=== markthomas|away is now known as markthomas === med_ is now known as Guest471 === Lcawte|Away is now known as Lcawte === Lcawte is now known as Lcawte|Away === med_ is now known as Guest61394 [07:29] Good morning. === ochoroch1 is now known as ochoroch === [1]badabum18 is now known as badabum18 [10:38] zul, jdstrand: is it possible to put the libvirt templates generates for instances into complain mode? trying to generate some updates for using hugepage memory === kickinz1 is now known as kickinz1|lunch [12:20] jamespage: yes but i forgot how === kickinz1|lunch is now known as kickinz1 [13:12] jamespage: yes, modify /etc/apparmor.d/libvirt/TEMPLATE.qemu to have "profile LIBVIRT_TEMPLATE flags=(complain) {" [13:12] jdstrand, awesome - thankyou [13:26] np [13:37] jdstrand, I've probably learn't and forgetten this already but is there a nice way to parse the complain entires into an profile update? [13:51] jamespage: there is a program called aa-logprof, yes, but it isn't smart enough to know to update the template and it may have trouble finding the vm profile in /etc/apparmor.d/libvirt [13:51] jamespage: if you were keen on doing that, you could try passing the '-d' directive to logprof, then do a diff between the template and the profile [13:52] but, handcrafted in this case is probably faster [13:52] jdstrand, ack [13:53] jdstrand, and do we have a nice mechanism for providing local profile updates for the generated profiles? [13:53] I was trying to figure it out but got stuck... [13:55] jamespage: if you want all VMs to pick up the change, you can modify /etc/apparmor.d/abstractions/libvirt-qemu. if you want only a single vm to have the change, you can modify /etc/apparmor.d/libvirt/libvirt- (note, *not* the one with the .files extension) [13:55] jdstrand, ok - ta [13:55] np [13:56] jamespage: I think you're set now, but fyi, http://wiki.apparmor.net/index.php/Libvirt [13:56] I see it needs a couple updates [13:57] * jdstrand does so [14:20] jdstrand, how does that look - http://paste.ubuntu.com/12071349/ ? [14:20] I suspect the top two rules are generically applicable - the third is dependency on where hugepages gets mounted [14:22] the rules are written well for the denials. however, I think they are too lenient [14:23] so, qemu only complained about the dir for cpu, so this might be better: [14:24] /sys/devices/system/cpu/ r, [14:24] I think this would be better for node: [14:24] /sys/devices/system/node/ r, [14:25] /sys/devices/system/node/node[0-9]*/** r, [14:25] actually, even better [14:25] /sys/devices/system/node/node[0-9]*/meminfo r, [14:26] as for the hugepages mount point, there is already this rule in libvirt-qemu: [14:26] owner "/run/hugepages/kvm/libvirt/qemu/**" rw, [14:26] it that not enough? (I feel there was a bug on this) [14:27] https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1001584 [14:27] Launchpad bug 1001584 in libvirt (Ubuntu) "libvirt-bin support for hugepages: apparmor needs a rule for hugetlbfs" [Medium,Expired] [14:28] https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1250216 [14:28] Launchpad bug 1250216 in libvirt (Ubuntu Saucy) "apparmor policy prevents using hugepages" [High,Fix released] [14:29] that last bug suggests that virt-aa-helper should be updated to generate the paths. is the mountpoint expressed in the domain xml? [14:29] jamespage: ^^ [14:29] jdstrand, reading [14:32] jamespage: here is a cleaned up summary of the above: http://paste.ubuntu.com/12071434/ [14:35] jdstrand, thanks [14:36] jdstrand, so for the owner "/run/hugepages/kvm/libvirt/qemu/**" rw, bit, I just need to make sure that hugepages is being mounted in the correct location I think [14:36] right, that is what I was thinking [14:36] # /run/hugepages/kvm/ [14:38] I don't know what the domain xml looks liek for hugepages, but if it says where they are mounted, then virt-aa-helper (the thing that stitches together the template, the uuid and uuid.files files) can be made to add the paths to the .files file automatically [14:38] if it isn't in the domain xml, it would require a bit more work I think [18:00] so someh ow I created a file called \ [18:00] how do I delete it [18:00] when I put it in rm it just says missing operand [18:01] n.m [18:01] rm \\ [18:01] :) [18:35] whats the new package for dhcp server...i tried sudo apt-get install isc-dhcp-server on ubuntu server 15.04 but says cant find package..also tried sudo apt-cache search dhcp didnt find anything about server [18:41] fellayaboy: http://packages.ubuntu.com/vivid/isc-dhcp-server [18:41] your sources must not be correct [18:41] think i need to do apt-get update [18:42] let me also check apt/sources.list [18:45] thanks qman that solved it [19:10] Howdy :-) - I want to install syslog-ng on my servers using puppet....Now I see that syslog-ng has a open bug with this since ano 2006 - cannot install syslog-ng because it has unmet dependencies ( syslog-ng : Depends: syslog-ng-core (>= 3.5.3), ....) [19:10] I can install syslog-ng by first installing syslog-ng-core and then install syslog-ng. But thats not easy in puppet.... [19:12] Danskmand: how are you trying to install it? apt-get install ought to handle all