[00:40] <_Techie_> is there a group a user has to be in to allow them to listen on the network? [00:41] Hmm.. Hello all [00:41] Seems my server is.. slightly unwell [00:41] "Gave up waiting for root device. " blah blha.. [00:41] ALERT! /dev/mapper/tank-root does not exist [00:43] Whats odd is this is a RAID1 mirrored over 3 disks [00:46] Slyboots: nothing wrong with that [00:47] Slyboots: not playing with zfs, are you? [00:50] no? [00:50] its been running for months.. powered it down.. powered it back up [00:50] AAHGHGHGHGHHGHGHGHGHGHG!! [00:50] And so on [00:52] Slyboots: no hardware failure? [00:54] I thought a disk was dead but.. [00:54] SMART shows OK across the board [00:56] New bug: #699665 in openssh (main) "sshd crashed during a rsync" [Undecided,New] https://launchpad.net/bugs/699665 [01:00] Slyboots: is that a "fake" raid controller ? [01:00] Linux softraid; MDADM [01:01] with some other layer on top or below then? [01:01] LVM [01:01] Slyboots: sorry, I mean that a three-way RAID1 is not odd [01:02] The BIOS shows the SMART status of each disk.. comes up OK [01:02] Im wondering if a kernel update might have hosed the modules or something [01:03] Slyboots: what's the status of the software raid device ? [01:03] No idea; trying to create a live bootable disk to go in and check [01:03] right now as soon as it hits grub it bombs out witht "Gave up waiting for root device" [01:03] heh? [01:04] so it's grub that complains? [01:04] I get a black screen with a flashing "_" under it [01:04] Then after about 30 seconds [01:04] "Gave up waiting for root device" [01:04] ALERT! /dev/mapper/tank-root does not exist [01:06] that's inside the initrd, so you should be able to see if the raid is assembled or not? [01:06] Not sure.. basiclylafter that it goes (initrtfs) and .. [01:06] Well nothing; tapping keys on the keyboard does nothing [01:07] hm, that's not very useful indeed ☺ [01:07] try booting an older kernel? === Tankenator_afk is now known as Tankenator [01:07] Dont have any.. least I dont think so [01:07] I dont evne really get a grub prompt [01:07] (Although i never clearn out /boot.. so it should be there in theory [01:08] you can enter the grub menu with Esc (grub1) or Shift (grub2) [01:09] xMm.. [01:09] "Grub loading.." [01:09] Then it just continues on [01:10] just keep Esc or Shift (depending on grub version) pressed down during boot [01:10] Tried that... [01:10] Does Caps Lock also work, as in extlinux? [01:10] That way at you can just push it and leave it down [01:10] Its a wireless keyboard.. [01:11] Doesnt have indicator lights for caps lock [01:11] Slyboots: then it probably won't work at all [01:11] I doubt that grub has a bluetooth driver [01:11] Its wifi through a usb interface [01:11] I guess that might work if grub has a USB driver [01:11] I doubt it's WiFi ;) [01:12] OK; got grub using another keybarod :P [01:13] but an USB keyboard might need a BIOS setting [01:13] Its RF or something; but thats outside the fact.. grabed another keyboard nad loaded up the last good kernel [01:13] Well its doing something; ubuntu Usplash has come up [01:13] more likely plymouth [01:14] twb: depends how old the server install is ;) [01:15] Okay; got fed up waiting for usplash to vanish and logged in via the cli [01:15] I guess if he sees "grub loading" than it isn't that stupid new "0 wait time" one [01:15] Not sure how to continue from here though [01:15] Should I.. fsck the disks or..? [01:16] Slyboots: what was the problem again? [01:16] Well it was going "AGGH! Root not found" [01:16] Loading an old kernel fixed that.. [01:17] But now I got run-parts: /etc/update-motd.d/90-updates-availabel exited with an return code 2 [01:17] Everal times [01:19] Mm.. brb. hold on [01:22] Mm.. [01:23] Right [01:23] Well back into the server.. not 100% if everything is OK though [01:24] First things first should be to remove that "bad" kernel from grub.lst right? [01:30] not sure where grub.lst is though [01:33] Slyboots_: /boot/grub [01:33] Slyboots_: just remove that kernel, even [01:34] Mmm.. using apt-get ? [01:34] ALthough if it were me, I'd probably try to work out WHY it wasn't working [01:34] Slyboots_: yes [01:35] Seems like its not loading LVM or something [01:35] And since its 2am in the morning I just want to fix it and go to bed [01:35] or play Dawn of war.. minions of Chaos to kill and whatnot [01:39] Mm.. [01:39] Okay; there is one marked -virtual [01:39] Not sure why that is there [01:43] Slyboots_: I don't know what you're looking at. [01:44] Grubs list of kernels [01:44] I've removed one marked "virtual" see if that makes any difference [01:45] I believe the "virtual" flavoured kernels have many features removed because they are intended only for use as guest OSes in VMs [01:45] Aye; Thats what Im thinking [01:45] Im not sure what its on my server at all but I've removed it and updated grub; a reboot will tell [02:04] I'm running Ubuntu 10.4 and installed BIND using this howto: https://help.ubuntu.com/community/BIND9ServerHowto [02:04] I keep getting these messages in my log: Jan 6 18:46:59 jupiter kernel: [1462769.470917] type=1503 audit(1294361219.337:72660): operation="open" pid=13550 parent=1 profile="/usr/sbin/named" requested_mask="ac::" denied_mask="ac::" fsuid=105 ouid=105 name="/var/log/query.log" [02:06] Any thoughts? [02:06] My home machines all use that machine for DNS and when my wife goes to facebook, BIND is not resolving some of the servers that host facebook's css files and a few other things. [02:13] One of the DNS names that aren't resolving is b.static.ak.fbcdn.net. When I do an nslookup on that address, I get some of those kernel messages in my logs. I'm hoping this is the issue. === loadedwithwyrms is now known as loadedwith|work [04:15] kaje, looks like apparmor is denying it access to that file [04:16] yeah, I got it worked out in the bind channel. Thanks for the help though. [04:16] ah, ok === hikeonpast_ is now known as hikeonpast [06:26] what command do i use to see how many cores a vps has available to it. i am using ubuntu server 10.04 [06:32] cat /proc/cpuinfo [06:40] Depending on the virtualization technology, that won't tell you how many your VPS is allowed to use [06:57] New bug: #695985 in mysql-5.1 (main) "/etc/mysql/debian-start exposes debian-sys-maint users password to any users on the box via ps(1)" [Medium,Confirmed] https://launchpad.net/bugs/695985 === guampa is now known as guampabed [07:01] Is there a way to gete byobu to save the tabs you have open over re-boots? I have a shell, root shell and media server running and owuld like their tabs to be kept open over re-boot [07:18] In logrotate.conf, is i.e /var/log/myapp/* a valid descriptor, when myapp contains subfolders, that contain the logfiles? [07:42] martinjh99: screen cannot save state between reboots. [07:43] milligan: you could configure .screenrc to *launch* programs whenever it starts. [07:43] i think you mean me - How would you do that then? Point me to some docs? [07:46] Sorry, yes [07:46] martinjh99: say you wanted to run bash and top and "w3m google.com". [07:46] martinjh99: you would put in three lines like "screen"; "screen top" and "screen w3m google.com" [07:46] martinjh99: in your .screenrc, I mean [07:47] wouldn't that run 3 instances of screen though? [07:47] No [07:47] .screenrc takes screen commands, not sh commands. [07:48] hmm ok [07:48] In any case, the sh command "screen foo" will create a new tab if it is run within screen (i.e. if $STY is set) [07:48] It is unfortunately an extremely confusing arrangement for newbies [07:51] certainly sounds like it and why when i google "configuring screen" i get a load of windows pages ;) [07:55] Your best references are 1) the manpage/infopage; and 2) #screen channel on Freenode [07:56] You will have better luck (but still not good) in google by using `"GNU Screen"' rather than `screen'. === RudyValencia- is now known as RudyValencia [08:01] using /bin/screen instead of screen might help too [08:01] Interesting idea [08:04] ah ok thanks - just searching ubuntu