[00:22] Hello everybody! [00:23] I need a little help with Ubuntu Server 17.04 setting Bind9 up to work with Apparmor if anyone can spare a few mins pls msg me [00:24] are you getting any DENIED messages in your dmesg? [00:25] well I want to setup bind9 to work with apparmor but I have no clue on how to tell bind which directories to use [00:26] just configure bind however you normally would [00:41] cpaelzer: sigh, it took all day, but i think i finally got our snap to use pristine-tar and gbp (from upstream) only (not using the system one). I just pushed to master so it should build shortly. Meanwhile, i'm doing a virt-manager import from scratch on my VM at home now with basically nothing but git-ubuntu installed [00:45] rbasak: i'm starting to think, even though we're a classic snap, that our snap application wrapper scripts should not use any system paths at all (right now they prepend to the system paths, if set). Thoughts? [00:45] rbasak: our intention is we have all the deps in the snap, so that mostly allows workarounds. [00:45] rbasak: i'd test it in my vm first, as well, just to be sure we don't happen to need some system dependency that's not correctly expressed in the snap [00:45] rbasak: but right now, my VM doesn't even have git installed and it's working :) [01:07] rbasak: also, perhaps i should import your ssh key(s) to the bastion, so you have a snap-based place to import from? [05:33] good morning [05:34] nacc: sad to hear that it turned out so complex [05:34] cpaelzer: so virt-manager imported except patches-applied (in my vm) [05:35] cpaelzer: try it yourself, if you can, and you can use --skip-applied if need be [05:35] cpaelzer: i'll check in again tmrw [05:40] nacc: you didn't push that right? [05:40] nacc: a clone now did still point ubuntu-devel to 1.3.x [05:40] * cpaelzer importing locally [05:45] nacc: with and without the --skip-applied I fail at the verify step with ambiguous argument [05:46] let me ensure the snap is refreshed and/or run it in a VM/Container [05:59] nacc: the latest snap in artful with and without --skip-applied is failing on the ambiguous error [05:59] trying from git now [06:10] nacc: still the same from git [06:10] nacc: but please go to bed [06:11] nacc: I can provide a valid upload tag commiting onto the last upload-tag and ignore importing "for now" [06:11] nacc: I'll file a bug thou so we can keep track [06:39] Good morning === sammyg is now known as hakra [07:56] Hello! Can someone take a couple minutes and help me troubleshoot samba. Ubuntu Server 17.04 (base system tools, Budgie desktop, Bind9) [08:06] What is the problem? [09:18] Hi I need a help on the Libvirt Snapshot mangement [09:20] imraan: any more details [09:22] I am using the Command: virsh snapshot-create-as --domain MWIF_18 bkp2_2017-08-31_14.42.15.986080 --disk-only --atomic --no-metadata to create the snapshot To commit I am using the command: virsh blockcommit MWIF_18 vda --active --pivot --shallow --verbose --timeout 120 But sometimes i am observing the error like below what is the reason how do i recover: Error: 1. : block copy still active: disk 'vdb' alread [09:22] Good Morning, I've got a question wrt. Landscape (standalone) and MAAS. My aim is to use autopilot to deploy OpenStack. In my initial MAAS node commisoned nodes, i only had single nics. Landscape/Autopilot complained, so I hooked up one more network, recommissioned that node. However, Landscape/Autopilot did not detect the change. So I then removed the node, and started it from scratch, and commissioned it.. MAAS detected the new network automat [09:23] @cpaelzer I had posted my scenario [09:23] any help is greatly appreciated [09:25] FOR WORKAROUND: Currently using below: virsh blockjob MWIF_18 /mnt/alcm/packages/MWIF_100018/MWIF_18/MWIF_100018.bkp2_2017-08-31_14.42.15.986080 --abort virsh blockcommit MWIF_18 vda --active --pivot --shallow --verbose --timeout 120 But this also dont work all the time [09:26] @cpaelzer: Got my question? [09:26] yeah just a second [09:26] Okies..waiting [09:29] parlos: Your message got cut off at "MAAS detected the new network