the deps for you without trouble. [19:14] sarnold: Yeah, I know - but running "apt-get install syslog-ng" results in "The following packages have unmet dependencies:" [19:14] " [19:16] "syslog-ng : Depends: syslog-ng-core (>= 3.5.3)" [19:16] (3 more depends, 5 recommends) [19:16] e: Unable to correct problems, you have held broken packages [19:17] that's interesting, never seen that before [19:17] try this: dpkg --get-selections | grep -v install [19:19] When I then do a "apt-get install syslog-ng-core" it installs that and when I then repeat the "apt-get install syslog-ng" it installs without a problem ! [19:19] So it is really that syslog-ng-core ,,,, [19:20] The .v results into - nothing - [19:22] hmm. I don't have any held packages on my systems, I assumed the 'install' would go away if it were held.. maybe try dpkg --get-selections | grep held or | grep hold ? [19:23] Yeah, - nothing - [19:30] OKay....have installed a new ubuntu-Server as a vm @ home and run "apt-get install syslog-ng" - there it will install is without any problems ! - the new system is on 15.04 - the problem-server is on 14.04 LTS.... [19:30] I suspect a brand-new 14.04 would also work fine, there's something strange on your existing system.. [19:36] I see theres a 14.04.3 release - how can I find out the release I am running down to the "3"-number ? [19:37] Forget it - I overread it - I have a 14.04.1 release running... [19:38] How do I upgrade to 14.04.3 release ? [19:39] sudo apt update && sudo apt full-upgrade [19:39] Danskmand: full details here https://wiki.ubuntu.com/Kernel/LTSEnablementStack [19:40] I just dont wann upgrade away from the LTS-release... [19:41] the 14.04.1 kernel will be supported for the full five years; if you upgrade to 14.04.3 you'll need to upgrade to the 14.04.5 kernel in august 2016. [19:44] Hmm....okay, so upgrading is not the solution if I dont wanna loose the LTS and wanna solve my problem.... [19:45] just keep running apt-get update && apt-get -u dist-upgrade on a regular basis === zerick_ is now known as zerick [20:03] Running the apt-get dist-upgrade looks to me like I will leave the LTS-path.... [20:03] Which I cannot... [20:06] the usual 'upgrade' command refuses to install new packages, which means e.g. kernel updates never happen. dist-upgrade can install new packages, so you'll get new kernel packages. it only changes the actual distribution if you modify the 'trusty' to 'vivid' or 'wily' in your /etc/apt/sources* files === markthomas is now known as markthomas|away [20:12] Okay...running dist-upgrade right now.... [20:12] Lets hope we have a server afterwards.... [20:13] We have a problem with our servers hdd-controller of some kind... [20:13] you should, i've run apt-get dist-upgrade a few thousand times in 15 years and only had two or three problems, back in the days when I ran debian unstable, and that kind of thing was expected. :) [20:14] Yeah, and I can still choose the old kernel at boot... [20:16] Its a known problewm with the hp....380 Servers... [20:26] Ooops ! - No more system I think ! [20:28] "Gave up waiting for root device. common problems: ......... === medz is now known as samsn [20:29] ALERT! /dev/disk/by-uuid/add.........many numbers does not exist. Dropping to a shell! [20:29] samold: [20:30] Danskmand: did you need to build an out-of-tree kernel module to mount your storage array? [20:30] NOw my prompt is "initramfs"..... [20:30] Hmmm...dont know - wasnt there when it was built..... [20:31] Danskmand: maybe you need to rebuild your initramfs with a specific kernel module to allow mounting the array? [20:31] I am running on a Busybox right now... [20:32] Doing a reset right now.... [20:33] HP proliant... [20:35] Smart array b320 i controller... [20:37] ata6: ATA4: Sata link down (SStatus 0 SControl 300) [20:38] Gave up waiting for root device. [20:38] So now I think I know what I should not have done to that server !!!!!! [20:39] Sh**.....Can you help me on this ? [20:39] Please ? [20:39] you need to install kernel updates; there's just no way around that. You've got to figure out how to make that system reliable enough to reboot whenever you need with a new kernel. [20:41] So in other words, the system is wrecked now - I need to reinstall .... [20:41] No way back, right ? [20:41] i'm sure there's a way forward [20:41] i've just never had one of those machines [20:42] there's a note that the dynamic smart array needs to be disabled on one specfic generation, but I don't know if that's something that would affect you or not http://www.ubuntu.com/certification/hardware/201401-14512/ === Guest61394 is now known as med_ [21:04] So....Now I booted from the old kernel....and it works.... === markthomas|away is now known as markthomas [21:45] having trouble with networking in an ubuntu server vm. I just added a new eth adapter, the system is detecting it but I can't bring it up [21:45] http://puu.sh/jAxgQ/619c45f8e6.png this is some of the output from lshw -C network [21:56] ...problem solved [22:14] Samold: You still here ? [22:16] Of course - I have just that one server that has this problem with the smart-array controller .... [22:17] A HP proliant DL360e gen8....