[00:28] is there dh* suite tool for grub2 updates? [00:29] maybe dh_autoconf? [00:29] to save the trouble of scripting it myself and crafting a package.postinstall thing [01:47] is there a method to redirect logs collected and managed by upstart to rsyslog? [01:48] I didn't think upstart 'managed' logs.. do you have an example? [01:49] sorry - rather the logs that find themselves in /var/log/upstart/ [01:54] yoink: aha. hrm. upstart doesn't look very configurable there http://upstart.ubuntu.com/cookbook/#console-log === CodeMouse92 is now known as CodeMouse92__ [01:59] sarnold: thank you. My research led me to the same cookbook recipe. :/ Time to start building our packages for 16.04 I guess! :) [02:19] yoink: yeah, there's more options there === Malediction_ is now known as Malediction [06:35] I was wondering if anyone knows of a way to have an encrypted folder/partition/container on an openvz vps? [06:36] what's the goal? [06:38] I want to have a folder with private data that requires user input to decrypt the data if it is either unmounted or the vps/host is restarted. [06:40] it's just that if the domain or the kernel is cracked, it's not _really_ safe.. perhaps it'd be overlooked by hasty automated data scrapers or something, but the machine's owner really _owns_ it.. [06:40] can you use ecryptfs from within openvz? [06:41] if not, how about truncating a huge file and doing LUKS on a loopback mount? is that allowed? [07:43] Hello [07:43] I decided to use lxc for virtualization at least to has contaninerized server. [07:44] My host machine has like 192.168.1.150 local ip. and netmask 255.255.255.0 . I think I am able to give ip from same network of host to guest. [07:44] Am I right [07:45] for example. can I assign 192.168.1.200 local ip to guest [07:45] And should I start from where? should I search "lxc bridge network" [07:47] temhaa: hopefully useful https://insights.ubuntu.com/2016/04/07/lxd-networking-lxdbr0-explained/ [07:49] sarnold: thanks for replying. so lxc, lxd which one of them should I use? [07:52] temhaa: probably lxd; I think its simplified a lot of what was learned via lxc to work well vs not work well [07:58] sarnold: I got it [09:11] jamespage: hi, in the past to fix up bug 1495895 qemu-block-extra was added as dependency [09:11] bug 1495895 in qemu (Ubuntu) "Unable to attach rados block device to instances" [High,Fix released] https://launchpad.net/bugs/1495895 [09:11] jamespage: I wonder if we could drop that to a recommends [09:12] jamespage: it really is not a depends and it has an -extra in its name for a reason [09:12] jamespage: Debian would go the "recommends" with me and we could drop some delta [09:12] jamespage: IMHO this should still solve all issues we had [09:12] jamespage: but at the same time allow users to install qemu without (if they want) [09:13] jamespage: since we default to install recommends we should be fine right [09:13] jamespage: Debian is integrating the recommend right now, and I want to use that next re-merge in a few days [09:13] jamespage: but an ack from you as the original reporter would be nice, to ensure there was no hidden context I don't see [09:14] jamespage: might be traveling so also highlighting coreycb and rharper who worked on that^^ [09:16] cpaelzer: that sounds OK to me [09:16] Hi, (automatic) upgrades is killing my server during kernel upgrades (more specifically recreating the initramfs images) with OOM messages. Server only has 256M. Odd thing is, it also has 1G of swap and it only uses 4M of that ever. Any ideas? [09:17] jamespage, cpaelzer good morning [09:18] jamespage, we officially have a new 13.1.3 tag in the git repo for Nova [09:19] jamespage, I have a question, when I do `gbp import-orig` where is exactly that you download the tarball from on the openstack website ? or do you generate the tarball using the git repository of nova at the given tag ? [09:20] zioproto: tarballs.openstack.org is normally used as that is where all of the release tarballs endup [09:20] jamespage, https://tarballs.openstack.org/nova/nova-13.1.3.tar.gz [09:20] great [09:20] zioproto: the git packaging repo and the project repo are distinct [09:20] yes sure, the packaging repo is in launchpad [09:21] ok I try to do some work refreshing this nova package [09:23] thanks for the ack jamespage [09:23] freakynl: depends what your applications do with the memory [09:24] freakynl: are you interested (independent of your upgrade issue) to test to get it swapping more? [09:26] freakynl: stress-ng --vm 2 --vm-populate should give you ~512MB pressure that is swappable - in a second console you can check if it does indeed swap while these are running [09:31] cpaelzer: Actually have 2 of these VM's. They run powerdns