[00:04] hello people .. I have one question .. I'm running 10.04 LTS with kernel 2.6.32-35 64 bit.. I'm looking for a safe way to upgrade the kernel to 2.6.35 without compiling it myself .. [00:04] I know about the http://kernel.ubuntu.com/~kernel-ppa PPA .. but I'm not sure that it contains the server edition of the kernel [00:05] can anyone tell me if adding the http://kernel.ubuntu.com/~kernel-ppa PPA to my sources list will upgrade the kernel to 2.6.35 server version or will i get the standard desktop version of the kernel ? [00:36] anyone have experience with ircd-hybrid? [01:06] Anyone know what disk /dev/dm-xx is? [01:07] dm is the device mapper [01:07] It's used for LVM, LUKS, and a few other things (md raid?) [01:08] Oki, well, got both lvm and md raid. But didnt have these before. And now they are there. [01:08] Strange. [01:08] You probably just didn't notice them. [01:08] Maby, === _Techie_ is now known as _TechAway_ [01:39] Hey guys, I read that Convirt 2.0 is now in the ubuntu partner repo - I can't seem to find it on my lucid installation though... [01:40] hey guys could anyone give me a hand recovering a raid-5 dm raid partition? i can't get it to mount after moving the drives to a new computer [02:55] So i have lamp installed, but how do i get my site on the web. ( I can view my site on my network, but my friend says that he can't at his house.) [02:56] do you have consumer internet or business class internet? [02:56] consumer. [02:57] go into your router and port forward a nonstandard port say 9999 to port 80 of your internal webserver [02:57] kirkland, it would be nice if you'd push the on 'iamcli' at https://launchpad.net/ubuntu/natty/+queue?queue_state=0&queue_text= for me. [02:57] then give your friend your external ip with the port number [02:58] if they use ie they will have to put http://yoursiteip:9999 [02:58] so wait. it is asking me for the start and end ports [02:58] just make them the same [02:59] start 9999 end 9999 [03:01] gonna throw my question up again...hey guys could anyone give me a hand recovering a raid-5 dm raid partition? i can't get it to mount after moving the drives to a new computer /dev/md0 shows up in mdadm.conf with the same uid but when assembling it says "mdadm: /dev/md0 assembled from 1 drive - not enough to start the array." but the drives are there and have the uuid listed [03:02] should I change the port in my config file to 9999? [03:03] crazifyngers, should I change the port in my config file to 9999? [03:04] you can if you want but if you do then it is port 9999 both internally and externally. if you leave it at port 80 and forward port 9999 to 80 externally then you don't have to change it [03:06] how do i forward to 80? there is no setting for that in forwarding [03:07] well it's different for different routers [03:09] what should I type in for server ip address? my network ip? [03:10] hmm... it won't let me type in the other one... it must be my local ip [03:10] what do yoiu mean server ip address? like what should you give to your friend? [03:11] it asks for server ip address. i think thats what i give to my friend [03:12] where does it ask for that? [03:12] in port forwarding [03:14] no the server ip is the internal ip [03:14] define internal? [03:15] you mean what i get when i do ifconfig? [03:15] for eth1 [03:15] yes [03:16] internal=inside your lan [03:16] ok [03:16] external=outside your lan [03:16] and it should be TCP/UDP right? [03:17] depends on the website. some streams do udp. but you should be ok with both [03:17] so i give him my external with port like this: xx.xx.xx.xx:9999 [03:19] yea [03:19] find your external if you don't know it go to ipchicken.com [03:20] yeah i already have it. but i dont think it's working. [03:28] should i try another port? [03:29] i would check your router [03:29] if you can get to it internally then maybe it is the port but doubtful [03:30] my router is a piece of crap netgear [03:30] i might as well throw it out the windows and let my dogs chew on it. === _TechAway_ is now known as _Techie_ [03:33] maybe it is because i am on wireless connection? should i switch to wired? [03:34] no a connection is a connection === squishy is now known as rbniknej === rbniknej is now known as jenkinbr === jenkinbr is now known as squishy [03:50] jdstrand: the i386 qemu-kvm for natty is back up [03:51] alas, armel build failed, will have to keep looking into that :( [03:52] GAH! it just ran out of memory [04:02] Does anybody know the ETA of getting Convirt 2.0 into canonical's partner repo? [04:02] i saw the announcement but it doesn't seem to be there at the moment [04:05] My ext3 root filesystem's "last mounted time" is in the future [04:05] This causes mountall to halt the boot process and just hang forever. [04:06] How can I set the last-mount-time parameter to now? tune2fs doesn't appear to have such an option. [04:09] Hm, I'll try dropping into busybox to do it === _Techie_ is now known as _TechAway_ [04:14] anyone use rssdler in here? i had it all setup, and it now only seems to run once instead of always running & checking on an interval... [04:14] i guess i could just setup a cron job to start it. [05:56] does anyone know, if I send data from one EC2 server to another in the same zone but using the public IP address, does AWS count it as data transfer out? [06:32] New bug: #676830 in munin (main) "pid_file in munin-node.conf is ignored." [Undecided,New] https://launchpad.net/bugs/676830 [06:42] New