=== Exagone314 is now known as Exagone313 === Lcawte is now known as Lcawte|Away [01:19] If I’ve got a 12.10 box I need to run up to the latest LTS, is do-release-upgrade going to go straight to 14.04 or through all the in between releases one at a time [01:25] it looks like from /etc/update-manager/release-upgrades it will be on the normal path instead of lts [01:26] I assume if I change that to lts, it will stop at 14.04 instead of running all the way to 15.10 [01:42] grilled-cheese: pretty sure you'll have to do it release by release. === dw2 is now known as dw1 === devil is now known as Guest48289 === Guest48289 is now known as devil_ === devil_ is now known as devilzz === devilzz is now known as devil__ [08:46] Good morning. === Lcawte|Away is now known as Lcawte === cpaelzer is now known as cpaelzer_afk === cpaelzer_afk is now known as cpaelzer [10:12] rbasak, morning - can I bring https://bugs.launchpad.net/juju-core/+bug/1517258 to your attention? [10:12] Launchpad bug 1517258 in juju-core "juju 1.24.7 precise: container failed to start and was destroyed" [High,Triaged] [10:12] rbasak, I think its something todo with the lxc version in the cloud-tools pocket for the cloud-archive [10:14] cloud-tools is quite out-of-date - https://launchpad.net/ubuntu/+source/lxc/1.0.7-0ubuntu0.9 is the latest trusty version afaict [10:18] jamespage: could you tell me when the first maintenance release will happen for openstack liberty packages in the UCA? [10:19] Trying to get liberty running and there are some important bugfixes in the stable/liberty branches (mostly neutron and nova) that we really need! [10:22] toMeloos, the plan was to track any point releases issued by the individual projects - I see neutron popped a 7.0.1 in the last 24 hrs, I'll findout what nova are going todo [10:23] toMeloos, we might be able to squeeze an stable update in pre-christmas but given that mitaka b1 is also in progress right now, the packaging team might be streched to make that [10:23] so most likely early Jan [10:23] but I need to checkin with them first [10:25] that's too bad. nova for example has a serious regression which means ceph and libvirt won't work together. bugfix is merged, but waiting to be packaged... [10:27] and neutron seems to have a lot of L3 DVR and HA fixes as well [10:28] jamespage: can't building packages from point releases be automated by the way? [10:29] toMeloos, well we can pick and high priority bugs earlier if need be [10:29] toMeloos, like I said I need to catchup with coreycb and see what they have capacity for [10:29] toMeloos, can you point me at what is hurting you? [10:31] jamespage: this one right now: https://bugs.launchpad.net/nova/+bug/1508230 [10:31] Launchpad bug 1508230 in OpenStack Compute (nova) "regression in cloning raw image type with ceph" [High,Fix released] [10:33] and as I'm sure you can understand getting liberty working is a work in progress, so I can't guarantee nothing else will pop up after getting past this hurdle... [10:34] toMeloos, ack [10:34] coreycb will be around in ~3 hrs or so [10:44] jamespage: I see about half a dozen bugfixes each for cinder, glance, nova and neutron. Slightly less for ceilometer, horizon and keystone. Haven't tested heat yet, but worried seeing that this list of merged patches contains a lot of bugfixes... https://review.openstack.org/#/q/status:merged+project:openstack/heat+branch:stable/liberty,n,z [10:45] I'll be around in ~ 3 hrs :-) [10:54] in usually daemon configs, can I use more than one bind address to allow access to servers? e.g redid server? === cpaelzer is now known as cpaelzer_afk === cpaelzer_afk is now known as cpaelzer [12:06] jamespage: for bug 1517258, what's enabling the cloud tools pocket on the Juju bootstrap node? Is that Juju itself doing it? It's been a while since I've had to care about that pocket! [12:06] bug 1517258 in juju-core "juju 1.24.7 precise: container failed to start and was destroyed" [High,Triaged] https://launchpad.net/bugs/1517258 [12:06] rbasak, I think its juju on 12.04 installs [12:06] only... [12:06] I think [12:08] OK, so I think the bugfix is in precise-updates but agree it's missing from the precise/cloud-tools CA pocket if that's what you're saying? [12:08] So we should push Trusty'x lxc back into precise/cloud-tools? [13:31] toMeloos, jamespage, ceilometer and neutron are the only projects that have tagged point releases for liberty at this point. nova doesn't have a point release out so the best we can do at this point is cherry pick patch(es). [14:08] cpaelzer: do dpdk consumers currently have to specify which split DPDK libraries they should link with? [14:30] rbasak: they can but don't have to IIRC [14:30] What happens if they don't? [14:31] rbasak: they can link against the