forums to see if they have anything there [08:04] Pfft [08:04] web fora are just people too stupid to use usenet [08:05] Rather, too stupid to configure a newsreader [08:06] or a mail client ;) [08:08] hehe well i found something I can use via Google... Thanks all [08:10] JanC: I suppose so, although I prefer newsreaders for reading mailing lists :P [08:25] New bug: #699737 in autofs5 (main) "automount[1275]: syntax error in nsswitch config near [ syntax error ]" [Undecided,New] https://launchpad.net/bugs/699737 [09:23] hi guys... how can i view my raid setup on an ubuntu server [09:23] twister004, hardware or software raid? [09:23] hardware [09:24] that depends on the kind of raid you have then [09:24] what is it? [09:24] it's RAID1 [09:24] no I mean the brand [09:24] the controller [09:24] I'm used to array-info for smartarray (hp/compaq) for example [09:25] 82801G(ICH7 Family) IDE Controller [09:25] that's the adapter model [09:26] hmmm, google tells me this is an audio controller [09:26] it is? [09:26] :S [09:26] :D... my bad [09:26] twister004: lspci -nn | grep IDE [09:26] or maybe there's two products with the same name [09:26] ICH7 is just a southbridge [09:27] Probably you've got an ICH7R or so [09:27] sorry... it's the N10/ICH7 Family SATA IDE CTRLLER [09:27] ok [09:27] IOW fakeraid [09:27] twister004: stick to md RAID [09:28] ill keep that in mind.. right now, in the "Disk Utility"... i see two HDDs under the Adapter.. are these the two RAIDed HDDs? [09:28] Who cares? [09:29] or will I be able to see only one of the RAIDed hdd? [09:29] I want to know if there's a third hdd [09:29] twister004: why are you using a GUI [09:29] twister004: open the case, and count the drives, then [09:29] :D [09:29] im remotely located [09:29] Blergh [09:29] pastebin the contents of /proc/mdstat and /proc/partitions [09:30] partitions?.. is there a command called partitions? [09:30] i cant find it [09:30] hrmm [09:30] these are files... [09:31] just cat them, or use pastebinit to do it faster ;-) [09:32] oh.. here you go http://pastebin.com/jkR4HyJT [09:33] twister004: you have two disks set up with per-partition RAID1 arrays [09:33] sory.. it's software raid [09:33] ok... but what about the third.. i remember it was in there [09:34] There's no third disk, as far as linux sees [09:34] There are three ARRAYS [09:34] is there something like a "devfsadm" for linux? [09:34] I do not know what that is. [09:34] The tool to manage md arrays is mdadm. [09:34] it's probing for devices(newly connected) [09:35] under solaris [09:35] udev is infrastructure that responds to device events (including them being connected) [09:36] e.g. you can tell it "whenever the USB mass-storage device with the serial number XXX is connected, mount it and start backing up /srv to it" [09:36] But more generally it creates device files in /dev, and a bunch of desktop wankiness like mounting stuff in /media [09:36] twb.. ok [09:36] thanks [09:37] looks like there is no 3rd disk [09:37] wierd [09:37] i clearly remember there was a 3rd [09:40] Maybe it's not cabled properly === eremyja___ is now known as eremyja [09:59] twb.. yeah.. looks like that's the problem.. ill have to go onsite [09:59] thanks for all your help and advice! [10:06] Well, I AM a genius [10:16] http://xkcd.org/844/ :) [11:10] win 17 [11:10] sorry [11:28] Hey, I wanted to outsource my web-server so I created a virtual machine. Now I do not want to install Webmin, phpmyadmin again. Is there a software I can use to manage (add domain/add database) the virtual machine from the host? [11:32] gobbe: that's not out yet [11:33] Frenk: ssh [12:06] s === guampabed is now known as guampa === doko_ is now known as doko [12:42] how do i install http://pastebin.com/7u9AzMEh php extensions? i knew 2 so mentioned their packages, what about rest? [13:18] I want to otain a SSL certificate but I read that if it is password protected I need to enter the password each time a service is restarted. I have a monitoring solution (monit) which restarts the services