[03:15] Is there anybody instearesting in beaglebone? [03:41] hi, with ubuntu 12.04 and dovecot 1.2.0.19. I created UTF-8 (non-english chars) name directory but it doesn't work. I can not subscribe those folders. how to fix this issue? thanks. [03:47] I think you should find where is the probmlem. you can just use english cahrs and see whether is normal. [03:52] xiaorong, I used english chars. it works fine. are you from China? basically, I created chinese directory name. it doesn't work. english name works fine. [03:53] xiaorong, I used "maildirmake.dovecot home/vmail/$domain/$user/.". in evolution, I tried to subscribe to those directories. it doesn't work. [03:56] yes,i from china,Can you find something useful in the log [03:58] gave me some time.i will find how chinese people do for you problem. [03:59] xiaorong, not really. after I failed to subscribe those directories from evolution client, I manually added one directory to subscriptions file in the server. then the dovecot log shows one line: " Removing invalid entry: .已发送". that's all. [03:59] xiaorong, thanks a lot. [03:59] I have a 12.04.2 install. apt-get upgrade wants to install linux-generic-lts-quantal, but also has linux-image-generic-lts-quantal "kept back". Can someone clear up what's going on? [04:09] Hello all,Have any one upgraded 10.04 server to 12.04 using Alternate CD-ROM upgrade? [04:10] While trying to upgrade using CD-ROM, the system turns to be ubootable, it starts but a blank dot only appears === sgo11 is now known as sgo11_away [04:26] Hello Guys,could some body suggest me Alternate CD-ROM updgade for ubuntu server 10.04 to 12.04 works or not, === thumper is now known as thumper-afk [05:15] Riddell: For kcron, changelog needs fixed, it's got the issue with debhelper/compat version, it also has the -dbg depends issue, the first line of debian/copyright needs checking. Accepted. [05:15] Oops, wrong channel. [05:41] how can i setup a ftp account for a specific directory? === smb` is now known as smb [07:48] Good Morning Serverites! [07:51] Daviey: dum dum dum duuuuuum [08:28] Am hoping someone can help me. I have an issue with intermediate cert and apache ssl config - IE and chrome work fine but firefox is missing the intermediate cert. Looks like IE and Chrome are following the Authority Info Access to get the intermediate CRT. I installed the intermediate cert bundle into /usr/local/share/ca-certificates and ran update-ca-certificates and it's showing in /etc/ssl/certs/. On the apache2 end I have specified the interme === thumper-afk is now known as thumper === Will_ is now known as Guest76793 [10:48] hey guys i have made a bootable usb and i want to try and run memtest is anyone encountering an error message when trying to run it something about being unable to boot because an old ram disk [10:48] zul, adam_g: http://people.canonical.com/~jamespage/ca/havana/ [10:48] ceph update for havana CA === sgo11_away is now known as sgo11 [11:14] jamespage: +1 [11:15] jamespage: i need to find out why xen failed in the cloud archive [12:45] zul, thanks [12:45] yes indeed - looked like missing deps? [12:45] jamespage: nah ftbfs [12:49] jamespage: xen debian/control has something that is not in precise so kind of missing deps === fwereade__ is now known as fwereade [12:58] zul, What xen ftbs (the Precise version)? And where was the fail again? [12:58] 4.2.1 [12:58] That would be r or s [12:59] smb: its failing because it has "Built-Using" in xen-utils-4.2 part of the debian/control which precise debhelper doesnt recognize [13:00] Question would be what is pulling in the -4.2 dependency as Precise should be 4.1 [13:01] smb: im buidling 4.2 for precise in the cloud archive since its a dependency of qemu-kvm [13:02] zul, Hm, not the other way round? [13:02] smb: libxen-dev is a dependency [13:02] zul, Fun... [13:03] smb: not really [13:03] zul, That was meant a bit more sarcastic than it may have sounded :) [13:04] smb: mine was equally sarcastic [13:04] We need sarcastic tags then. :) [13:04] smb: indeed [13:06] jamespage: trade ya? http://people.canonical.com/~chucks/ca/ [13:45] jamespage: http://people.canonical.com/~chucks/ca/ [13:46] zul, no change? [13:46] jamespage: nope removed a line in the debian/control [13:47] zul, no - its still has the same error as last time i looked [13:47] see PM [13:47] jamespage: missed PM had to reboot [13:47] zul, new PM [13:52] jamespage: fixed === wedgwood_away