bug: #676832 in eucalyptus ""Ghost" instance using a slot + elastic IP" [Undecided,New] https://launchpad.net/bugs/676832 [06:42] New bug: #676833 in munin (main) "snmp__* plugins don't like host names." [Undecided,New] https://launchpad.net/bugs/676833 [08:14] morning [08:15] can anyone tel me if this result of 'top' looks unusual in respect to memory usage. I keep getting screen sessions dropping out and am wondering if I keep running low on RAM: http://paste.pocoo.org/show/292944/ [09:52] hi there, i've got a 10.10 server machine i want to build VMs on. i've tried with virt-install, but get an error ("failed to retrieve chardev info in qemu with 'info chardev'"), now looking at vmbuilder - but if i understand correctly vmbuilder pulls an install over the network, right? [09:56] pvh_sa: correct, it will d/l packages from archive.ubuntu.com, unless you specify a local mirror [09:57] Hi guys... I'm seeing some reference in the 10.10 release announcements about gluster and ceph... does anyone know stable they are for real-world use in a cloud environment? [09:58] TeTeT, ok, not enough network here for that. back to trying to get virt-manager to work. thanks [10:01] pvh_sa: good luck. I use virt-manager on Lucid quite often and it worked fine for me so far [10:02] TeTeT, i've googled this error and it comes up quite often but often with different causes. i suspect what might be happening is *another* error which means this 'info chardev' is just a symptom [10:04] pvh_sa: never seen it, sorry === ihCiFriS is now known as SirFiChi [10:25] can anyone tel me if this result of 'top' looks unusual in respect to memory usage. I keep getting screen sessions dropping out and am wondering if I keep running low on RAM: http://paste.pocoo.org/show/292944/ [10:34] j0nr: Looks fine to me [10:39] has anyone here created a ubuntu vm using virt-install on 10.10? if so could you offer the commandline you used? thansk [10:39] thanks even [10:47] using 10.4 pvh_sa [10:47] yann2, i'm going to install 10.04 just now to see if it works [10:51] on 10.10, i'm trying to install with a slightly altered version of one of the virt-install examples: virt-install --connect qemu:///system --name demo --ram 500 --disk path=/var/lib/libvirt/images/demo.img,size=5 --network network=default,model=virtio --disk path=/root/ubuntu-10.10-server-amd64.iso,device=cdrom --vnc [10:52] * [diablo] is looking forward to his meeting with Canonical today at their offices... we plan to deploy Ubuntu Server in our data centre :-) bye bye CentOS [11:09] [diablo]: Congratulations. [11:10] Daviey: Have you ever used Eucalyptus in SYSTEM mode? [11:47] So here's a question for you lovely people... does UEC run Windows images alright? I can't seem to find much documentation either way... [11:51] pws: did you google? [11:51] theres some howtos on howto deploy window images in eucalyptus opensource version [11:52] yes, I can find some vague references to 'I think it should be possible' and 'I'd like to', but nothing definite [11:52] Ah, that'll probably be a good thing to search for then, I'll go and look :-) [11:52] Thanks [12:02] soren: the mode where instances get their IP from the host network dhcpd? [12:06] soren: the mode where instances get their IP from the host network dhcpd? [12:09] Daviey: Yes. [12:09] Daviey: I think I found my answer now. [12:09] soren: groovy [12:10] what do you use that cloud stuff for? [12:10] Daviey: I was just wondering what sort of IP Eucalyptus would claim they had. A couple of forum posts suggest 0.0.0.0. [12:11] soren: Yeah... i can't see how it could know [12:12] soren: If they were really devious they could do an arp lookup [12:12] I have used SYSTEM mode, but didn't observe the IP field output tbh... mainly that it was working :/ [12:16] Daviey: Oh, I didn't expect it to know. [12:17] Daviey: I was just wondering if it actually assigned one internally ( and exposed it in euca-describe-instances) or if it actually didn't assign any. [12:17] soren: I could check :) [12:18] Daviey: I couldn't possibly ask you to do anything with Eucalyptus. I like you. [12:18] :o [12:19] soren: Now you mention it... my rig is currently tied up with other software atm. [13:25] * bond hi [13:30] Daviey: someone uploaded the new eucatools yet? [13:40] zul: no... smoser has it in hand tho [13:41] Daviey: if you want me to review it and upload it let me know.. [13:42] zul: Well i did a pass of smoser's branch yesterday; but you may have seen that all of his patches have now gone upstream (\o/)... so i imagine smoser will track that for this upload.. [13:42] If smoser can get that done today, i'm sure he'll appreciate a second pair of eyes [13:42] cool... [13:48] zul: I will add you on a write-up of the SRU thingies you will give me [13:48] hggdh: sure [13:48] (that is, in my BP) [13:49] zul: do you want to do the same with the admin of the dailies? [13:49] hggdh: sure [13:49] heh [13:50] Daviey: Ok, so I want to get the packaging branch in order. [13:50] soren: "in order" ? [13:50] Daviey: Yes, as in "get it sorted out". [13:51] soren: Oh awesome, what are you thinking? [13:51] Daviey: My current inclination (which changes by the hour, it seems) is to have a repository with just debian/* in it. [13:53] Daviey: That makes it simple to merge it into whichever branch of Nova you feel like building packages out of. [13:53] that would be good for me i think [13:53] and then you guys can cherrypick anything you guys want [13:54] Daviey: ...and if