automat". [09:29] imraan: if you have long tests/log s to share use pastebinit [09:29] ... , but still L/A complains. It seems that L/A does not check/is not updated that the MAAS nodes have changed.. Any pointers? [09:29] and report the link [09:30] sry I meant parlos with the latter [09:30] okies [09:31] cpaelzer link to what? [09:33] parlos: if you have a long log/text you can do like "cat log | pastebinit" which gives you a link [09:33] Thnx [09:33] parlos: then on caht you can "hey I have a maas question ...link" [09:34] imraan: I assume that is on Xenial? [09:35] Hey, I've got a maas/landscape question: https://pastebin.com/kxvD67SJ [09:35] cpaelzer: I am using ubuntu release: 16.04 LTS [09:35] :) [09:35] yeah ok [09:35] imraan: once in the "bad" state does it resolve over time or is that job always considered active? [09:36] cpaelzer: it is always considered active. [09:37] imraan: is that a production environment or better "how much debug can you do"? [09:37] there were a few related simimlar issues [09:37] but so far all ended non reproducible enough to finally solve them [09:38] cpaelzer: It is not resolved over time.... After applying workaround only i am able to recover... [09:38] I have a few (now outdated but still testable) xenial libvirt builds, obviously testing on newer Ubuntu would be great - but that depends a lot on your case (how often it triggers and how mcuh you can do) [09:38] that is why I ask "how much" you can debug that [09:38] cpaelzer: In my environment i have 2 servers in one server it is continously reproducible [09:40] cpaelzer: Can u give me direction how to debug further...I am just struck not able to know where to check further [09:41] cpaelzer: If u can give me guidance on which logs or some steps to execute i can try out and get u back the inputs needed [09:41] imraan: let me collect a few things what you could do [09:41] okies [09:45] let me know what all things u need [09:45] none yet, I try to repro the crap I want to recommend before leading to you bullsh§$% [09:47] cpaelzer: Okies [09:53] imraan: lets start comparing what libvirt and qemu think on active block jobs [09:53] I assume the issue is only on vdb as you reported, please go to others as well if needed [09:53] the libvirt view would be [09:53] virsh blockjob vdb --info [09:54] while we can "ask" qemu what it thinks with [09:54] virsh qemu-monitor-command --pretty '{ "execute": "query-block-jobs" }' [09:54] imraan: ^^ [09:54] cpaelzer: the issue is not only with vdb it is with many other volumes also.. [09:55] i had just posted the sample [09:55] imraan: I guess it is a race of some sort [09:55] imraan: when you have the issue again, you can check with those commands [09:55] imraan: and from there one can debug/think further [09:56] imraan: if you have a test environment that you could use to recreate you could do so until you feel that you have a way to 100% recreate - once you did you could try https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/2620 and https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/2619 [09:56] okies...u have any idea on which scenrio the blocking of jobs can happen....if some operations ongoing on the volume os that sort? [09:56] those contain (now outdated) libirt builds around a similar issue [09:56] so you'd need to downgrade [09:57] but it makes no sense to try these if you are not convinced to have a way to "from fresh sys to 100% sure to repreoduce" [09:58] imraan: it is just a thought/theory - but on the test builds it was about soe jobs not being considered jobs [09:58] imraan: so they could race against each other [09:59] imraan: but backporting the fix is very invasive, so without a way to reproduce/test that isn't going to SRU [09:59] imraan: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_39a8dbb93caf4ec889f8a1b7f69885db/bileto-2619/2017-03-21_11:13:47/xenial_libvirt_content.diff [10:01] imraan: and OTOH the fix is available in any later release (without the uncertainty to accidentially breaking a lot of people on upgrades= [10:01] imraan: and many people use the Ubuntu Cloud Archive, so