linker script, the combined library, individual libraires or static build [14:31] rbasak: depending which way they choose they have to specify subibs or not [14:31] rbasak: static also being available as split or combined [14:33] Let's ignore static, since the end result will be the same I think? I presume it's easiest for static to just follow what happens with shared. [14:33] rbasak: yes [14:33] Red Hat want the combined library out, right? [14:34] Are they saying that the linker script should become unnecessary, or that the linker script should be used to end up with shared library dependencies on the split libraries? [14:35] rbasak: I understood it as "linker script replaces combined library" [14:35] rbasak: the linker script auto-refers to the individual libraries then [14:35] rbasak: so the second part of your question IMO [14:35] rbasak: "he linker script should be used to end up with shared library dependencies on the split libraries" [14:36] If a library consumer uses the linker script to end up with shared library dependencies on the split libraries, will it end up with dependencies on just the set of libraries required or all of them? [14:37] rbasak: I understood the mail discussion that he will only end up with those needed, btu I never seen it in action [14:39] coreycb: I can live with that, but is there any chance of a maintenance release in the very short term? [14:40] coreycb: recommend having a look at heat too. lot of bugfix commits merged in stable/liberty for it! [14:40] cpaelzer: OK, thanks. [14:40] cpaelzer: I'm drafting a reply, I'll ask for a review from you before sending. [14:40] rbasak: sure [14:40] toMeloos, those are questions for upstream. they are the ones deciding on when to cut point releases. [14:42] toMeloos, as for the existing point releases for ceilometer and neutron, we'll do our best to get those into the proposed pocket of the liberty cloud archive before the holidays [14:49] coreycb: thanks! that would be a great start [14:50] toMeloos, ok good. and I just poked into #openstack-nova to ask if they have a point release coming soon. if not I'll try to sru a fix for the but you mentioned earlier. [14:50] s/but/bug [14:52] coreycb, jamespage: by the way, are there any plans for packaging Octavia? We need it to deploy Neutron LBaaS V2 and it should be declared stable with Liberty but it's not in the liberty cloud archive [15:04] toMeloos, no immediate plans [15:35] thanks bradm [15:55] cpaelzer: first draft: http://paste.ubuntu.com/13826699/ [16:07] cpaelzer: toned down a bit: http://paste.ubuntu.com/13826999/ [16:07] (interesting URL diff) [16:13] rbasak: hehe, reading the new one [18:19] Hi, 14.04 proftpd has a major flaw [18:20] critical [18:21] you must be aware of it [18:22] ProFTPD Version 1.3.5rc3 [18:22] davidic657: http://www.cvedetails.com/cve/CVE-2015-3306/ ? [18:22] davidic657: can you prepare a debdiff for us to sponsor? there's a few bugs that need fixing, http://people.canonical.com/~ubuntu-security/cve/pkg/proftpd-dfsg.html [18:22] just get on with it [18:22] lol [18:22] fix the darn thing [18:23] davidic657: protftp is community maintained; I think no one actually uses it, since no one has prepared updates yet. [18:23] your supposed to be the top notch distro [18:23] its a major exploit === cpaelzer is now known as cpaelzer_afk [18:23] proftpd has nothing to do with the ubuntu server [18:23] davidic657: more information on how to prepare an update is available at https://wiki.ubuntu.com/SecurityTeam/UpdatePreparation [18:24] unless you mean this has been fixed upstream and the package needs updating [18:24] Sling: yeah [18:24] Sling: that's probably the case; see the people.canonical.com url I pasted earlier for details [18:25] ok, then I said nothing [18:25] davidic657: feel free to run vsftpd, which is in main and supported by the ubuntu security team [18:25] sarnold: proftpd is built into my CP [18:25] hate taking anything off default :) [18:25] davidic657: what's that? [18:26] Virtualmin [18:29] yikes, I hope you've got _that_ firewalled off to just your home IP too... [18:29] sarnold: that to me :) [18:29] I guess not === cpaelzer_afk is now known as cpaelzer === cpaelzer is now known as cpaelzer_afk === cpaelzer_afk is now known as cpaelzer === tekku is now known as tekk [20:40] hi all. i'm trying install ubuntu server on a ts140 (2 disks, 3 tb each one). i installed the system and i configured the raid, but now i've problem with grub. seems i cannot install it (fatal error). [20:40] anybody can help me? [20:43] bios configured as: quick start disabled. secure boot disabled. boot mode legacy first. disks AHCI (i used the software raid during the installation). [20:48] the image i used is: ubuntu-14.04.3-server-i386 [20:48] plm: what error