you don't want to use upstream code from bzr (which I, somewhat ironically (being upstream and all) don't), you can just check out this branch from the top-level dir, fiddle with things, and bzr commit does what you expect. [13:54] Daviey: Instead of having the root of the repo /be/ (rather than /contain/) debian/, in which case debcommit, for instance, no longer works. [13:54] soren: i entirelly agree that the packaging branch should not have the code in [13:55] hmm [13:55] soren: i was thinking ROOT:/debian/ [13:55] so otherwise empty [13:56] this means that debcommit does still work [13:56] Daviey: I'm not sure I understand. [13:56] soren: a bzr tree with only one top level directory [13:56] Yes, that's what I'm suggesting. [13:56] i agree with that [13:57] Cool. [13:57] soren: Do you have plans for having a consistent location for per commit tarballs? [13:57] (/me is thinking, not hosted on LP) [13:57] Daviey: Already have that. Since yesterday. Try to keep up :) [13:57] ooooooo [13:57] what is it? [13:57] Daviey: http://nova.openstack.org/tarballs/ [13:58] soren: and that includes the magical fairy dust, that bzr doesn't - right? [13:58] Daviey: Indeed. [13:59] soren: Ok... great [13:59] soren Daviey: the ubuntu-natty has config file changes that i had to do in order to get it working for natty [13:59] zul: Yeah, some of the defaults changed, too. We need to account for that somehow. [13:59] zul / soren: There probably needs to be two tree's now... one for Maverick and one for Natty [13:59] We should also make all the various components use the same config file. [13:59] Daviey: Yes. [14:00] soren: agreed [14:00] soren: Do you want to clean the tree out? [14:00] the apport hook is already done as well [14:00] zul: damn.. you are fast. [14:00] Daviey: I do, I just wanted to get the bzr repo layout nailed down first. [14:00] Daviey: i had a 3 hour plane ride ;) [14:00] Daviey: But that seems to be a done deal now. I'll whip up a new repo so we can start from scratch. [14:01] If I didn't (and just bzr removed everything except debian/), merging it into code trees would be very awkward. [14:01] soren: Ok.. great.. We are still all peer reviewing each other [14:01] right? [14:01] Oui. [14:01] \o/ [14:02] there's an openstack-ubuntu-packagers team on lp. [14:02] yeah, saw it [14:02] I'll make ubuntu-virt a member, along with Monty. [14:02] soren: A clean tree makes sense IMO [14:02] I just didn't get around to it yet. [14:02] Daviey: I just need to make sure I can actually merge it then (i.e. it won't complain about lack of common ancestry). [14:02] Daviey: I'll work it out. Give me 20 minutes. [14:03] soren: hmm.. should it ever need merging? [14:03] Daviey: "need" is such a strong word. [14:03] I think it's a valid use case. [14:03] It's needed to make recipe based builds work. [14:03] bzr branch lp:nova ; bzr branch lp:~openstack-packagers/etc debian ; cd nova ; ln -s ../debian debian [14:03] ? [14:04] hmm [14:04] that is probably unwise actually :) [14:04] I can see how it can be convenient. [14:04] If you just want to track bleeding edge upstream and packaging, it's kind of handy to be ably to just use bzr for that. [14:05] It's not super important, but if it's easy enough to accomplish, I might as well. [14:05] soren: You can merge without common ancestry, in some situations... but TBH... working out how (i'm not clear myself), is probably a higher price than the reward [14:06] Daviey: I have a trick up my sleeve. [14:06] Daviey: You'll see. [14:06] :) [14:06] (if it works) [14:06] * Daviey hopes it's not rude. [14:07] soren: example reciepe https://code.launchpad.net/~vcs-imports/couchdb/trunk/+recipes [14:07] zul: Oh, we're already using recipes to build nova. [14:07] I just don't want to do that anymore. [14:07] For a couple of reasons. [14:08] such as? [14:08] a) They're only built once a day. I want a package per commit. [14:08] TBH.. i'm not sold on recipes myself tbh. [14:08] so like a hudson hook or something? [14:09] b) We're building tarballs per commit. [14:09] b1) The tarballs have stuff in them that are not in bzr. [14:10] b2) The reason we resort to building from vcs is because upstream fail to provide tarballs per commit. We provide them, they should be used. [14:10] (I use "we" very confusingly to mean "Ubuntu" and "openstack" as I go along) [14:11] gotcha....what are in the tarballs that bzr doesnt have beside the .bzr directory (assumingly) [14:11] soren: With Mythbuntu we build daily (via cron fired script) for ~3 ubuntu releases and 2 upstream version snapshots (trunk and stable/fixes)... that seems to work well [14:11] users are normally happy. [14:12] b3) Even if we were to build from bzr during the dev cycle, once there's a release out, we should use the tarball. vcs snapshots are used to be able to track development when releases are too infrequent. It's error prone to have one way to build packages during dev and another once we near release (i.e. when the tarball is released). [14:13] b4) Everyone besides Ubuntu will be using the tarballs. Having Ubuntu also build from tarballs feeds into OpenStack's QA. We can be sure everything needed to build Nova properly is in the tarball. [14:13] c) Upstream is kindly asking that this is to be so. [14:13] that's a bit more than a couple, but I hope that's ok :) [14:18] soren: I agree with your comments fwiw. [14:18] zul: There's a ChangeLog. [14:18] zul: More stuff may turn up. [14:19] soren: right [14:19] zul: just noticed