if something happens. Is it very insecure to have a SSL-cert without password or is it easier to configure monit with the SSL-password? [13:27] Frenk: i don't think you can decrypt a certificate with a monitoring program [13:28] Frenk: it is standard to not encrypt certificates on services that need to come up unattended [13:30] I mean the watchdog needs to start the service somehow ... even if ssl-cert is password protected... | Okay I try to set it up without password - but "The CSR key must have a length of 2048 bit" has nothing to do with wether its encrypted or not? [13:31] Frenk: no [13:34] having a password on the cert is nice, if you think your box will be rooted [13:35] not having a password is fine, as long as your not rooted [13:35] then it can still be *ok* (depending on your definition and level of ok), if you revoke the cert, assuming you know you where rooted [13:38] how do i install these: http://pastebin.com/7u9AzMEh php extensions on ubuntu karmic? i have listed the two known ones, don't know packages for rest, tried apt-cache search name-here | grep php but no results === oubiwann is now known as oubiwann_ [13:40] patdk-wk: pmatulis thx [13:48] how do i install these: http://pastebin.com/7u9AzMEh php extensions on ubuntu karmic? i have listed the two known ones, don't know packages for rest, tried apt-cache search name-here | grep php but no results [13:49] I have postfix installed for sending e-mail. Does my reverse dns has to be mail.domain.com or can I just use domain.com? [13:52] Frenk: the name that your MTA will expose to the internet should be both forward and reverse resolvable [13:53] Frenk: forward is more important but some MTAs may refuse your mail if reverse is missing [13:55] Frenk: but of course your actual domain name needs to be (at least forward) resolved as well [13:55] Ill check it =) [13:57] how do i install these: http://pastebin.com/7u9AzMEh php extensions on ubuntu karmic? i have listed the two known ones, don't know packages for rest, tried apt-cache search name-here | grep php but no results [13:58] And I have a strange thing happening pmatulis, I have cyrus and denyhosts, every time denyhosts bans any ip the permission on /etc/hosts.deny are changed and cyrus cant read the file = refuses all connections. I dindt find any permission settings in deny-hosts config. === Tankenator is now known as Tankenator_walk [14:02] Error404NotFound, they are installed by default in php-common (I think) [14:02] patdk-wk: pcre is, not sure abour json, couldn't find its config in php.ini [14:03] heh, fail [14:03] check with phpinfo [14:03] there are no config options for json, so it won't be in php.ini [14:06] patdk-wk: :P [14:06] thanks :) [14:10] Hi. I followed a broken tutorial online for setting up Drupal with a LAMP stack on Ubuntu and I think I made some bad changes to my /etc/apache2/httpd.conf file. I'm wondering how I can either restore it to the original settings or reinstall apache with the original .conf file [14:13] Daviey: can you have a look at 697753 its pretty simple [14:17] Arcitens: use at your own peril, also, check the path to the deb: 'dpkg --force-confnew -i /var/cache/apt/archive/apache???.deb' [14:17] 'archives' [14:19] zul: so re nc fix, will users of previous releases be burned in any way? [14:19] pmatulis: as in? [14:20] zul: well, as i understand it, change to nc was to enable simultaneous connections to a libvirt session, will continue to be like this? [14:20] zul: w/o extra hoops? [14:20] pmatulis: the "-q" functionality was put back but with a warning message [14:21] zul: i guess i don't grok the issue, i thought the 'q' thing was the change we put in [14:21] zul: that caused the bug [14:21] New bug: #699845 in php5 (main) "php5 affected by http://bugs.php.net/53632" [Undecided,New] https://launchpad.net/bugs/699845 [14:25] pmatulis: right it is...but its very tied into libvirt, users can use netcat-traditional if they need regular functionality [14:28] zul: but what if they need both libvirt *and* traditional