do you get? irc works best with specifics.. [20:50] the exectuion of "grub-install /dev/sda /dev/sdb" failed. this is a fatal error. [20:50] dang that's it? how about log files? [20:51] i'm in the installer. how can i access the log file? [20:52] maybe alt-f2 or control-alt-f2? [20:52] . === cpaelzer is now known as cpaelzer_afk [20:55] ok. [20:57] grub-install: warning : this GPT partition label contains no BIOS Boot Partition; embedding won't be possible. /// grub-installer: grub-instal: error: embedding is not possible, but this is required forRAID and LVM install. /// grub-installer: error: running 'grub-install --force "/dev/sda/"' failed. [20:57] plm: You need to fix partitioning. You need a small partition of GPT type ef02 (I think) at the start of each disk. [20:58] md: delaying resync of md2 until md1 has finished (they share one or more physical units). [20:58] what's ef02? [20:59] plm: The GPT partition type ID for "reserved BIOS boot partition." [20:59] plm: When installing a BIOS GRUB loader (not UEFI) on a GPT-partitioned disk, which yours are because they're >2TB, you require that partition in order for GRUB to have somewhere to embed itself. [21:00] but now, after did raid i cannot edit partitions. how can i do? have i rewrite the partition table? [21:00] plm: Reinstall would be the easiest way. [21:01] ok, i'm rebooting the server. [21:05] ok, now i'm in the partitioning step. how should i proceed? [21:06] plm: Create a small (1MB is fine) partition at the start of both disks. Set the type to "reserved BIOS boot partition." [21:06] Then partition the disks normally other than that. [21:08] the partition tool shows me the raid [21:08] i need to delete it... before. [21:09] trying to create a deb of nodejs 4.3.2 and running debuild after having this rules file seems to not include usr/bin/nodejs in the deb only the doc's any idea what is going on? https://bpaste.net/show/3a9c2cf852c7 [21:12] quantic, i removed the previous partitions. but when i tryed to create a 1 Mb partition it says it's too small [21:13] What did you enter at the size prompt, specifically? [21:13] I normally do 8mb [21:13] I think 1mb is just enough these days [21:14] 1M is quite massive for a GRUB partition. [21:14] On my servers, it's exactly 1MB. [21:16] but the partition tool don't allow to use a so small partition. while, if i ask to the partition tool to autocreate a partition table, it creates 1mb (free space) ; 1mb biosgrub ; 3tb / ; 17mb swap. can it be a solution? [21:16] or the 1mb empty space is a problem? [21:17] plm: The 1mb biosgrub is what I'm talking about creating. o.o [21:18] of course, i understand. but it's not at the begin. there is 1mb free space before. [21:19] is it a problem? [21:21] quantic, have i create a md for biosgrub also? [21:30] plm: No, it won't work. Just leave it as a bare partition. === CiPi is now known as cipi [21:30] no [21:30] it can be anywhere [21:30] ideally it should be within the first 500megs of the disk, or 8gigs or whatever bios limit you have :) [21:31] but otherwise it could be at the end of the disk [21:34] quantic, so i've to partition the two disks in the same way, and i'll obtain something like: so the partition table would be something like: 1MB free; 1MB biosgrub; 3TB RAID; 17GB RAID & 1MB free; 1MB biosgrub(?); 3TB RAID; 17GB RAID [21:35] but the 1Mb of biosgrub i've in the sda, what it would be in sdb? [21:36] a mirror image [21:37] if you expect to boot from it [21:38] so, patdk-wk i've to select also in the second disk biosgrub? [21:38] yes [21:38] however i cannot flag this partition as bootble. [21:38] is it normal? [21:38] and in your grub config, select it also [21:38] heh? [21:38] it should be flaggable as bootable [21:39] GPT doesn't have a bootable/active flag. [21:39] So, no, you shouldn't be able to flag it as bootable. [21:40] so, it's normal. ok. cool :D [21:47] so at the end i've sda1 and sdb1 biosgrub (1mb each one); MD0 3TB for Root; MD1 17GB for Swap. right? [21:50] quantic, patdk-wk i'm going to arrive to the grub installer.. pls don't leave me now. :/ [21:56] plm: Sounds about right. [22:09] So. This is a real shot in the dark here, but is there any chance anyone here has written a CIS-benchmark compliant preseed for 14.04 ? [22:18] ok, finally i'm at the grub installer tool. it asks me to install it on mbr. but if i right understood, maybye i've install it on sda1 and sdb1 right? [22:21] No. [22:21] Install to MBR. [22:25] ok, finally it was succesfully installed thank you very much [22:27] Cheers. :) [22:28] sda1 and sdb1 are partitions, sda and sdb are where you can install bootloaders === Jordan_U_ is now known as Jordan_U === arosales_ is now known as arosales === Lcawte is now known as Lcawte|Away === cipi is now known as CiPi