smoser is on holiday today and tomorrow... so re: euca2ools... it might make sense to leave until Monday, there is no huge rush, i guess [14:19] zul: it's like how you typically don't see the configure script in vcs for C projects, only the stuff needed to generate it. [14:19] Daviey: ack [14:19] zul: Part of making a tarball is building that configure script. [14:20] zul: ..and other sorts of things. Right now, it's just the changelog and maybe I leave out a bzr plugin that generates said changelog. [14:20] zul: I'm not sure. [14:20] soren: cool...i would love to see the script eventually for my own projects [14:21] zul: It's mostly done. [14:23] soren: throw it into a bzr branch when you are done if you havent already [14:25] Daviey: My trick failed me. :( [14:25] soren: what was it? [14:25] Daviey: See in #zr. [14:26] #bzr, that is. [14:26] k [14:26] Scroll back about a pageful. [14:26] Does anyone recommend a good NMS to monitor Asterisk [14:27] soren: oh aye,.. good idea [14:27] Daviey: Not good enough. [14:27] Lord_Rahl: depends exactly on your requirements [14:28] Lord_Rahl: nagios works well for checking it's working :) [14:29] Logwatch doesn't seem to want to read my samba logs, anyone have a reason for this? [14:30] Daviey, I have try Nagios very labor intensive to setup. maybe I can find a plugin to add nodes, editing all the configs by hand sucks :) === jdstrand_ is now known as jdstrand [14:43] Daviey: This kind of pisses me off a bit. [14:44] Daviey: It seems like a perfectly reasonable thing to be able to do. [14:45] soren: ack... svn allows it :) [14:45] infact, svn allows you to have external depends, that get pulled in when you co [14:46] Daviey: That's coming for bzr, too, I believe. [14:47] Daviey: Mostly driven by the need to be able to import such things from svn (and git, which also supports it). [14:47] This is why we can't import libvirt to bzr on launchpad. It nests gnulib. [15:06] New bug: #677019 in backuppc (main) "package backuppc 3.1.0-9ubuntu1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/677019 [15:11] Daviey: Not sure how to resolve this. There's not way to make up a common ancestry, apparantly. [15:11] Daviey: bzr-builder can deal with it, as it turns out. It has a "nest-part" option that does the right thing. [15:12] Daviey: The problem is if someone wants to manually merge them. It's doable (bzr merge -r 0..-1 lp:blahblha), but not obvious. [15:13] Bah, let's just do that . [15:15] can anyone recommend any SNMP s/w to use with IPv6 ? [15:36] Daviey: lp:~openstack-ubuntu-packagers/ubuntu/natty/nova/ubuntu [15:38] Daviey: One thing that irks me a bit, though.. [15:38] Daviey: If we push a fresh version to Natty, that pretty much blows my chances of putting the Austin release of Nova into Maverick. [15:38] Daviey: But I'm not completely sure that I care, really. [15:39] zul: Can you branch lp:~openstack-ubuntu-packagers/ubuntu/natty/nova/ubuntu and apply those changes of yours to that and do a merge proposal? [15:39] soren: sure [15:47] Daviey, zul: Awesome. "bzr bd -S" just works. It grabs the tarball from http://nova.openstack.org/tarballs/ automatically. [15:47] sweet well do that going forward probably [15:48] Yup. I'm looking forward to see how it works out when there's a new upstream revision. [15:50] jono: Mr. O'Bacon! [15:50] hey soren [15:50] jono: did I tell you one of my colleagues thought you were "John O'Bacon" when we were at CLS? That still cracks me up :) [15:50] lol [15:51] not the first, and probably not the last :) [15:51] There would have had to be an Irish guy with a first name of Bacon at some point for that to work. That'd be pretty awesome. [15:59] soren: http://paste.ubuntu.com/533899/ [16:00] oh wait...i left ubuntu-virt i think [16:02] soren: yeah i did [16:02] hiya guys. just did a do-release-upgrade from 9.04 to 10.04.1. my /etc/lsb-release still refers to jaunty. I looked in the log for the upgrade and it said it installed the package lsb-release [16:03] i also reinstalled lsb-release with no effect. [16:03] is that safe to purge and install? [16:09] zul: Should I add you back to ubuntu-virt, or should I add ~motu to the packagers team? [16:09] zul: I'll do the latter regardless. [16:10] zul: Try again. [16:12] zul: Err.. Hang on, you were supposed to file a merge proposal, by the way. [16:12] zul: Not just push. [16:18] Daviey: should I expect Maverick's UEC to boot and work a Natty image? [16:20] soren: yeah i was trying to push to a seperate branch in the openstack-packagers team [16:20] could you add me to the ubuntu-virt team again? [16:23] zul: Just push to something under ~zulcss. No need to make it a team branch when it's being merged anyway. [16:24] soren: sure hold on [16:24] zul: but you're a member of openstack-ubuntu-packagers anyway now (by way of ~motu). [16:25] ack [16:25] zul: ...and now ~ubuntu-virt, too. [16:25] soren: i dunno if you want to add ~motu there though [16:27] zul: they can upload it anyway. [16:28] soren: well the merge request is there now [16:28] zul: no point in not letting them push to bzr first. [16:28] soren: i should be able to do merges in the future correct as well right? [16:29] zul: Yes. [16:30] soren: oh crap i just need to clear out one thing [16:30] zul: You need to clear out everything, it seems. [16:30] zul: Errr... What the? [16:31] soren: gimme a sec [16:31] zul: You've filed a merge proposal to get the branch I just created merged into lp:nova/ubuntu. [16:31] That's not what