even on 16.04 they are on newer qemu/libvirt levels quite often [10:01] imraan: anyway - sorry that this is no easy "do this and you are good" case [10:02] imraan: I hope that will help to get better info next time [10:02] imraan: in case my response latency is very high, I'm soon out for lunch [10:05] Thanks cpaelzer.. [10:05] For ur inputs [10:07] I will further check on this have ur lunch... :) [13:20] Should the fsck option from the recovery menu be failing becasuse I have a Samba / cifs mount in my /etc/fstab file? It seems that it tries to mount everything read-only, cannot do a DNS lookup on my cifs server, then fails. === caribou_ is now known as Caribou === Caribou is now known as Guest42097 === Guest42097 is now known as caribou [13:27] Is the entry marked to be checked? [13:42] openstack folk: what is the intended difference between xenial/pike and xenial-updates/pike in UCA? most deployers seem to use only the latter, but that is lagging behind currently. see e.g. https://github.com/openstack/kolla/blob/master/docker/base/sources.list.ubuntu#L13 [13:43] the updates column here looks pretty red, too: http://reqorts.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/pike_versions.html [13:44] jamespage: coreycb: ^^ [13:45] frickler: yes we've not release yet [13:45] frickler: xenial-proposed/pike has latest atm [13:45] frickler: cpaelzer and I are working on a libvirt hang that we really don't want to push to -updates :-) [13:46] lordievader: the recovery menu seems to claim it makrs everything to be checked. [13:46] jamespage: so with the release, xenial-updates would get synced and people installing only from there will be correct again? [13:46] frickler: yes [13:46] jamespage: great, thanks for clarifying [13:46] TafThorne: It shouldn't, I get the feeling every entry is simply marked to be checked. [13:47] TafThorne: Did you read the documentation of fstab? [13:50] lordievader: when I added the cifs share 2 years ago I probably did. I iwll give it another look [13:53] lordievader: my sixth field is set to 0. fsck should be leaving the network mount alone.. [13:54] the / mount point is 1 so it should be getting checked first. Maybe it only takes a couple of seconds and I am missin git? [13:54] Hmm, indeed. That would suggest it does not get checked. [13:55] Does it succeed if the entry is commented? === JanC is now known as Guest85497 === JanC_ is now known as JanC [14:21] lordievader: I'll give that a check this evening. Rebooting my main machine is rather disruptive in office time. [14:40] cpaelzer: sorry, i meant to say, also use --no-fetch [14:41] cpaelzer: i did get some failures from the cron job, i'm guessing related to what you see [15:18] nacc: ok, thanks for the info [16:12] hello. i need some help installing python anaconda in my ubuntu x64 [16:20] heistheDude: https://docs.continuum.io/anaconda/install/linux.html looks fairly straightforward? What issues are you running into? [16:26] cpaelzer: ok revised a little - but I think this works OK - http://paste.ubuntu.com/25439408/ [16:27] cpaelzer: I just revert back to always doing the deferred processing path [16:35] coreycb: cpaelzer: https://launchpad.net/~james-page/+archive/ubuntu/libvirt-testing has the patched version for Xenial [16:37] andol, i did this exactly still i cannot see anaconda in the application list [16:43] jamespage: that seems to do the trick! i made it through 20 hard reboots without a problem. [16:44] how can i run installed anaconda software in terminal? [16:44] coreycb: basically I ripped out the in-context call that caused the problem and switched back to deferred calling which works OK [16:44] I think [16:45] jamespage: seems like a good work-around at least until upstream has something better [16:45] coreycb: cpaelzer: patch attached to https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1714254 [16:45] Launchpad bug 1714254 in libvirt (Ubuntu) "libvirtd hangs after # instance creation events" [Critical,Triaged] [16:50] coreycb: i get a clean tempest smoke with that patched version installed [16:50] jamespage: excellent [17:03] rbasak: figured out the issue, i think [17:04] or a issue, at least :) [17:08] rbasak: stupid typo on my part [17:10] nacc: ah. The hardest type of bug to debug :) [17:12] heistheDude: Well, that installation page mentions the command "conda list". Can you run that from the terminal? [17:17] no andol [17:17] says conda command not found [17:19] heistheDude: What about step #9? [17:20] that workss fine. [17:20] source ~/.bashrc i mean [17:21] program is getting installed fine, but i cannot find it anywhere [17:33] heistheDude: you ran exactly those steps? [17:43] Will live kernel patches feature be in the community version of Ubuntu server? Looks like its a Canonical service 'paid' https://www.ubuntu.com/server/livepatch [17:46] joeyjoerobinson: as mentioned on that page (perhaps in the video) you get 3 free machines [17:46] joeyjoerobinson: and it's already available, afaik [17:47] Yeah, I mean truly free. Not paid after 3 [17:47] joeyjoerobinson: no [17:47] joeyjoerobinson: not afaik [17:47] joeyjoerobinson: also not entirely sure what you mean by 'community version'? there is only one ubuntu server. that there is a canonical supported stack above it doesn't change that (afaics) [17:49] nacc the Canonical supported version offers this as a PAID feature. It should be obvious that I am looking to know if it will be a feature that is supported by the Ubuntu community without any requirement of payment [17:50] joeyjoerobinson: as i said, the community has 3 free [17:53] That sux and is antithetical to FOSS. [19:01] tried installing server on an old laptop and it couldn't find the CD ROM. in this case it has a CD ROM but then the rest of the install goes to shit. [19:02] why does it need a CD ROM? never had that issue on other server boards without a CD ROM. [19:06] it needs a cdrom if you're installing from a cdrom [19:07] Yea....nothing should EVER cahrge for anything...it all should be free.......damm those developers who want to eat.... [19:08] what a moron [19:08] Ussat: ? looks like #ubuntu-offtopic to me [19:08] Yea well...fair buff [19:08] fair nuff [19:10] in installing from USB drive. and I'm not fully installing I'm just trying to see if the installer recognizes my new USB Ethernet adapter [19:10] It can't boot to USB stick? [19:10] madLyfe: if you're seeking help with this issue, you should describe more closely how you installed, which version and variant you installed, which error messages you ran into in which context [19:11] don't 16.4.3? whatever the most recent version is and used the HWE option during install [19:11] the most recently released ubuntu version is 17.04 [19:12] sorry most recent stable [19:12] that's the most recent stable release. releases are always stable. [19:13] (at least that's the goal) [19:13] most recent LTS. that should do it. [19:14] Ubuntu Server 16.04.3 LTS then [19:16] hmm changed USB ports and regular install(not HWE) and no issue this time [19:16] still not seeing my USB Ethernet adapter tough [19:16] we [19:17] wtf. it notices the laptops onboard LAN and wireless though. lol [19:19] guess it just didn't like USB Ethernet adapters [19:22] ya odd. the HWE install doesn't like not having a CD ROM [19:22] https://usercontent.irccloud-cdn.com/file/YSnCfbj2/irccloudcapture1753526775.jpg [19:23] https://usercontent.irccloud-cdn.com/file/RfWQO85y/irccloudcapture1408436233.jpg [19:23] https://usercontent.irccloud-cdn.com/file/bw5g39gW/irccloudcapture1261440775.jpg [19:26] also I wonder if I need to plug in an Ethernet cable into the USB Ethernet adapter for it to recognize [19:26] shouldn't need to [19:27] it should be the chipset it detects, not the link [19:29] thats crazy that server doesnt find two different usb ethernet chips but finds laptop wifi chip. [19:29] why is that crazy [19:29] are the chipsets supported by linux, if so, are the supported by the kernel you're using [19:29] USB and ethernet seem more along the lines of server than wireless [19:29] does it see the device at all at a dumb level [19:30] no it doesn't [19:30] usb is supported, ethernet is supported [19:30] usb ethernet is a niche thing