and mysql which usually leaves enough room in memory. The difference is, one is running under vmware the other under hyper-v. The one under hyper-v uses much more swap and gets through the kernel upgrades just fine. Both have 256M RAM. The vmware one has 1G of swap, the hyper-v one only 512M. Both are running ubuntu 14.04.5 LTS server and [09:31] have little differences besides open-vm-tools on one and the hyper-v tools on the other [09:32] Don't get why one uses the swap for recreating the initramfs images and the other doesn't however :) [09:32] freakynl: some sort of memory balooning maybe [09:33] freakynl: it is common in virt environments to try to adapt guest sizes instead of swapping [09:33] freakynl: maybe something like that prevents it from behaving as expecte? [09:33] +d [09:34] cpaelzer: could very well be. Can't even run dpkg --configure -a, it will kill my SSH session and nearly every other proces on the server [09:34] freakynl: can you cat /proc/meminfo to see roughtly where mem is? [09:34] pastebinit maybe [09:37] cpaelzer: might be something with the open-vm-tools and the kernel, I updated the kernel to xenial on the 14.04 instances [09:37] Not sure what that does with the open-vm-tools [09:40] cpaelzer: Hmm was getting those results but had to reboot it first for fresh values, 3rd time I rebooted it, completed the upgrades now. Only 2.6M in swap now [09:45] freakynl: I read a bit, it really could be the ballooning which has issues with your too constrained system [09:46] freakynl: usually you'd rather slighly oversize a guest and balloon it down (+swap in the host if ever needed) instead of swapping in the guest [09:47] freakynl: there can be weird cases of double faults if the guest swaps in itself - onto a page that is swapped out by the host, to be swapped in by the host , ... [09:47] freakynl: so I can't blame anybody to try to avoid, but I'd think in your super small system that might cause problems [09:47] freakynl: I'm happy you could pass the update [09:47] cpaelzer: system runs fine when not updating :). Really doesn't do that much. 18MB mysql database + powerdns. Normally doesn't hit swap, but needs some more ram during initramfs creation. It's not swapping in the hypervisor at all [09:47] freakynl: but you might try the workload I sent and disable ballooning to see if your system could swap fine then [09:48] freakynl: depending on that insight you can then either keep it balooning but increase size or disable it or ... (whatever you choose) [09:48] cpaelzer: Yea I'll disable ballooning, the host has plenty of memory anyways, vmware charges for RAM assigned to VM under VSPP though :) [09:49] freakynl: I'd hope they do not charge for what is ballooned down [09:49] freakynl: I like the business case - you have to pay in your supermarket for a full cart, no matter how full you make it :-) [09:50] because you could fill it up [09:52] cpaelzer: well under VSPP you get charged by the memory assigned to a VM. Or more accurately the *reserved* RAM for an instance, but if nothing is reserved they'll count a minimum of 50% of the RAM. Billing is capped at 24G, so a VM with 128G of RAM would still only cost 24G of RAM. But that's the only thing you pay, no CPU's, storage, etc. They have to base it on something I guess. Flat fee per [09:52] instance wouldn't be doable either [09:53] yep, with that details it sounds ok [09:59] hi, i just ran some updates and now mysql won't start.. systemctl just shows this > Failed to start MySQL Community Server. [10:00] https://paste.ubuntu.com/24089202/ [10:00] i can't take a dump and do anything else without it [10:01] ren0v0: databases are under /var/mysql usually, you could take copies of those, but it's binary form. Try starting mysqld manually [10:01] ren0v0: might be /var/lib/mysql btw [10:01] freakynl, how to start it manually ? [10:02] ren0v0: run mysqld as root (sudo mysqld OR sudo -i mysqld) [10:02] Good morning [10:02] doesn't do anything [10:03] ren0v0: no output at all? [10:03] freakynl, the only mysql process i have is > root 2907 0.0 0.0 9596 2604 ? Ss 10:02 0:00 /bin/bash /usr/share/mysql/mysql-systemd-start post [10:03] nope, no output at all [10:03] if i kill this process above it just comes back, clearly this is some major issue with systemd [10:05] ren0v0: not really, systemd is supposed to watch services and restart them when they die unexpectedly which is pretty much what happens if you kill it [10:05] not that specifically, the fact my whole mysql server broke when i ran updates (which i think included new systemd units) [10:06] ren0v0: should go away if you run: systemctl stop mysql (not sure what the service name is) [10:06] yea just did that ^ [10:06] so, trying to find out how to run