we want. [16:31] hold on [16:34] JamesPage: so... let's get Hudson rolling? :-) [16:35] hggdh: sounds like we have the required approvals so yes! [16:37] soren: thats what it was defaulting to in launchpad. should be fixed now. [16:37] Don't see any merge proposals.. [16:37] hi [16:37] zul: hey - I plan to remove myself from the administrator list for ubuntu-server-bugs [16:37] i just clicked on the button ;) [16:37] zul: anyone else to replace me? [16:37] hggdh: do you want to spin up the ec2 instance or shall I? [16:37] zul: you'll be the only admin then [16:38] zul: I checked after you said it. [16:38] mathiaz: hggdh maybe? [16:38] ive got a server that does not have a correct /etc/lsb-release [16:38] JamesPage: either way... [16:38] zul: me what? [16:38] hggdh: do you mind if I set you as an administrator for ubuntu-server-bugs? [16:38] I trying to install ubuntu server on hp dl120 g6 but it didn't recognize logical drive for SATA Array configured [16:38] mathiaz: not at all, please go ahead [16:38] do release-upgrade appears to have me up to 10.04.1 [16:38] Is this the best channel or is there another for running Ubuntu as virtual machine (standalone or in cloud)? [16:38] soren: launchpad is slow maybe [16:38] hggdh: which email address? [16:39] i just got the merge request like 2 minutes ago [16:39] mathiaz: hggdh2 at gmail dot com [16:39] zul: You filed it against the maverick branch. [16:39] zul: Try again :) [16:39] looks like ubuntu server doesn't recognize array controller b110i [16:39] fuuuuuuc [16:39] any suggestions? [16:40] jgjones: either here or on #ubuntu-cloud [16:40] hggdh: thanks [16:41] soren: cross fingers :) [16:42] Daniel__: it detects none of your disks at all? [16:43] no it detects both fisical disks, but it doesn't detects the logical one 1+0 raid [16:43] oh, it's not a real raid card [16:43] soren: should be there now [16:44] the array card is hp smart array B110i [16:44] ya, it's a softraid card [16:44] Daniel__: its a fakeraid then [16:44] zul: It would really be helpful if you could actually describe the changes you're making. [16:44] consider using os level software raid instead [16:45] soren: noted [16:45] zul: I mean... "Update build dependency" isn't helpful when I have to review things. [16:46] yes it use a software provided as a small bios for HP [16:47] soren: I have updated the commit message in the merge proposal [16:47] and will have better commit message in the future ;) [16:48] hi - i need django 1.2.3; i notice it's in sid [16:48] is there a recommendation for getting this onto a lucid machine? backports? mirror sid and pin it? [16:48] mgolisch you mean I can't use the builtin raid system, instead I have to use ubuntu software raid? [16:49] Daniel__: yeah or search for a driver [16:49] as its fake/software raid anyways id just use md raid [16:50] mgolisch I found a driver for redhat linux but I don't how to install it with ubuntu [16:55] zul: I'll comment on it after I've eaten. [16:55] soren: sure [16:55] i need to go eat as well [16:57] what about performance? It would be better to try find a driver or ubuntu server raid will do the job with same performance? [17:27] Daniel__: performance with s/w raid is fine [17:27] Daniel__: fakeraid is just, bleh [18:09] ok guys, thank you very much for your help! [18:16] Nov 18 10:08:12 ltsp1 kernel: [ 3943.664469] nfs: server fs1 not responding, still trying [18:17] this is causing my machine to hang... any ideas on troubleshooting this? fs1 is pingable [18:17] don't hard mount it, or use hard + intr [18:17] man mount :) [18:19] checking manpage [18:20] the strong intr does not exist in the manpage... ? [18:21] http://www.faqs.org/docs/linux_network/x-087-2-nfs.mountd.html [18:21] =] [18:21] oh heh, they moved the nfs stuff to man nfs [18:22] the file server, has 7.X loads... [18:22] mine gets up to 30+ at times [18:23] but it does take some work, to get it that high [18:24] if I throw more processor/ram at it, will it help? or just delay the inevitable? [18:24] it would help (very little) [18:24] it's probably disk i/o it's waiting on [18:25] hmmm [18:25] possibly. any good way to test this? [18:26] top :) [18:26] vmstat :) [18:26] i never understand the output of vmstat [18:26] it's the same as top :) [18:26] r = processes that are attempting to run [18:26] b = processes that are blocked? (never see this >0) [18:26] swap drive usage [18:27] free ram [18:27] buffer memory (to be written todrive) [18:27] cache (drive cache usage) [18:27] si/so swap in/out k/s [18:27] bi/bo disk in/out k/s [18:27] in interrupts/s [18:27] cs context switchs/s [18:27] us user cpu usage [18:27] sy system cpu usage [18:28] id idle cpu usage [18:28] wa wait cpu usage (this shows slow harddrive performance) [18:28] http://pastebin.com/gbmdN6bs [18:28] so the higher the wa is, the faster you need to make your drives go, basically [18:28] but I'm not sure what "sane" numbers would be, to compare to. will any machine do? [18:28] to benchmark it against [18:28] doesn't matter [18:29] it's what ever is applicatable to you, and your usage [18:29] generally why benchmark numbers don't mean much [18:29] unless your doing the exact same thing as that benchmark [18:29] vmstat: http://pastebin.com/9KZEsWxi [18:30] that is useless [18:30] you need like a lot of lines of output from vmstat [18:30] seemed like it. it needs to update continually [18:30] yeah :) [18:30] vmstat 10 [18:30] let it for for like 10min or so [18:30] the first time is a