[00:05] can anybody tell me why chroot vsftpd is not working when I login as an testuser because I have the jailed list setup but testuser can still see top level folders but its not locking to testusers home directory [00:09] TimRe: are you missing these updates? https://bugs.launchpad.net/ubuntu/+source/vsftpd/+bug/1160372 [00:09] Launchpad bug 1160372 in vsftpd "Login is not possible" [High,Fix released] [00:09] that link doesnt work sarnold [00:10] TimRe: hrm, uvirtbot loaded it alright... try just looking up the bug number 1160372 [00:12] so its an bug then [00:15] TimRe: well, there was a bug in vsftpd that has had fixes released, I'm curious if the symptoms look like yours or if you've got a different one.. [00:20] well I did the patch one what was listed here: https://bugs.launchpad.net/ubuntu/+source/vsftpd/+bug/1160372 but now I cant even connect to it now either by port 21/22 [00:20] Launchpad bug 1160372 in vsftpd "Login is not possible" [High,Fix released] === Will_ is now known as Guest61878 [03:08] zul, jamespage more! http://people.canonical.com/~agandelman/ca/folsom/swift_1.7.4-0ubuntu2.2~cloud0/ [03:09] adam_g_: +1 [04:04] Hey all -- trying to start up 12.04 on EC2 + default VPC and I feel like I'm missing the "set up DNS, /tmp, mirrors, etc" script or something. Can anyone answer a few newb questions? [04:05] does the stock syslog that comes with ubuntu server support forwarding logs to a remote host over udp:514? [04:15] Skaag, ubuntu uses rsyslog by default === thumper is now known as thumper-afk [04:42] I managed to get it to work [04:42] thanks [04:53] zul, this one too http://people.canonical.com/~agandelman/ca/grizzly/swift_1.8.0-0ubuntu1.2~cloud0/ === marlinc is now known as marlinc|away === smb` is now known as smb === hXm is now known as hxm === marlinc|away is now known as marlinc [08:16] adam_g_, zul: swift uploaded to grizzly staging [08:16] adam_g_, also flushed through the 2013.1.2 updates to proposed === marlinc is now known as marlinc|away === thumper-afk is now known as thumper === marlinc|away is now known as marlinc === marlinc is now known as marlinc|away === marlinc|away is now known as marlinc [13:28] jamespage/yolanda: https://code.launchpad.net/~zulcss/python-keystoneclient/ftbfs-new-deps/+merge/170614 [13:31] utlemming, ping. [13:31] wondering if it would be useful for us to consider changing our ebsroot and instance store images on EC2 to be a partitioned disk and use the other pvgrub loader [13:34] ie, using krn/pv-grub-hd00_1.03-x86_64.gz rather than krn/pv-grub-hd0_1.03-x86_64.gz [13:35] the reason for this would be that then a.) grub might compleain less as there is a MBR to install to (although it still would'nt be read) and b.)if that worked, users could more easily take the disk image out of EC2 and boot it on kvm [13:37] there would likely be some changes necessary to do this even in cloud-init's grub config module [13:37] zul, nearly done with the havana ca updates [13:37] all aside from quantum uploaded [13:37] How do you remap CapsLock to Ctrl? I know how to do this on Ubuntu versions with a GUI, but no the server. [13:37] jamespage: sweet thanks === huats_ is now known as huats [13:58] zul: http://people.canonical.com/~jamespage/ca/folsom/ [13:58] can I get an ack on the libvirt update as well please [13:59] jamespage: perhaps...+1 [14:00] jamespage/yolanda: can you have a look https://code.launchpad.net/~zulcss/quantum/quantum-refresh/+merge/170626 [14:00] zul, almost there - http://reqorts.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/havana_versions.html [14:01] mm, i don't have permissions to do reviews now [14:02] yolanda: ?? [14:03] with the branch change i do not have the possibility to approve it [14:05] *sigh* [14:06] roaksoax: https://code.launchpad.net/~zulcss/python-keystoneclient/ftbfs-new-deps/+merge/170614 and https://code.launchpad.net/~zulcss/quantum/quantum-refresh/+merge/170626 === wedgwood_away is now known as wedgwood [14:17] jamespage, rbasak: https://code.launchpad.net/~yolanda.robla/ubuntu/saucy/exim4/update_smtp_banner/+merge/170631 [14:22] yolanda: do you still need to patch src/EDITME, or is that not required any more now that you're patching Configure-config.h? [14:23] you are right, i think it isn't needed [14:23] let me do a test build without that anyway, but it should be ok [14:23] yolanda: and do you need to refresh EDITME.exim4-light.diff? It would save having a