it manually, i tried "sudo /usr/bin/mysql" but think thats client [10:06] ren0v0: yes the daemon is mysqld [10:07] ah > /usr/sbin/mysqld [10:07] doens't do anything when i run it like that [10:07] ren0v0: should normally be in the $PATH [10:07] ren0v0: try backing up /var/lib/mysql and removing and reinstall mysql-server after that [10:08] freakynl, would much rather get it running and take a proper dump, thats a last resort really [10:09] ren0v0: yes the main goal of that would be to get it running, the back-up is just in case, afaik removing mysql-server doesn't remove the databases [10:09] ren0v0: just a safety precaution [10:10] i see ok [10:10] here are some logs from syslog https://paste.ubuntu.com/24089229/ [10:11] ren0v0: hmm seems to have issues with the container. Did you reboot the machine after the updates? [10:11] yea, i'll try it again now [10:11] mysql5.7 is kept back [10:11] not sure if thats related [10:11] ren0v0: I'd personally migrate to mariadb, usually drop-in replacement [10:12] thats what i'm trying to do! [10:12] but i need my DB backups first [10:12] this is for zoneminder [10:12] ren0v0: they're binary compatible [10:12] freakynl, i took a snapshot of this container luckily, when i upgraded my ZM database was gone. [10:17] freakynl, ok backed up, force installed 5.7 and it works [10:17] mariadb or mysql? [10:17] mysql [10:17] now i can move onto install mariadb [10:17] thanks for the help [10:19] ren0v0: last time I did it it was just removing mysql-server and installing mariadb-server === JanC is now known as Guest86538 === JanC_ is now known as JanC [11:01] hello. I need your helps [11:01] I installed lxd . I started container from ubuntu [11:01] I want to access network of host to container. I tried bridge network but I couldnt success it [11:40] coreycb: ok, over the unrelated image issues now - verification done and bug update [11:40] +d [11:40] it seems I loose a lot of trailing "d" today :-/ [11:57] jamespage, coreycb I was able to compile the nova package versione 13.1.3 I pushed stable/mitaka and pristine-tar branches to https://code.launchpad.net/~zioproto/ubuntu/+source/nova/+git/nova How does it work for the merge request on launchpad ? I have to do 1 MR for each branch ? I will be testing the new package on our staging cluster. I will do the merge request when I am sure it is okay. [11:58] zioproto: yeah its one merge for each branch - so rather than do that I'd suggest that you just propose the review for the stable/mitaka branch, and note on the review that the upstream and pristine-tar branches will also need a review [12:00] great, the gbp import magic also updated the upstream branch ad added a tag, you are right === disposable3 is now known as disposable2 [12:29] I am trying to setup a new machine for building and I have a problem with debcheckout [12:29] W: Unable to locate package nova [12:30] I was pretty sure I never did anything special [12:30] to have debcheckout download sources [12:31] ops, I think I am missing a lot of deb-src stuff in my list files ... [12:50] anyone ever had a problem like [12:50] gbp:error: Couldn't commit to 'pristine-tar' with upstream 'upstream': [12:53] zioproto: if the branch called upstream is not there or has a different name [12:53] cpaelzer, it is there [12:53] it is called upstream [12:53] that is why I dont get it [12:53] it gets also updated [12:53] but the pristine-tar does not [13:01] cpaelzer, I found the problem, I was missing a package calles pristine-tar [13:01] apt-get install pristine-tar [13:01] but I have no idea why this was not installed automatically [13:01] I had it already in another xenial system [13:36] coreycb, jamespage for Nova Newton the ubuntu packages are updated at nova_14.0.2 but there is already a 14.0.4 release. In this case I just import the 14.0.4 tarball and I ignore the missing 14.0.3 release ? [13:36] zul, ? [13:36] zioproto, yes that should be fine [13:36] zioproto: yes skipping a release is OK [13:36] ok great, so I will refresh and newton now [13:36] yes [13:37] I guess most of the work is to figure out what to drop in debian/patches [13:37] zioproto: headers on patches may help there [13:37] hopefully the answer is 'all of them' but its depends [13:37] :) [13:38] for the mitaka refresh I could not drop anything [13:38] diff origin/stable/mitaka stable/mitaka shows that really a small portion of the code changes [13:53] hi, is there lately a error known with lxc + systemd (host, and container ubuntu server 16.04) - I'm getting the error 'Unexpected status from systemd "Failed to connect to bus: No such file or directory\n"' if trying to start the service I want to run in the container === mwsb is now known as Guest70562 [16:05] coreycb, I