from centos that ubuntu [11:17] it says urecognised service/command not found [11:18] xenaxon: apt-get install mysql-server [11:18] that will install it for you [11:18] I did [11:18] did you double check? :p [11:18] ok [11:18] no [11:20] no such file or directory [11:20] and I installed it right now [11:22] TeTeT, I'm planning to use kickstart or preseeding ... isn't there any where I could define incremental IP addresses for my machines and usernames as well ? [11:22] in any of the 2 methods [11:22] pm me the output of what it said when you did apt-get install mysql-server [11:22] any way* [11:23] how can I check if mysql is running [11:32] xenaxon: sudo /etc/init.d/mysql status [11:32] what mysql client should I use [11:33] joschi do you recommend any mysql client? [11:33] interface* [11:33] xenaxon: whichever you like... [11:33] depends what you want to do with it [11:35] on windows I was just using the phpmyadmin [11:35] I need it to do manual database operations [11:35] like importing a database/table/entries [11:35] and also creating/editing table structure [11:42] xenaxon: try http://dev.mysql.com/downloads/workbench/ [11:42] xenaxon: or install phpmyadmin if you liked it and it did what you needed [11:48] tgalal: IP addresses should go in your DHCP server and dealt out from there. On usernames I fear you have to script [11:48] tgalal: e.g. you can script that they get each a different user in the %post section of the kickstart file [11:51] thanks joschi. I installed the workbench. [11:58] is it possible to import a database with the MySQL Workbench? [12:00] xenaxon: does it not have any documentation? [12:01] where are the database files kept in mysql [12:01] I just want to copy the database from my windows folder to the linux mysql folder [12:02] can I just copy the database table from windows mysql to linux mysql? [12:03] xenaxon: Try /var/lib/ for the DB data files and /etc/my.cnf (or similar) for cnfig. [12:03] But there are lots of docs around for MySQL. Have a look e.g. [12:04] http://dev.mysql.com/doc/ [12:04] on windows my dataase has it's own folder [12:05] I can't find folders for any of linux's databases [12:20] does anyone know how to import a database with MySQL Workbench? [12:20] it doesn't have any help files [12:20] ... [12:27] TeTeT, I don't want to use a dhcp server because it will be a single point of failure ... and I want the machines to be networked even of the dhcp server is down .. so I want to use static IP's ... any idea ? [12:29] Whereas with static network configuration, you have *two* SPOFs [12:30] how is a static network a single point of failure ? [12:30] Machines will still have network if the DHCP server is down. They typically just won't be able to reboot while the DHCP server is down. [12:30] tgalal: because both ends need to be statically configured to talk to one another. [12:31] this doesn't make it a SPOF .. because if a machine falls down .. the rest will be able to communicate [12:31] I have 100 machines [12:31] Anyway, there's no reason you can't have multiple DHCP servers using the same ethers database. [12:32] too much hassle [12:32] I need static IP's :-) [12:32] As a rule, I don't help people to do stupid things [12:33] this is not stupid .. this is the best for my setup [12:33] multiple dhcp servers is a stupid idea [12:33] 100 dhcp servers on 1 network ? [12:33] are you sure this is wise ? [12:34] 0 is a network. It has 2³² IPs in it. [12:34] ? [12:35] 0 as in CIDR 0. a.k.a. 0/0 or 0.0.0.0/0.0.0.0 if you want to be verbose. [12:35] <\sh> oh well...make the dhcp servers redundant e.g. pacemaker .. and use dhcp relay features of your switch [12:35] \sh: DHCP already works by broadcast [12:35] twb, I learned this in my network basics class [12:36] If you're using fixed IPs with a shared ethers database, it doesn't matter which one responds first. [12:36] hmm [12:37] Of course, distributing ethers to hosts direct and forgoing DHCP entirely wouldn't work too well, because you wouldn't have enough network to access the LDAP or NIS database that had the ethers in it. [12:39] <\sh> twb, you could use the dhcp relay features of your switch, when you want to use only one dhcp service for any networks in your local infrastructure...works very nicely when you have several native vlans and you need ip.. [12:43] does anyone know how to import a mysql database? [12:50] What the hell .. if I use kickstart I will still have to provide the 100 machines with the installation CD's ? [12:50] TeTeT, ^ [12:51] tgalal: no, you can install it conveniently from a PXE boot server [12:53] https://help.ubuntu.com/community/PXEInstallServer [12:53] TeTeT, thanks :-) [13:00] I need someway to generate incremental usernames and passwords for all the machines [13:00] machine1 password1, machine2 password2, ... etc [13:01] TeTeT, can I do so with kickstart ? [13:02] tgalal: you have to script it somehow and give each host a different ks.cfg [13:02] how do I give the hosts a certain kickstart file ? [13:02] tgalal: for example providing ks.cfg via a cgi script [13:03] will you please elaborate a bit ? [13:04] as far as I can understand the as soons as I boot the client machines in PXE mode .. they detect the server and run the designated kickstart file ... so how can I specify which file every client uses ? [13:04] tgalal: ks=//