reviewer verify that it doesn't change anything :) [14:24] yes, these 2 patches can be skipped if EDITME is not needed [14:24] i had to patch the -light.diff because it was giving conflicts [14:24] but with the new approach EDITME should be skipeed [14:25] Oh I see, of course. Now I see why you had to refresh it before. [14:26] i'm just building without the EDITME changes, let me see [14:27] it's quite strange how exim4 config is working, isn't it? [14:27] zul: done! [14:28] roaksoax: thanks [14:28] It does seem rather obtuse [14:28] I wonder what the reasoning was for doing it this way [14:29] the custom buildconfig parser looked strange to me [14:40] jamespage/roaksoax: https://code.launchpad.net/~zulcss/keystone/keystone-refreshed/+merge/170639 [14:47] rbasak, i pushed version without EDITME, it works fine [14:56] jamespage/roaksoax: https://code.launchpad.net/~zulcss/keystone/keystone-refreshed/+merge/170639 [15:26] if i wanted to configure a development server to face the WWW but not be accessible by the public should i just host it at www.mysite.com/dev and put a password up and hope hackers dont find it? whats the procedure on something like that? [15:26] yolanda: lsb_release -si says only "Ubuntu". So the banner says 220 ESMTP Exim 4.80 Ubuntu with no release information. Is that intentional? [15:27] rbasak, yes, we just need to show Ubuntu there === marlinc is now known as marlinc|away [15:28] rbasak : https://blueprints.launchpad.net/ubuntu/+spec/servercloud-s-server-app-banner-updates [15:40] yolanda: lgtm. I've commented on your MP, but I can't sponsor the upload. [15:42] mm, i can't see the line wrap... [15:43] you mean i exceed the max chars [15:43] ? [15:48] i updated it and i did a minor fix in patches, to add the BP url [15:48] so jamespage, zul, can you sponsor it? https://code.launchpad.net/~yolanda.robla/ubuntu/saucy/exim4/update_smtp_banner/+merge/170631 [15:49] also dovecot is pending: https://code.launchpad.net/~yolanda.robla/ubuntu/saucy/dovecot/update_dovecot_name/+merge/170071 [15:50] Yeah - convention is to wrap after <80 chars. [15:50] It doesn't matter much though [15:50] fixed, i just checked pep8 [15:51] getting this when trying apt-get upgrade The following packages have been kept back: [15:51] mysql-client-5.5 mysql-server-5.5 mysql-server-core-5.5 [15:52] 0 upgraded, 0 newly installed, 0 to remove and 3 not upgraded. [15:52] so, total newb. what does this mean [15:52] I am guessing it does not want to update MySQL [15:55] LargePrime: "upgrade" won't pull in any new packages to fulfill dependencies, so if dependencies have changed then it refuses to upgrade packages if the upgraded packages won't ahve all dependencies fulfilled [15:55] LargePrime: "dist-upgrade" will pull in new packages as needed. [15:55] rbasak: ive always just ignored when it says that [15:55] will that upgrade my distro? [15:55] "itll fix itself later" [15:56] o really? [15:56] primiarly because i didnt know what it meant [15:56] It won't upgrade your distro unless you have messed with your sources.list or sources.list.d/* [15:56] pull one thread and look what happens [15:56] It will also tell you what extra stuff it is pulling in and prompt you first. If you see a huge number of packages, then perhaps stop [15:56] can i tell it to just install the new dependancies [15:57] Yes - by using "dist-upgrade" [15:57] Do i know this update will not screw with my DB and things [15:57] rbasak: will the eventually come forward or do you have to prompt them? [15:58] resno: it depends, but normally I wouldn't expect anything to change over time. If "upgrade" can't do a package, it won't magically be able to in the future unless that package changes again. [15:58] LargePrime: of course it'll screw with your DB. That's the whole point of upgrading. You do have backups, right? [15:58] LargePrime: you can also run it with --simular [15:59] simulate* [15:59] And you know how to restore the backups and have tested this, right/? [16:03] welp, i should had some backups before doing that updat lol [16:08] rbasak: I thank you [16:08] also not need backups [16:08] you dont? === acidflash_ is now known as acidflash [16:16] roaksoax: https://code.launchpad.net/~zulcss/keystone/keystone-refreshed/+merge/170639 [16:18] jamespage: when we upgrade a charm, the config-changed hook is called, but not the install hook right? [16:18] roaksoax, no - the upgrade-charm hook is called [16:19] roaksoax, or