is now known as wedgwood [14:04] zul, does it build against staging now? [14:05] if so +1 [14:05] jamespage: it does [14:05] thanks [14:12] adam_g, that ceph charm issue you raised a MP for is due to a change in behvaiour in ceph-disk-prepare [14:12] it used to always zap the disk - it does not anymore! === Aww is now known as EvilAww [14:29] hallyn: hey, just had this one time failure in lxc tests, http://pastebin.ubuntu.com/5795530/ the next run passed though. [14:30] hallyn: i could watch it for a couple of more runs and see if i need to report a bug [14:30] psivaa: thanks [14:31] * hallyn looks for an obvious indicator [14:34] my guess is a transient archive error, but the logs don't show that [14:46] hallyn: ok, ill keep watching this job then [14:46] hallyn: hey, I was noticing in that lxc test that it pulls the lxc package source and rebuilds the package, but this seemed to be only needed until the version we have now landed in raring. Can that bit be pulled out now? [14:50] what's the lp uri again? [14:50] http;//launchpad.net ? [14:50] the version i have in +junk/lxc-test just grabs the bzr source to get the api python test, but doesn't build it [14:50] Daviey: no, for the utah server tests :) [14:50] :-) [14:51] Anyone else getting vagrant errors? http://paste.ubuntu.com/5795698/ [14:51] i get this for the 32, 64 of precise. [14:55] plars: yeah looks like that can be dropped. [15:00] hi all === Aww is now known as EvilAww [15:19] psivaa: plars : when i run sudo utah -r lp:ubuntu-test-cases/server/runlists/lxc.run by hand it locks up my terminal and after ctrl-c i have to /usr/bin/reset. is that expected? [15:23] hallyn: i have not tried with lxc tests but when i ctrl-c some other utah tests i did not have to run reset [15:23] Daviey, or anyone else that owuld know this ... [15:23] i'm interested in getting fix for bug 1187429 into ubuntu [15:24] debian now has a fix, via update of python-requests from 1.2.0-2 to 1.2.3-1 [15:24] (in unstable) [15:24] that should/would normally just flow into ubuntu, right? [15:24] im trying to set ssh to disconnect user(s) after 30 secs and its not working. I have ClientAliveInterval 30 ClientAliveCountMax 3 [15:24] (as ubuntu's version was just 1.2.0-2) [15:25] ie, guess i'm asking when / how often the auto-sync bot runs [15:38] hi all. any idea what 'IPv6 addrconf: prefix with wrong length 48' means? it shows up in this server's kernel log. I found some links on the net that told me to set net.ipv6.conf.all.{accept_ra|autoconf} to 0 to fix this, but apparently it didn't help. server's on 12.04.2 LTS with kernel 3.2.0-43-generic [15:44] zul, how much does kvm/qemu get in the way of block io for a vm? [15:44] jamespage: no idea ... hallyn? [15:46] jamespage: well qemu emulates the hardware, so there it gets in the way [15:47] hallyn, but the layer is quite thin when a vm device is backed by an actual block device right? [15:47] plars: actually i think we may have problems with api incompatibility if we don't build [15:47] I think it should be rather thin with virtio [15:48] jamespage: yes [15:48] hallyn: using virtio or other drivers? [15:48] jamespage: why are you asking? Is there some perf # I can hand you? [15:48] hallyn, I was considering the performance charateristics of ceph on bare metal and when containerized [15:49] conatinerized should have zero impact [15:49] hallyn, and when virtualized under kvm [15:49] hallyn, and you can directly pass through a block device to a container right? [15:50] yup [15:50] hmmm [15:51] jamespage: I think a perf # woudl be very interesting [15:51] hallyn, agreed [15:51] i don't think there is an io path htat doesn't include qemu in the middle at least a bit [15:52] so there'll be some data copying for naught - i *think*. [16:25] plars: ~. [16:25] sorry, net hiccough [16:25] plars: the utah tests don't get run under < raring right? [16:26] (else i need to add a check to build on < raring) [16:27] hallyn: no, not at the moment. should be safe there [16:27] plars: thanks [16:27] hallyn: we do precise tests, but those use the old usit stuff [16:35] I have a 12.04.2 install. apt-get upgrade wants to install linux-generic-lts-quantal, but also has linux-image-generic-lts-quantal "kept back". Can someone clear up what's going on? [16:35] Why are some kernel packages set to be kept back, while others aren't? === matsubara is now known as matsubara-lunch [16:37] shodan45: kernels get bumped when you do dist-upgrade and not just upgrade [16:38] genii: right, I mostly understand that [16:42] genii: http://paste.ubuntu.com/5795976/ [16:43] shodan45: is uname-r reporting 3.5.0-27-generic as your current kernel? This is what http://packages.ubuntu.com/precise-updates/linux-image-generic-lts-quantal reports as a depends [16:44] genii: 3.5.0-23-generic [16:44] shodan45: So then it's held back because your current kernel is -23 and not 27 [16:45] ( would be my guess ) [16:46] genii: my understanding is that I can either stick with the 12.04 kernels (apt-get upgrade) or upgrade to the 12.10 kernels (apt-get dist-upgrade). Is that correct? [16:46] both are considered "LTS" iirc [16:46] shodan45, not quite - lemme dig out the page [16:48] hah. just realized I'm already running the quantal kernel... I guess that came with the ".2" install media? [16:53] looking at the output of dpkg -l | grep -i kernel, I think I need an overview of how ubuntu names & packages their kernel packages [16:56] shodan45: grep linux instead [17:04] ugh. stupid wifi. >_< [17:14] Trying again: hi all. any idea what 'IPv6 addrconf: prefix with wrong length 48' means? it shows up in this server's kernel log. I found some links on the net that told me to set net.ipv6.conf.all.{accept_ra|autoconf} to 0 to fix this, but apparently it didn't help. server's on 12.04.2 LTS with kernel 3.2.0-43-generic [17:15] RoyK: do you need to set that sysctl for the specific interface? [17:16] sarnold: let's see [17:17] how can i setup a ftp account for a specific directory? [17:17] i have vftp [17:17] shankstaBytes: yes, the best way is usually rssh [17:18] should work well with vsftpd [17:19] i actually just install ftpd [17:19] i dont want ssh access [17:20] im trying to give a friend access to upload files [17:20] store [17:20] but i only want him to have access to a specific folder [17:20] is that hard to setup? [17:22] RoyK: any idea? [17:23] shankstaBytes: sftp or rsync with rssh will inhibit ssh access, but allow your friend to use ftp-over-ssh to access the data, thus ensuring security [17:24] just google rssh setups - should be rather trivial [17:24] forget about what I said about vsftpd - not relevant [17:24] ok [17:24] you can chroot vsftpd as well, but the passwords will be sent in cleartext, which isn't good [17:25] old ftpd should not be used, it's not secure [17:25] really, ftp shouldn't be used at all unless it's over an encrypted link [17:26] ssh+rssh ensures security and allows the client to use stuff like filezilla for file transfers, which should be trivial [17:32] Hmm. I have a physical server that has, twice in as many weeks, gone unresponsive and required a hard reset (not network connections, no video output, not even any result with Alt+SysReq+REISUB). I can't find anything remotely relevant in /var/log/syslog or kern.log, anyone have any ideas how I'd start to track something like this down? [17:33] keithzg: use a network console - should tell you what's wrong if the box gets a panic and shuts off local storage [17:35] RoyK: Both times I've had SSH sessions open, and have returned to work the next morning to see nothing more than "Write failed: Broken pipe". [17:35] then setup ssh keepalives on the client [17:35] ServerAliveInterval 5 in $HOME/.ssh/config [17:35] should do [17:36] putty probably has similar settings if windows clients are used [17:37] ssh tends to die without them in certain NAT setups [17:37] Naw, even though the company I work for is a Windows software developer, my desktop is purely Linux :) [17:37] hehe :) [17:37] keepalives are nice and evil [17:37] they keep your nat connections alive [17:37] but if you don't need them, they will kill your connection if you like suspend your laptop [17:38] This is within a local LAN, just a few switches inbetween my work desktop and the server in question. [17:38] the only thing that would kill it then, would be a network outage, or a stateful firewall [17:38] and it would have to be a network outage when something was transferring over the connection [17:45] patdk-wk_, RoyK, have you guys tried mosh yet? [17:46] I'm liking the instant-update still-connected thing just between my laptop and a machine in the same room. :) [17:46] mosh? [17:47] RoyK: it does some clever rsync-like things for terminal state and uses udp. It starts the session via ssh, so no changes needed there. [17:47] Keep meaning to try