summary line, since boot [18:31] defently use smartctl on your drives tocheck their health too [18:31] okay, can do. [18:32] this is a dell poweredge 2950 server, so it's not just a homebrew machine, too... [18:32] raid? [18:32] raid5 [18:32] then smartctl probably won't work [18:32] but also running DRBD, if you are familiar with it [18:32] make sure you check the raid health [18:33] drbd slows everything down to a crawl [18:33] cause it has to confirm writes on the other machine [18:33] <_ruben> depends on the drbd config tho [18:33] it's set in dual primary mode [18:33] well, using non-write confirmation isn't that safe :) [18:33] <_ruben> ah, you like living on the edge [18:34] i only write to one still, currently... [18:34] it still has to transfer that write from the first machine, to the second [18:34] confirm it's written [18:34] get that confirmation back, and tell your nfs client about it [18:34] <_ruben> indeed, if any way possible, disable the sync for a short while and see what happens [18:34] <_ruben> too many layers currently where it could be going wrong [18:35] indeed [18:35] ya, and with drbd like that [18:35] either one could be slow, not both [18:35] so we have to figure out what machine also still :) [18:35] <_ruben> true ;) [18:36] so, the setup is like this... [18:37] ubuntu 10.10 host, xen virtualization.. ltsp server (ltsp1), fs1 drbd with fs2 on another physical server [18:37] ldap for authentication throughout, and it shouldn't matter, but there is an asterisk server hosted on here too... [18:37] ltsp fs1 are all xen guests? [18:37] yes [18:37] <_ruben> sure hope not ;) [18:37] <_ruben> ouch [18:37] yuk [18:38] another layer of issue :) [18:38] recently, the thin clients (ltsp1) have been freezing a lot. has worked great for 2 years... now, ltsp1 can't see fs1 on occasion.. happening more often.. [18:38] when the thin clients freeze, ltsp1 has 70+ loads, and fs1 has 6-7. [18:38] you have lots and lots of stuff to look at [18:38] <_ruben> could very well be that the load reached a tippingpoint .. seen it far too often [18:39] <_ruben> could be a minor increase in load triggering it [18:39] the users have not grown... same amt of users, too.. [18:39] <_ruben> perhaps some decided to multitask (more) [18:39] perhaps. [18:39] I doubt even that [18:40] 10.10 wasn't out 2 years ago [18:40] on vmstat, id = 100 the whole time. :) [18:40] no, 2 years ago, it was 8.04 [18:40] just it's usage of cpu caches has changed I bet [18:40] and could of done this [18:40] if you where close already [18:40] the swap to 10.10 was about 4 months ago [18:40] <_ruben> even more living on the edge eh? [18:40] I noticed on my system atleast, switching from a 1m to 2m cpu cache speed it up a crapload [18:41] <_ruben> it was release like 5 weeks ago [18:41] hey people, I have a question largely because apache directory has not worked out for me. I am looking for a guide out there that will help me config LDAP on Ubuntu Serv 10.04 to push windows updates and other software to like 20 computers. I failed in my google effort. any ideas? merci! [18:42] <_ruben> burntoutlamp: push windows updates using a linux box? doubt that's gonna work (never looked at it tho) [18:42] blast I need an active directory I cannot go on like this -.- [18:42] <_ruben> burntoutlamp: for stuff like that, yes you do indeed [18:43] my bad: root@talon:~# cat /etc/issue [18:43] Ubuntu 10.04.1 LTS [18:43] * burntoutlamp backs up the question trolley [18:43] so is it possible to push out windows updates using a linux server to 20 windows laptops -.- merci beaucoup [18:43] <_ruben> burntoutlamp: there might be alternatives, tho i doubt they'd bring any joy to the person admin'ing it ;) [18:44] <_ruben> burntoutlamp: same answer, i highly doubt that, and if there's a way, it sure won't be a pretty one [18:45] <_ruben> hell, even pushing out updates for 20 linux laptops is quite the challenge i bet ;) [18:46] I am the one adminning it [18:46] * burntoutlamp cries [18:47] why not allow the machines to update themselves? trying to save the bandwidth? [18:47] <_ruben> burntoutlamp: given that the windows options are quite complex on their own (which doesnt really tell all that much really), i really doubt there's any FOSS alternative capable of the same/similar [18:47] trying to distribute them locally, for speed? [18:48] yeah [18:48] right now I have to do it manually at each machine [18:48] right, but you could set them to auto, or run boot scripts? [18:49] if bandwidth is the issue, you can use squid3 to cache the files :) [18:49] hmmmmmm [18:49] I'd never thought of those alternatives [18:50] <_ruben> then again, windows update doesn't like proxies much [18:50] agreed. when I did that, I set it up at the router level. [18:50] so Windows update didn't know the difference. :) [18:50] <_ruben> transparent proxy? that'd work i guess [18:52] ah very cool [18:52] you guys are giving me some options I have not even thought of. I don't have nor am I able to set up a windows server here [18:53] my vmstat output: any advice? http://pastebin.com/v0JfgEGR [18:55] <_ruben> nothing out of the ordinary there [18:56] aisexec is using lots of resources, it seems.. [18:56] server is still at 7.0 load, with no traffic to it.. [18:59] I can reset fs1, but in a few hours, it will happen again. [19:10] Daviey: ping let me know when you are around. === JFo is now known as JFo-food [19:45] New bug: #677153 in dhcp3 (universe) "dhclient assigns all interfaces