functionality? [14:28] pmatulis: good point lemme think about it [14:28] pmatulis: then they use nc.traditional [14:28] mdeslaur: o_0 [14:29] pmatulis: oh, zul updated the patch to the one debian just changed....so nc-openbsd now behaves like the upstream one [14:29] ie: behaviour is the same if no -q is given [14:29] mdeslaur: ok, that makes sense then [14:30] but the default behaviour between nc-traditional and nc-openbsd is different, and that has nothing to do with the -q patch [14:30] pmatulis: Sorry, I went afk. And I'm not sure what exactly you're suggesting I do there, sorry. [14:30] fedora uses nc-openbsd by default, and we do too now...so if anything, we'll be consistent [14:30] Arcitens: it's a command [14:31] pmatulis: and I should replace ??? with apache version? [14:31] Arcitens: to the real path of the package that provides the conf file your're talking about [14:31] pmatulis: ah, ok. thanks. [14:31] pmatulis: It wasn't to enable simultaneous connections to libvirt. It was to enable more than one *ever*. If you had connected to it with nc.traditional once, you could never, ever connect to it again. [14:32] Arcitens: to be sure: 'dpkg -S /etc/apache2/httpd.conf' [14:32] soren: ah [14:33] hi soren! [14:33] pmatulis: says not found when I ran the second command. (Sorry I'm being a total noob here. I appreciate the help.) [14:33] Arcitens: so the path you gave is not right [14:34] pmatulis: it must be. I'm staring at the file in nautilus and I can open it with 'gedit /etc/apache2/httpd.conf' [14:35] Arcitens: symlink? [14:35] pmatulis: nope [14:35] mdeslaur: dude. [14:35] mdeslaur: :) [14:36] :) [14:36] pmatulis: I'm fine just starting over completely. is there a way I can uninstall the whole apache-mysql-php stack including the config files and then reinstall and work with a fresh slate? [14:37] Arcitens: i'm not sure you can back out completely from a tasksel task [14:37] pmatulis: that's unfortunate :( [14:37] Arcitens: you'll need to google 'round. maybe view what the task does and remove individual packages [14:38] Arcitens: there's a file that explains the tasks [14:38] Arcitens: man tasksel [14:38] pmatulis: Hmm ok. what about just the apache part? Can I uninstall apache and purge the config files for that? I have a feeling it's apache I screwed up in. [14:38] Arcitens: sure [14:39] Arcitens: 'aptitude purge apache' should do it [14:39] pmatulis: thanks very much for the help. [14:39] Tasks don't "do" anything. [14:39] They're just a set of packages. [14:40] soren: possible to remove associated packages in one fell swoop? [14:41] I suppose "sudo apt-get --purge remove taskname^" should do it. [14:41] (Note the ^ at the end) [14:41] right, ok [14:41] is it best to use the VM image or ubuntu server for vm? [14:41] AndyGraybeal: say what? [14:41] aaah nevermind [14:41] AndyGraybeal: there is no more JEOS version available [14:42] f4 on install and it pops up for vm install! [14:42] i'm a bit behind the times. [14:42] aah, you mean that [14:42] i have ubuntu 10.04 server install and i hit f4 on what type of install i want to do .. and i see the virtual machine choice now, it didnt show up earlier becaues i was doing something wrong.! [14:43] thanks gobbe and pmatulis [14:43] :) [14:43] this is the recommended way to start out with a VM correcT? [14:44] AndyGraybeal: you want to create a KVM guest? [14:45] pmatulis: yes [14:45] AndyGraybeal: use virt-manager is you are just beginning [14:45] s/is/if [14:45] oh no no, i've made many images with virt-install but i have used the 'normal' install frm ubuntu server disc [14:46] i just learned yesterday aboutthe VM install! [14:46] i learned kvm about 2 years ago with 8.10 [14:47] well learned enough to make it from point a to point b... throw a ringer in the mix and i'm confused.. but from point a to b and i'm fine. [14:47] AndyGraybeal: i never used a VM install explicitly in the installer. i assume it just makes a minimal install + the virtual kernel [14:47] i wouldn't know how to do that manually, so