do you mean when we upgrade openstack? [16:20] jamespage: err, sorry i guess i should have said when we change the config and call upgrade-charm [16:21] roaksoax, ok - so for openstack series upgrades - yes - its just the config-changed hook === james_ is now known as Guest67947 [16:21] as the change in openstack-origin triggers the upgrade (at least it does right now) [16:22] jamespage: ok cool! thanks! [16:22] resno: I ment that I did not NEED them for this upgrade [16:31] jamespage/roaksoax: (sorry for being a pain in the ass today) https://code.launchpad.net/~zulcss/cinder/cinder-refresh/+merge/170663 [16:40] zul, +1 on that cinder refresh [16:40] zul, just hit a great bug backporting ovs to 12.04 [16:40] nothing todo with 12.04 but on my saucy laptop the pid of init is not 1 [16:41] huh? [16:41] zul, its the upstart user thingy [16:41] I get two init processes - one for the system, and one for me [16:41] oh crappers [16:42] the schroots see my user based one [16:42] openvswitch has a test which expects the pid of the init process to = 1 [16:42] #bag [16:42] #bang [16:42] whoops! ;) [16:42] good idea to get it fixed and contribute back upstream === matsubara is now known as matsubara-lunch [17:11] zul: http://paste.ubuntu.com/5784315/ [17:11] opinion sought [17:11] I'm checking for a process named 'init' rather than pid 1 [17:12] wheres init.pid come from? [17:12] zul: are we still supporting essex? [17:12] roaksoax: yes [17:12] for security updates [17:13] ack [17:13] jamespage: so charms should continue to support essex then? [17:13] roaksoax, YES [17:13] yes [17:13] sorry - accidental caps [17:13] lol :) [17:14] jamespage: but yeah looks good [17:15] adam_g_: so for the templating, templates/ should be duplicated across all the releases that hve the same file, but only added if one release config file has different config? [17:17] roaksoax: well no i think essex uses python-gflags [17:17] roaksoax, templates/ is one that has stayed the same across all releases [17:18] roaksoax, templates// is one that has changed format since [17:19] adam_g_: right os if essex changed format, but not the rest, I still have to have templates/{essex,folsom,grizzly,havana> right? [17:20] roaksoax, if the essex format is different, but all releases since folsom use the same format, you can have templates// + templates//. all releases after folsom will fall back to the most recent templates/ [17:21] is there a place that lists PPAs? I am looking for MySQL 5.6 [17:21] roaksoax, this approach may not be the most intuitive and im happy to change it if theres an easier way [17:23] adam_g_: that makes sense to me [17:24] adam_g_: but maybe, since we already now what release to install, it would be worth to add the release to the context and use it? [17:24] roaksoax, use it for what? [17:28] adam_g_: in the templates, but I guess that's just adding more complication [17:30] roaksoax, ya, by the time those are called the template loader is already constructed. my idea was to have the context generators be as generic and re-usable as possible. [17:31] roaksoax, the logic of supporting multiple release templates is isolated in the renderer + template classes, and each charms templates can be responisble for taking what they need from the generic contexts [17:54] adam_g_: ok so does this make sense? http://paste.ubuntu.com/5784442/ === ^rcaskey is now known as robertj === marlinc|away is now known as marlinc [18:04] roaksoax, some indentation issues in the paste but generally looks goo [18:04] roaksoax, oh and +#!/bin/bash [18:04] ;) [18:06] adam_g_: lol:) didn't even notice :) === matsubara-lunch is now known as matsubara [18:14] adam_g_/roaksoax: https://code.launchpad.net/~zulcss/cinder/missing-dep/+merge/170677 [18:19] adam_g_: ideas of why this might happen? http://pastebin.ubuntu.com/5784524/ the relation is not being set in db_joined or something? [18:21] roaksoax, its looking in your charms config for database + database-user, but the config.yaml doesn't contain those [18:22] adam_g_: ohh i see! === marlinc is now known as marlinc|away [18:22] roaksoax, up until now, each charm had their own database config value (cinder-db, nova-db, etc). i'd like to standardize on just 'database' + 'database-user' across all of them. [18:24] adam_g_: ok, so i should just change glance-db to database and so on [18:24] this won't create any issues on upgrade? [18:24] roaksoax, im not sure TBH. [18:26] adam_g_: well I guess if relations are re-generated on upgrade, it shouldn't [18:35] Hi, is it meant to be that '$ w' will not show users that are currently rsync'ing to a server? [18:35] (Because it doesn't). [18:37] sw: I believe the intention of 'w' is to show which users are on which terminals [18:37] sw: no terminal would be used for rsync, so no 'w' entry [18:38] sarnold: Oh, woops. Is there a command like that but would show anything connected to the server (whether by rsync, SSH or whatnot)? [18:39] sw: I'm not sure. 