connected to a bridge addresses" [Undecided,New] https://launchpad.net/bugs/677153 [19:52] New bug: #677152 in dhcp3 (universe) "dhclient started on a bridge tries to assign addresses to all bridged devices (dup-of: 677153)" [Undecided,New] https://launchpad.net/bugs/677152 [19:53] hi all, I'm a linux newbie... and I ran into a real problem! I ahve been backup server 10.04 for the last few months with Clonezilla... now my boss wants me to install re-install our setup onto a new server with a different processor, raid system etc... and I'm getting a gazillion errors...!! ubuntu doesn't like all these changes... is there an old school method to use (TAR ??) [19:54] instead of a disk image, would this TAR method of backup and restore work? or is that just used for data files and not OS files... like will it resore my mysql installation, php stuff etc... or just the data files [19:57] mathiaz: ping where are we at the package set stuff? === JFo-food is now known as JFo [20:00] need some help I am trying to enable php5-pdo and pcntl.so for lilac. I have all the ini set but it still come up that I do not have have enable. [20:00] bmw: usually making a database dump and copying the data files for other apps should do well [20:01] fyi it need them for php-cli [20:02] RoyK: problem i'm worried about is that php for this particular software on the server etc.. took hours to configure... i want that all to come across exactly [20:02] most of php is configured out of the box if you just install it from the packages [20:03] Lord_Rahl: you need what of php-cli? [20:03] pdo & pcntl [20:03] zul: IIRC the DMB has taken ownership of the ubuntu-server-devs team [20:03] zul: and handles upload privileges for the package set [20:03] zul: the next step is to review the packages in there [20:04] zul: and that I don't know how to review the 350+ packages in there [20:04] seriously 350+ ? :) [20:04] RoyK, I have both .so but they are not loading is there a ini for php-cli [20:05] Lord_Rahl: /etc/php5/cli/php.ini [20:05] thanks I will check it out [20:05] that's on 10.10, but IIRC it's the same on lucid [20:08] RoyK, Ok I added the to be loaded. Is there a way to reload the cli with out a reboot? [20:09] Lord_Rahl: no need for a reboot [20:09] hey guys [20:09] question for yall why am i getting this error [20:09] RoyK, I restart apache but that did not do it [20:09] Nov 18 20:47:51 eagle dovecot: IMAP(jonathan): fchown(/home/jonathan/Maildir/.Sent/dovecot-uidlist.tmp, -1, 114(dovecot)) failed: Operation not permitted (egid=1000(jonathan), group based on /home/jonathan/Maildir/.Sent) [20:09] Lord_Rahl: are you using php-cgi? [20:09] yes [20:09] not fastcgi? [20:09] everything works fine sending and recieving but any emails i sent it cant add them to the send folder how can i fix the error above [20:09] no [20:10] cgi is something that belongs in the ninetees [20:10] cgi is _bad_ [20:10] lol [20:11] I once DoSed one of Norway's largest news sites because of silly CGI [20:11] trying to fix a poll :þ [20:11] cgi is simple, fork out a php/perl/something process on each request [20:12] fine, but what if there are 10k requests, and each process has a memory footprint of 3 megs........ [20:12] eagles0513875, does the user own the folder? are you using maildrop? [20:12] Lord_Rahl: yes user and group is jonathan [20:12] with drwx [20:12] permission [20:12] and im using dovecot + postfix [20:15] did you do a mkdir or maildirmake [20:16] eagles0513875, did you do a mkdir or maildirmake [20:16] New bug: #677161 in openssh (main) "tunnelled clear text passwords" [Undecided,New] https://launchpad.net/bugs/677161 [20:18] mkdir Lord_Rahl but i think i have isolated the problem [20:18] which is the group being dovecot === luis__lopez is now known as luis_lopez [20:18] eagles0513875, sweet [20:20] still not working lool [20:21] whops meant Lord_Rahl [20:21] Lord_Rahl: Nov 18 21:18:28 eagle dovecot: IMAP(jonathan): fchown(/home/jonathan/Maildir/.Sent/tmp/1290111508.M7146P6450.eagle, -1, 114(dovecot)) failed: Operation not permitted (egid=1000(jonathan), group based on /home/jonathan/Maildir/.Sent) [20:22] eagles0513875, do you have a tmp folder under .Sent [20:23] Lord_Rahl: ya i do and all permissions are the same on all folders [20:26] eagles0513875, Dont know maybe someone else can help. I use maildrop for that. [20:27] eagles0513875, I can give you may script for senting up a mail server if that will help [20:51] New bug: #677181 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 5.1.41-3ubuntu12.3 failed to install/upgrade: el subprocés seqüència pre-installation nova retornà el codi d'eixida d'error 1" [Undecided,New] https://launchpad.net/bugs/677181 [20:59] vsftpd won't start. says the process is running, but its not. http://paste.ubuntu.com/534005/ any ideas why? [21:01] GeekyAdam, how do you know it is not running? [21:01] raubvogel: check the code, i did a "ps aux | grep vsftpd" with no results. [21:02] Could it have started (PID 15984) and then crashed without cleaning up after itself? [21:02] Is anyone familiar with UEC? I have a few quick questions. [21:02] As in need to see the logs? [21:03] Hey guys - I notice that canonical and convirture have partnered to offer convirt through the partner repository - anybody got this running? I can't find it after updating? [21:03] raubvogel: logs empty :/ [21:03] After all, you said it claims to be running [21:03] Usually that means a lock or pid file somewhere [21:03] raubvogel: what