mosh, but haven't bothered to even figure out if it'd work on my phone (it'd work on my Android tablet, but I'd need to use a different client sadly) [17:48] Wait, why am I even questioning it? It's a commandline utility, *obviously* people have compiled it for the N9, heh. [17:50] I look at mosh, but never really cared to try it [17:50] I do most all of my ssh over vpn and it's nice [17:50] I can suspend my laptop [17:50] drive for a few hours [17:51] open laptop back up, and all my connections are still alive [17:51] patdk-wk_: really? wow [17:51] that's not been my experience :) hehe [17:51] ya, cause of the vpn, no nat issues [17:51] but if you turn on keepalive, that will close them all [17:51] suddenly I wonder if I've got keepalive turned on :) hehe [17:52] keepalive is great for nat issues, so good to turn it on for external facing ssh [17:52] but for over the vpn ones, I keep it off [17:52] nice hints, thanks :) === matsubara-lunch is now known as matsubara [18:01] Hmm. But no other ideas about how I might discover what's causing the computer to hang other than trying to get an SSH session to debug the condition for me? Guess I'll just keep a running tab in Yakuake until the next time it happens, heh. [18:02] kind of doubt it's a hang [18:02] as if it was, it would print something on the screen [18:02] more likely an infiniate loop somewhere or other [18:04] Shouldn't Alt+SysReq+REISUB break that then, though? Although I've noticed that as of more recent *buntu defaults, everything but the final B seems to be disabled. [18:04] Hi everyone, can anyone tell me what the most stable recent ubuntu version is? I have a large time trusting any recent versions after the instability issues I have experienced with ubuntu desktop in the past. [18:04] keithzg, no idea, never had to do it [18:04] define *stable* [18:05] keithzg: a great many sysctl keys are disabled now.. [18:05] patdk-wk_: Speaking as someone who for years ran an NVIDIA TwinView setup . . . I have much experience with my system hanging, heh. Ohhhh NVIDIA. [18:05] keithzg: see /etc/sysctl.d/10-magic-sysrq.conf for details [18:05] keithzg, I have always run with 4 monitors on nvidia, never had a hang [18:06] so nvidia causing you issues, never caused me issues in twinview [18:06] keithzg: (sorry to derail the conversation towards mosh, I hadn't seen the problem description..) [18:07] maybe your issue is something else? like the psu going bad? [18:07] that can completely drive you nuts [18:07] patdk-wk_: Haven't in years now myself (although very lately my card died and I switched to a Radeon) but it used to happen all the time for me. It was indeed fixed with a driver update. [18:07] Anything using compositing ended up being a crapshoot. [18:08] sarnold: aha, thanks for pointing me towards the right config file, I had been wondering. [18:08] :) [18:08] maybe, just for extra goodness, see if you can disable the screen saver in the kernel, so the console won't go black [18:09] Hmm interesting idea. [18:09] shodan45, sorry - here it is - https://wiki.ubuntu.com/Kernel/LTSEnablementStack [18:10] It's on a KVM switch anyways, so no worries about burn-in. [18:15] jamespage: thanks, that helps === ajmitch_ is now known as ajmitch [21:39] adam_g: http://paste.ubuntu.com/5796755/ [21:41] * keithzg finally bothered to install Mosh on his phone. By typing "apt-get install mosh" into the terminal. Could literally have done that a year ago, yeesh! [21:42] roaksoax, i think you can probably get away with not doing that and have the template set paste deploy flavor == keysotne {% if auth_host %} or something similar [21:43] adam_g: yeah that could also be a good option [21:43] roaksoax, its usually just unset by default . worht noting glance is currently the only project that sets that value in its .confs [21:43] yeah [21:44] doesn't really make sense [21:44] eh its definitely better than messing around in the api-paste.ini files [21:45] roaksoax, i just pushed haproxy + hacluster stuff to lp:~gandelman-a/charms/precise/cinder/pyrewrite [21:45] roaksoax, one thing worht noting is changing the restart_map that gets passed to restart_on_change() from a dict to an OrderedDict, so that we can just rely on restart_on_change() to restart api servers + proxies in correct order [21:46] roaksoax, doing so turns the peer hooks into 1 or 2 liners [21:46] need food back in a few === wedgwood is now known as wedgwood_away