'last' may show you want you want; 'netstat -ntup' shows 'active' network connections and their processes... [18:40] adam_g_: http://paste.ubuntu.com/5784572/ [18:41] adam_g_: so it seems that the relation data been passed from mysql to glance is passed after it tries to render the config, resulting in a config without sql_connection being configured [18:41] adam_g_: any ideas? [18:43] adam_g_: http://paste.ubuntu.com/5784579/ [18:45] What would be the cron line for 45 minutes past midnight? Is '45 00 * * *' correct? [18:45] sarnold: Thanks! [18:46] roaksoax, ya, i was expecting we might hit this. my idea is to raise some new exception from context.py,like OSContextIncomplete, that the hook calling config.write() can catch and exit 0, similar to how we were checking [[ -n "$db_host ]] && exit 0 [18:46] roaksoax, if you wanna have a go at that, feel free [18:46] adam_g_: ok cool, will look into it [18:47] sw: yes, that looks correct [18:58] Having some very odd problems with NFS mounts this week... [18:59] ...unable to write to them, and "ls" gives errors. [19:00] what error? [19:03] The "mount" cmd says: 10.1.0.1:/webapp2qa/temp on /webapp/temp type nfs (rw,vers=3,proto=udp,soft,bg,intr,addr=10.1.0.1) [19:03] But $ ls /webapp/temp gives: ls: reading directory /app/temp: Input/output error [19:04] sjbnz: tried to umount it and remount it? [19:04] anything in dmesg? [19:05] Happening since Monday. Concerned that "unattented-upgrades" may have bumped us onto a bade version... [19:06] Yup, getting some: [48534.516447] nfs: server 10.1.0.1 not responding, timed out [19:07] ...but other mounts are working fine. [19:07] sjbnz: to same server? [19:07] on that server? or other servers? [19:08] NFS server is an "i5" (AS400!) - this is the 13th of 13 mounts [19:10] Have 3 boxes with the problems - two prod one test. All working fine until Monday [19:10] can you leave an 'mtr 10.1.0.1" running for a while and check the stats? === hatch_ is now known as hatch === stevechch is now known as sjbnz [19:30] jamespage, FYI this is killing ceph 0.63.1 on precise for me ATM, at least with nova's default long hostnames [19:30] ? [19:31] adam_g_: so something like this: http://paste.ubuntu.com/5784732/ [19:32] Just changed IRC clients, so I have missed the lat 10 mins [19:32] sjbnz: nothing directed your way :) [19:33] mtr to my NFS server give avg:0.5, worst:9 [19:33] seconds? [19:33] ah, I think it reports ms by default. okay. good. :) hehe. [19:34] ms probably (whatever it is that mtr talks :-) It's on the same network, so it SHOULD be fast [19:35] Looking y'day saw lots of those timeout messages, but not really matching with our problem... [19:35] ...which is specific to one mount. [19:36] adam_g_: nevermind that's wrong :) [19:38] Actaully, some other mounts are also giving "funny" responce to an ls... [19:38] ls: reading directory /webapp/proofimages: Input/output error [19:39] sjbnz: tried to remount it? [19:39] sjbnz: and are there any new kernels installed lately+ [19:39] ? [19:40] nfs is in the kernel (usually) [19:42] yup, although the boxes hadn't been rebooted to have them take effect. [19:45] On 15/06: Upgrade: linux-server:amd64 (3.2.0.45.54, 3.2.0.48.58) [19:45] ...form /var/log/apt/history.log [19:46] but y'day 20/06: Remove: linux-server:amd64 (3.2.0.48.58), [19:48] and now uname gives: Linux acme.example.org 3.2.0-48-generic [19:49] (after I spent some time trying to reverse the updates) [20:00] thx. Signing off now. I'm blaming the "mainframe" i5 guys seing reverting this kernel at the Ubuntu cleint end didn't help. [20:11] My /etc/hosts says "127.0.1.1 server1. Should this be my real ip? [20:11] How can i set that from DHCP? [20:14] halvors: no, thats ok [20:14] halvors, I think is not necessary using DHCP [20:14] and yeah, it's ok, it's default value [20:17] adam_g_, whats killing ceph? [20:17] But my other slave dns server isn't able