do you mean by "lock or pid file somewhere"? [21:05] UEC Question: Say you have a cloud of 9 servers and you're running a VM within the cloud, if the node that's running that VM has a hardware failure, will UEC automatiacally move that VM to another node with no downtime? If not is that node recoverable without getting the down node back online? [21:05] VM* [21:08] GeekyAdam, it does not seem to have the pid stored in a file [21:09] I find it odd it did not even bother to write to its log file after you tried to start it manually [21:15] raubvogel: i removed vsftpd via apt-get, deleted the conf file just to have a fresh one, reinstalled vsftpd, and again, it says its running at the end of the install, but its not running. and there's no log file. wtf??? [21:15] raubvogel: wait, there's no conf file either now...? [21:16] if this were windows id be recommending a restart of the machine right about now... [21:16] Try to remove it using apt-get remove --purge and then reinstall it. [21:17] Then look for the pid it reports [21:17] good idea. purging... [21:18] raubvogel: that did it! running fine now. thx much. ^_^ [21:18] Sure thing [21:23] [21:49] anyone engaged Kernel: Neighbour table overflow issues? [22:07] anyone here have any experience with squid? [22:26] New bug: #677231 in mysql-5.1 (main) "Can't purge mysql-server" [Undecided,New] https://launchpad.net/bugs/677231 [22:46] smoser: ping [22:47] smoser: i tried to reproduce bug #676605 [22:47] Launchpad bug 676605 in byobu "byobu broken on natty" [High,In progress] https://launchpad.net/bugs/676605 [22:47] smoser: but couldn't [22:47] smoser: seems to be working correctly for me installed from today's daily natty server [22:47] smoser: when did you try? [22:51] how can I get the package that is related to an installed file? [22:53] Aison: can you give me an example ? [22:54] well, eg. I would like to know what package installed the file "/usr/bin/nano" [22:54] but that's just an example ;) [22:54] Aison: apt-file is a good tip [22:55] ok, nice [22:56] Aison: you can use "dpkg -S /usr/bin/nano" also [22:57] a nicer way [23:03] thx [23:22] anyone know why my ubuntu server wouldnt be reachable by its host name? my windows machines can all ping each other by their computer names, and my ubuntu server can ping the win PCs by name [23:23] is there something related to netbios that i have to configure? [23:25] entropy4: is there a dns entry for it ? [23:25] entropy4: are the windows boxes in an AD domain ? [23:27] ok so i have apache running, but firestarter (my firewall control gui thing) doesn't even show it on the current connections list. does this mean it isn't even getting on the internent? [23:27] *internet [23:27] snake: telnet localhost 80 see if it's running [23:28] what do you mean? is that a command? [23:28] snake: yes [23:29] it says that it's connected. [23:29] snake: ok, so you know it's running, that's a good start [23:29] snake: what are you firewall rules ? [23:30] so then why is it not working? I can't give my friend my external IP and let him view my web server? [23:30] ikonia, i'm not running a DNS server, and no domain either [23:30] snake: are you using a nat on your router ? [23:30] what is nat? [23:30] entropy4: the linux machine will need a dns server to resolve host names or an entry in /etc/hosts [23:31] snake: it translates a private IP address to a public one [23:31] ok how could i figure out if i have that [23:31] ikonia the linux server has no issue resolving hostnames. its other machines that cannot connect by name to it [23:31] snake: what is your machines IP address ? [23:31] entropy4: how is it resolving hostnames without dns or a host file ? [23:31] entropy4: what name service are you using ? [23:31] my internal ip is 10.0.0.5 [23:32] snake: is that the IP you are giving to your friend ? [23:32] no. i am giving him my external IP [23:32] snake: ok - so you need to get support for your router to find out how to do NAT on it, [23:32] snake: also make sure your router is not running a firewall [23:32] snake: while you are degugging disable the firewall on your ubuntu machine [23:32] ok [23:33] ikonia not sure if i understand what you mean. the linux box is pointing to my isp's dns server (in /etc/resolv.conf) is there more setup i need to do with /etc/hosts ? [23:34] entropy4: how does your ubunti machine know the ip/hostname mapping of your windows machines ? [23:36] ikonia sorry actually my router is functioning as a caching dns server.. i forgot [23:37] i think when hosts get an ip from the dhcp (from the routeR) it registers them in dns [23:37] entropy4: ok, so the ubuntu machine is probably not announcing it's self [23:37] so that might explain that [23:39] ok i found the NAT section in my router settings. all i have to do is tell it to switch to "Open" right? [23:39] snake: I don't know - thats up to you to configure your router [23:40] ikonia i guess one way around this would just be to setup static dhcp for my server [23:40] hmm.. alright thanks for the help anyway. [23:40] i will try it [23:40] entropy4: simple work around [23:40] entropy4: you could also configure dhclient to do a dns update, but I don't know what your router expects [23:41] me neither .. its a linksys but its running tomato firmware === dannf is now known as dannf_afk [23:58] hmm, still couldnt get it to work after doing static dhcp, then i found some post on ubuntu forums that said to install samba... i did and now i can ping it by name