sent merge requests and I assigned them to you, or is it zul working on it ? [16:05] Mitaka https://code.launchpad.net/~zioproto/ubuntu/+source/nova/+git/nova/+merge/318630 [16:06] Newton: https://code.launchpad.net/~zioproto/ubuntu/+source/nova/+git/nova/+merge/318628 [16:06] zioproto: im just about to upload it thanks though [16:06] zul, what are you uploading ? :) [16:07] zioproto; nova 14.0.4/13.1.3 [16:07] great, so we were working on the same thing ? :D [16:08] well you I tested my branch on our cluster. Can you confirm you did not drop any debian/patches in this refresh ? [16:09] zul, it might be worth to look at MR anyways, just tell me if I am doing the thing right, I can help out with future releases [16:09] zioproto: sure [16:10] I am doing this upgrade Mitaka to newton, if nova works fine next week I have to upgrade cinder and neutron. I might send more patches [16:11] zioproto:yeah looks good..in the future you would need to include the bug number [16:11] How do I find it ? You open a LP bug when the new release is out ? [16:15] zioproto, for any stable update we need to add the bug number to the d/changelog according to the SRU process [16:15] zioproto, you can find more reading about the process here: https://wiki.ubuntu.com/StableReleaseUpdates [16:17] zul, I have a comment about nova 14.0.4 [16:18] zioproto: basically https://bugs.launchpad.net/ubuntu/+source/nova/+bug/1668313 and https://bugs.launchpad.net/ubuntu/+source/ceilometer/+bug/1664306 [16:18] Launchpad bug 1668313 in nova (Ubuntu) "[SRU] mitaka point release" [Undecided,Confirmed] [16:18] Launchpad bug 1664306 in swift (Ubuntu Yakkety) "newton stable SRU releases" [Undecided,New] [16:19] zioproto: is it about the certificate warning? [16:19] zul, no it is about commit 1ad5c7305c37079ced24bf623810e63d5eac2661 [16:19] it is 1 commit on top of release in the nova upstream [16:19] I had to workaround my nova database [16:20] so I am not sure if you are safe with bug #1668310 with the current 14.0.4. You might want to cherry-pick this in debian/patches [16:20] bug 1668310 in OpenStack Compute (nova) ocata "PCI device migration cannot continue with old deleted service records" [High,Fix committed] https://launchpad.net/bugs/1668310 === vonsyd0w_ is now known as vonsyd0w [16:20] zioproto: loking [16:21] I managed to have that commit in the release for stable/mitaka, but there was not a release for stable/newton [16:21] I would check with Dan Smith if it is safe to run without that patch [16:22] zioproto: should already be in the 13.1.3 [16:22] yes it is, but it is not in 14.0.4 [16:23] ok ill have a look [16:23] basically after you upgrade from 13.1.3 to 14.0.4 you have a regression on [16:23] 1668310 [16:24] coreycb/jamespage/zioproto/anyone else: mitaka uploaded pending ubuntu-sru [16:24] zul, LGTM [16:26] zul: as you didn't respond earlier, can you confirm that you saw my two issues regarding new Newton pkgs? https://bugs.launchpad.net/ubuntu/+source/nova/+bug/1668676 https://bugs.launchpad.net/ubuntu/+source/neutron/+bug/1668578 [16:27] Launchpad bug 1668676 in nova (Ubuntu) "Newton package needs to bump dependency on python-rfc3986" [Undecided,New] [16:27] Launchpad bug 1668578 in neutron (Ubuntu) "Newton package needs to bump dependency on python-pecan" [Undecided,New] [16:27] frickler: sorry i had the day off yesterday [16:28] zul: np, thats why I'm bringing it up again ;) just wanting to avoid you having to make double releases [16:51] I'm trying to start a service (systemd unit file) so I can have jupyterhub - an ipython tool - to run a server constantly. I followed these steps https://github.com/jupyterhub/jupyterhub/wiki/Run-jupyterhub-as-a-system-service but when I do sudo service jupyterhub start - i get no error or output whatsoever, but the server is not running. How do I investigate this further? [17:41] How can I check if a service is running properly? [17:42] systemctl status service.name [17:42] £ [17:42] I would try that [17:51] madwizard: service.jupyterhub.service, Loaded: not-found (Reason: No such file or directory), Active: inactive (dead) [17:53] jayjo: try systemctl alone and search for jupyterhub [17:53] Maybe the name is different [17:54] I epect its jupyterhub.service rather that service.jupyterhub.service [20:46] server is stuck at here http://dpaste.com/0AK3DZD while rebooting from console. Is there a way to break this from console? [20:46] it rebooted almost all the way and all other services seems started.. but no network yet.. so cannot ssh to it either === beisner- is now known as beisner [21:12] axisys, have to know what is starting up after that [21:12] sounds like some service is stuck on wanting networking, but didn't specify so it was started before