to connect to it. [20:17] It connect's using the 127.0.1.1 ip :( [20:18] jamespage, well, first i thought it was a segfault in liblockfile1 stopping ceph-mon from starting. after fixing that, ceph-mon still isn't starting. [20:18] jamespage, failed to bind the UNIX domain socket to '/var/run/ceph/ceph-mon.server-78f8f3a8-d5c5-46f4-9e65-cfbf1deeff87.asok' [20:18] jamespage, trying to make sense of the upstart scripts to try to debug why its not starting [20:19] adam_g_, whats the length of that path? linux has a limit for opening sockets [20:19] 113 chars or something [20:19] jamespage, oh jeez, thats probably it [20:19] jamespage, the long hostname was causing the segfaul in liblockfile [20:19] adam_g_, somewhere in the depths its probably throwing a ENOENT or suchluck [20:20] * jamespage tries to remember [20:20] adam_g_, 103 chars [20:21] http://openvswitch.org/pipermail/dev/2013-January/024528.html [20:21] its possible to hack around it but its awkward [20:21] hmm [20:21] that path is only 72 chars [20:26] adam_g_, odd [20:27] jamespage, in other news i get a ugly core dump if i run 'ceph-mon --cluster=ceph -i server-78f8f3a8-d5c5-46f4-9e65-cfbf1deeff87 -f' manually or with a shorter id [20:27] pastebinit [20:27] ? [20:27] adam_g_, are you using charms to deploy? [20:30] jamespage, yes. up until bootstrap_monitor_cluster() [20:31] adam_g_, can I see directly? [20:31] jamespage, http://paste.ubuntu.com/5784893/ [20:31] im working locally here [20:32] zerick: What to do for setting the DHCP assigned address to /etc/hosts? [20:33] halvors, well, you have to edit your /etc/network/interfaces [20:34] halvors, set your ethX interface to "dhcp" [20:35] indeed, DHCP is not going to write to /etc/hosts, that file you can edit manually if you wish [20:35] zerick: I did. [20:35] Ah ok. [20:36] halvors, all concerning to IP configuration goes on /etc/network/interfaces [20:39] Anyway. anyone who knows how to set in /etc/network/interfaces that i want to receive a IPv6 prefix using DHCPv6 Prefix Delegation? [20:46] adam_g_, hmm - looks like cuttlefish introduced some charm incompatible changes [20:46] looking now [20:50] jamespage, ah [20:51] adam_g_, it would appear that /var/run/ceph and the mon directory itself must be present now [20:51] for the ceph-mon mkfs to actually work [20:51] /win 4 [20:52] I am running Ubuntu server in VMWare. Is there a way to make it so that the TTY is resizable? [20:52] I know that's probably baked into the kernel, but thought that there might be some fancy new thing for a resizable TTY. :) [20:53] jamespage, ill see if getting that created as part of install hook helps [20:54] adam_g_: i don't think we can add a OSContextIncomplete error class and catch the error elsewhere [20:54] adam_g_: at least catch that error in particular [20:55] roaksoax, why not? [20:55] adam_g_: so i've been trying this: two.py defines Error and raises it. in one.py I do try: except Error: and the error is not being caught [20:56] roaksoax, in one, do you 'from two import Error' ? [20:56] adam_g_: yeah! [20:57] I'm going to change the approach and see what happens [20:57] roaksoax, im not sure, i dont have it in front of me but im sure it can be done [21:00] adam_g_, http://paste.ubuntu.com/5784969/ [21:00] should sort you out [21:01] jamespage, cool. testing the same thing right now. thanks [21:03] ceph-havana/1: [21:03] agent-state: started [21:03] jamespage, ^ \o/ thanks [21:04] jamespage, ill propose something to the charm [21:13] adam_g_: ok IO'm hoping that this works: http://paste.ubuntu.com/5784994/ [21:15] roaksoax, thats what i had in mind. === adam_g_ is now known as adam_g [21:19] logrotate question: can I have the postrotate command in the global area of the /etc/logrotate.d/whatever config file? [21:32] hallyn: hey there. [21:32] hallyn: ever seen lxc-create hang ? [21:33] hallyn: http://paste.ubuntu.com/5785045/ - want a bug? I've left it for 15m plus before ^C'ing it. [21:46] Hi. Am I in the right channel for support on cloud images? 'Cause I managed to get a VM with the nightly saucy build running, but I don't even have a clue how to log in, that is, either I need user/pass (user being ubuntu, I guess) or a way to drop an ssh public key somewhere. [21:53] Baribal_: this may be useful: http://ubuntu-smoser.blogspot.co.uk/2013/02/using-ubuntu-cloud-images-without-cloud.html === hggdh_ is now known as hggdh === wedgwood is now known as wedgwood_away