this is great :) [14:47] AndyGraybeal: comparable to what vmbuilder does [14:48] i haven''t used vmbuilder since 8.10 [14:48] AndyGraybeal: so what's your question then? [14:48] i was asking about how to get to the virtual install; i hit f4 on the wrong screen, i eventually found it and answered my own question. [14:49] pmatulis: Well, google wins in the ned. All I had to do was add 'ServerName localhost:80' to 'httpd.conf' Now it seems to be working fine... Feel like a bit of a fool, but hey, learning experience, right? Thanks again for the help. [14:49] pmatulis: sory for keepihng yuo hanging. [14:51] good, looks like everybody is happy now === Tankenator_walk is now known as Tankenator [14:57] :) [15:01] New bug: #699855 in autofs5 (main) "autofs.schema in wrong location" [Undecided,New] https://launchpad.net/bugs/699855 [15:09] zul: I will ping mathiaz when he pops up re. uec-testing-scripts-devs (adding you) [15:10] (just noted you are pending there) [15:11] hggdh: thanks [15:13] the preseed value for d-i for partman-md/device_remove_md doesn't actually remove md [15:13] the install fails if md devices exist [15:13] so i have to mdadm -S /dev/md* && mdadm --zero-superblock /dev/sd* before i start the installer [15:14] and that can't be automated via early_command in a nice way, the mdadm udeb isn't loaded yet, and even if it was it can exit non zero eevn if the intended operatuion completes successfully on some of the nodes that * expands to [15:14] jdstrand: just to be sure - have all your bad kvm tests been using a snapshotted qcow2 guest? [15:14] this is 10.04 [15:15] also it seems the partition priority and sizxing doesn't behave as expected when doing lvm over raid [15:21] hallyn: a snapshot was always involved, yes. in the 'upgrade one/install another' test (comment #12), only the upgraded one was snapshotted as the other was still being bootstrapped [15:21] may have discovered how to fix the priority weirdness, apparently the order is dependent in this scenario [15:21] maybe the device_remove_md is order-dependent? anyone using this preseed value with success? [15:22] hallyn: also, I almost always was using a mix of i386 and amd64 installs, but I can't so with certainty I always did (and therefore can't say it is 'ok' with just one or the other) [15:22] hello! Could someone tell me what this means: Jan 7 15:14:32 leibniz kernel: [11819455.470672] type=1505 audit(1294413272.487:208): operation="profile_load" pid=19195 name="libvirt-4e12e041-2ec2-587c-4655-8c51167c15cb" [15:22] hallyn: and by 'installs', I mean 'guest installs' [15:23] yann2: apparmor is preventing [15:23] yann2: libvirt uses apparmor to confine virtual machines [15:23] gobbe: no, it isn't [15:23] ah, sorry [15:23] I just stopped apparmor [15:23] well, apparmor yes, not preventing :) [15:23] yann2, gobbe: that line is just telling you that the profile loaded [15:23] i didn't read enough good [15:23] damn I stopped whole apparmor before :( [15:24] yann2: no reason to stop apparmor for that line-- if apparmor is causing a problem, it will log a denial [15:24] and then you can tune apparmor [15:24] infact apparmor is working quite well [15:25] k, is a dev server anyway [15:25] profile_load and profile_remove are purely informational, and letting you know that everything is working properly [15:25] if you compare to selinux which might be pain in the ass [15:25] Jan 7 15:14:32 leibniz kernel: [11819455.470672] type=1505 audit(1294413272.487:208): operation="profile_load" pid=19195 name="libvirt-4e12e041-2ec2-587c-4655-8c51167c15cb" [15:25] arg [15:26] Jan 7 15:05:14 leibniz kernel: [11818898.104525] type=1503 audit(1294412714.157:188): operation="open" pid=18580 parent=1603 profile="/usr/lib/libvirt/virt-aa-helper" requested_mask="r::" denied_mask="r::" fsuid=0 ouid=0 name="/var/lib/kvm/oasouth-itadmin/root.qcow2" thats the one that got me concerned [15:26] yann2: virt-aa-helper denials are not necessarily fatal [15:26] yann2: look in /etc/apparmor.d/libvirt/libvirt-4e12e041-2ec2-587c-4655-8c51167c15cb.files [15:27] yann2: if it has /var/lib/kvm/oasouth-itadmin/root.qcow2, you are ok [15:27] yann2: virt-aa-helper is what generates the dynamic profile, tailored for your vm [15:28] jdstrand: btw i'm trying to use the security team docs examples to vm-clone etc - but when i vm-clone it starts the new machine but doesn't manage to connect t it over ssh to do the updates it wants. Is this known (.e. some script needs to add '.' to the hostname or something)? [15:28] mmmh could it be that apparmor is making issues if I put my VM in tmpfs? [15:28] yann2: it would show denials in the log [15:28] i know selinux could [15:28] i don't know much about AA but yea it would be in the audit logs [15:29] yann2: /var/lib/kvm/oasouth-itadmin/root.qcow2 should be allowed though, because of this line in virt-aa-helper's profile: [15:29] /**.qcow{,2} r, [15:29] yann2: is /var/lib/kvm/oasouth-itadmin/root.qcow2 a symlink? [15:29] nope, thats where I put my vms [15:30] the /var/lib/kvm/oasouth-itadmin/ is a tmpfs though [15:30] hallyn: I haven't used vm-clone in ages [15:30] when running the vm from disk instead of tmpfs it doesnt freeze anymore though :) [15:30] yann2: what are the contents of /etc/apparmor.d/libvirt/libvirt-4e12e041-2ec2-587c-4655-8c51167c15cb* [15:31] hallyn: vm-clone doesn't do snapshots [15:31] jdstrand: oh. ok. well, i'll keep trying, and will try a i386 one [15:31] hallyn: (which is why I don't use it anymore) [15:31] what do you use then? [15:31] just vm-start -s? [15:31] hallyn: that page has a section down below for using snapshots [15:32] kthx [15:32] hallyn: you looked at 'Cloned virtual machines'. I use 'Snapshotted virtual machines' [15:32] jdstrand, http://pastealacon.com/26508 [15:33] hallyn: someone from the team may use vm-clone-- perhaps kees or mdeslaur, I'm not sure. I used to, but don't any more [15:33] hallyn: I'm trying to debug that exact problem as we speak :) [15:33] I tried to use vm-clone this morning, and it's not working [15:34] mdeslaur: does vm-clone use nc? [15:34] jdstrand: I looked at that, but that's not the issue [15:34] hallyn: you're on natty, right? [15:35] yann2: from an apparmor perspective, it looks ok. I suggest looking in /var/log/libvirt/ [15:35] jdstrand: it's not mounting the images successfully I think...anyway, I'm still poking at it [15:35] yann2: if you want, you can add: [15:35] thats where I started :) [15:35] /var/lib/kvm/** r, [15:36] jdstrand, do you think that having /var/lib/kvm/oasouth-itadmin mounted as tmpfs changes anything for apparmor, than having it not? [15:36] to /etc/apparmor.d/usr.lib.libvirt.virt-aa-helper and reload with 'sudo apaprmor_parser -r /etc/apparmor.d/usr.lib.libvirt.virt-aa-helper' [15:36] for some reasons it doesnt work when mounted as tmpfs. It does work when not, but it's dog slow :( [15:37] yann2: but that won't fix your issue-- you can see the dynamic profile is correctly generated [15:37] yann2: having as tmpfs should make no difference [15:37] jdstrand, my real issue is that its a windows vm - and it freezes after a few seconds now [15:37] works fine without tmpfs, but am installing service pack on windows [15:38] yann2: what version of ubuntu is the host? [15:38] 10.4 [15:39] yann2: so, if you really think it is apparmor, you can sudo apparmor_parser -R /etc/apparmor.d/usr.sbin.libvirtd, then stop and start libvirt [15:40] I got no idea if its apparmor, its the only indication I could find in the logs :) [15:40] yann2: do you have enough ram? iirc a tmpfs can only take 50% of your ram. so that is half your ram for the disk, and then your vm still needs ram for itself [15:40] oh that would explain yes [15:40] host has 8GB and vm is 3.9GB :) host still had some ram left though [15:41] I ll try again on a host with more ram - thanks a lot for the info, thats probably it [15:41] sure [15:47] jdstrand: btw, as i was reading the securityteam/testenvironment page yesterday I *was* wondering why i'd want to first clone, then create a runtime snapshot :). oops [15:49] heh [15:49]