with odd chipsets [19:30] is the chipset supported under linux at all, is it supported with the kernel you're using, does it see the device but not know what it is [19:33] both of these adapters are recognized fine under the desktop live iso [19:34] so what chipset are they and what kernel module is driving them under the desktop ? [19:34] how do you know they are not detected under the server install [19:34] (apologies if I've missed your background to this) [19:35] ya i can get them to run after install. i want them to be recognized and run during install. i dont want to manually setup them up after. [19:36] how do you set them up ? [19:36] (what are the steps you take) [19:37] just check lsusb to make sure its recognized then add it to network interfaces [19:37] as in the network interface file [19:38] yes because it was not automatically put there. [19:39] so if you look at the syslog when you boot the server install do you see it detect the usb devices ? [19:40] how do you do that? [19:41] swap to the different virtual console and open the syslog in a text editor [19:42] why key to swap to a different virtual console? [19:43] alt+F$something [19:44] Stokachu :( no conjure-up versions --stable or -edge work for me localhost. --edge http://paste.ubuntu.com/25440328/ And with --stable I get http://paste.ubuntu.com/25440328/ [19:45] Would any of this be caused by me having manually created a controller and have a running machine attached? [19:45] those are the same links [19:45] Sorry [19:48] --stable http://paste.ubuntu.com/25440433/ stokachu [19:53] squashua: shouldn't [19:53] squashua: can you `touch ~/.cache/conjure-up/test.db` [19:55] Stokachu it didn't give any output [19:55] do you see it in `ls ~/.cache/conjure-up/` [19:56] Yes [19:57] ok so must be something else [20:02] Well then [20:05] ikonia: alt-f brings up the installer menu for me [20:05] Stokachu I stay @ a dead end with this.. It's so frustrating. [20:06] squashua: this is the first time you've come for help that i recall [20:07] https://usercontent.irccloud-cdn.com/file/KAerBpRP/irccloudcapture401038223.jpg [20:08] Stokachu sorry.. I can't keep my name cause I keep getting disconnected. This is Brazier. This is the same poweredge you have been on recently. [20:08] squashua: i see, so what happened? it was working [20:08] & it's still open to you [20:08] can you pm me the ip and username again [20:11] squashua: ^ [20:14] Stokachu sorry again. Not really. I never was able to launch I stance or anything. It just successfully deployed. So I refreshed today and these are the problems I'm having. [20:14] Instance* [20:14] braziercustoms-m: can you p m me the ip and username [20:14] Yeah [20:15] Can you pm me first stokachu [20:15] It's complicated on this phone lol [20:30] https://usercontent.irccloud-cdn.com/file/KH9G61vP/irccloudcapture777044987.jpg [20:42] Stokachu thanks for showing me my ignorance again :D. on the database is locked, what if it checked for another instance and asked what you want to do with it? [20:43] That would have kept a dummy like me off here longer :D [20:54] Stokachu http://paste.ubuntu.com/25440789/ [20:55] looking [20:58] brazier: it's going now, i need to look into this db lock stuff more [20:59] What happened? [21:00] the database was locked again [21:00] spells need to be udpated to make sure flush the db and close the connection [21:03] Hmm it didn't say db locked that time did it it Stokachu [21:03] yep in the logs [21:05] It does further up the log I see now.. guess I been caught up in the last error it showed. [21:07] Stokachu can I intern with you for a while :D [21:07] haha [21:09] Stokachu I thought the novalxd was the only one you could install on a single machine? That was the one I been trying to install [21:10] brazier: nah they all work on a single machine [21:10] i just forgot that was the one we were using before [21:10] I asked you before.. must have been a miscommunication.. [21:12] Stokachu will the whole conjure up cloud run inside of an lxd machine? [21:12] brazier: yea you gotta do some tweaks to the host container though [21:12] brazier: https://stgraber.org/2017/01/13/kubernetes-inside-lxd/ [21:14] Stokachu nice.. thanks. [21:16] Stokachu since you chose kubernets you think that would be the better one to try to learn/understand? [21:17] brazier: it is the new kid on the block [21:17] if it were me i would be learning kubernetes [21:18] Stokachu sometimes I have a hard time figuring out where to begin. [21:18] yea, it's a fast paced industry [21:19] Stokachu are any of the systems available capable of creating instances where the software can control specific hardware? [21:20] On the physical machine? [21:20] not sure i follow what youre asking [21:23] Stokachu Could I say create an instance that could communicate with a usb device attached to the physical machine [21:27] Stokachu btw it deployed successful [21:31] rbasak: so we have a case of an old/new mismatch causing an issue. ipsec-tools will fail to import for the time being, becuase the debian changelog object has no means to handle bad changelogs (it can either emit a warning and skip entries or make bad entries fatal). We are more lenient (in the shell parsing/dpkg-parsechangelog). Thoughts? [21:31] rbasak: http://paste.ubuntu.com/25440934/ [21:33] nacc: so is there a way to get the server installer to force the r8152 driver or whatever to recognize the chip? [21:34] madLyfe: we concluded the driver wasn't available on the iso, right? [21:34] cpaelzer: fyi, i am running the virt-manager import again now, to see if the same error is seen or not. Worst case, I'll do the reimport this afternoon [21:35] I'm not sure. I thought you guys said it was in that extra package [21:35] we got it working after the install but not during. [21:35] madLyfe: yeah, so the -extra package is probably not available/used during the install itself [21:35] i'm not 100% [21:35] madLyfe: the simplest ('simple') choice would be use a custom kernel and initrd [21:36] probably just an initrd, really [21:36] madLyfe: that does have the module in it [21:36] madLyfe: and then install the -extra package during install time [21:36] what does that mean. ELI5 please [21:36] madLyfe: you'll need a system that is working to work on [21:37] madLyfe: you'll uncompress the installer initrd there [21:37] madLyfe: then you'll add files to the installer initrd [21:37] madLyfe: recompress the initrd [21:37] madLyfe: and remaster the iso and burn to USB (normally you'd do this for a network install, but since you don't have one) [21:44] nacc: if it's entirely historical, could it be an easy fix to keep a patch for it? [21:44] nacc: if we ignore the issue, what would be the consequence in this case? [22:05] https://www.hugedomains.com/domain_profile.cfm?d=sarnold&e=com [22:05] wtf [22:05] sarnold: you are for sale :D [22:05] 24 monthly payments of $134 [22:05] thats cheap :) [22:06] rbasak: not sure what you mean? in theory, any changelog that has those entries in it (the ones that appear to be being dropped) will fail to import without being patched [22:07] nacc: HO if you're free? [22:07] rbasak: yep [23:03] cpaelzer: finally getting aroudn to implementing --reimport. I'm testing it now, I think it actaully will be pretty easy to do (presuming this passes) [23:09] rbasak: ok, so i'm still not 100% sure on what's going on, as running the importr by hand has gotten past that point. I'll debug that a bit (I do think it's the script calling to the cli, maybe). But in any case, I'm realizing anohter place we're not currently tracking (for eventual shell removal) is when we do fail to strictly parse a changelog -- there's no logging or any way to know we fellback to [23:09] shell parsing. HOw should we handle that? [23:31] cpaelzer: it would appear the reimport will succeed, but i'll let it finish first (it succesfully imported all of debian again). So I think the failure we're hitting now is the bug in the repository construction with the old vs. new pristine-tar. I'll make sure it integrates your two upload tags and then i'll update the snap (it's a hidden flag for now0 [23:40] rbasak: yeah, so the by-hand import of ipsec-tools, from my bastion, succeeded