[00:01] <_donvito> root@dreambox:/var_init/script> chmod -R 777 inadyn_script.sh [00:01] <_donvito> chmod: inadyn_script.sh: Read-only file system [00:01] <_donvito> how is it posible_ [00:02] <_donvito> how can i change rights === ejv_ is now known as ejv === phunyguy_ is now known as phunyguy [03:37] Can someone help me with "...error attempting to write lower page..."? [03:41] /n [03:41] ugh === Laogeodritt| is now known as Laogeodritt [06:54] I have a Ubuntu 12.04 server, configured in /etc/network/interfaces to use a static IP, that seems to at random decide to get a DHCP ip, and as such finds itself offline. [06:55] anyone seen something like it or have any ideas? I see dhclient is installed, but when trying to remove it, was also told I had to remove a ton of other stuff [07:35] ttx, thanks [07:35] ttx, glance now using milestone-proposed for CI builds; heat not in the lab or the archive for raring [07:35] probably need to fix that before we release === smb` is now known as smb [08:27] jamespage: heat is still inincubation, so you don't really need to "fix" that before O [08:28] err. S [08:58] kirkland: have you seen http://askubuntu.com/questions/244103/byobu-open-different-tabs-in-different-windows-of-the-same-session ? [08:59] ttx, yeah - the PTL did request that it be enabled in CI and we agreed to defer until it landed in debian experimental [08:59] but that never actually happened [08:59] Evidently it's a popular question [08:59] .... [09:04] jamespage: horizon milestone-proposed just cut [09:04] ttx:ack [09:39] rbasak, looking at your mongodb arm stuff now [09:39] Thanks [09:40] I do want to get more review, but I'm confident that this is an improvement over what's in the archive at the moment === neunon is now known as snoonan [09:43] hallyn: I opened #1157626 re my raring virsh migrate issues after removing openvswitch from the equation [09:49] rbasak, jon's original patch had some fixes in src/mongo/bson/util/atomic_int.h as well - any reason those where dropped? [09:50] jamespage: I'm not sure what jon's original patch was against. But in our version, atomic_int.h contains everything we need for arm using gcc builtins if they are defined, which they are for us. So the arm specific bits aren't needed. [09:50] rbasak, hmm - #elif defined(__GNUC__) && (defined(__i386__) || defined(__x86_64__)) [09:51] #else [09:51] #error unspported arch [09:51] that said - I'm not sure that code is actually used anywhere... [09:51] it uses mfence and xadd [09:51] jamespage: only Nova, Keystone left (+ Ceilometer) [09:51] ttx, **\o/** [09:52] jamespage: I'm generally not happy with the conditionals in atomic_intrinsics.h - there are paths that will lead to really broken behaviour and corruption if neither __i386__ nor __arm__ are defined for example - it's not robust [09:52] But I haven't reworked the whole thing so I left it mostly as-is for now [09:52] rbasak, tbh its a hack to add arm support right now [09:52] Yes [09:53] So for now I just did the minimal changes that will fix the issue so that I'm confident that I'm not introducing any new problems [09:53] rbasak, +1 for that approach [09:53] rbasak, I think atomic_intrinsics.h could be reworked to use gcc builtins pretty much everywhere [09:54] jamespage: indeed [09:54] jamespage: the risk there is that I break Intel, of course :) [09:54] And so I really didn't want to go there right now - certainly not without upstream approval [09:55] rbasak, actually I see what you mean in atomic_int.h [09:55] jamespage: is it good idea to wait 4 few more days if i need to test grizzly with cloud archive? [09:55] it looks fine as arm will have __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4 [09:55] Yes, exactly [09:55] koolhead17, actually if you want to test right now the grizzly-trunk-testing PPA is pretty stable [09:56] I've been using it for the last week and fixing up and issue I found [09:56] So effectively atomic_int.h has arm support upstream already [09:56] It's just atomic_intrinsics.h that needed fixing up [09:56] its tracking the milestone-proposed branches so its pretty much grizzly rc [09:56] So for that I fixed up jon's patch minimally [09:56] rbasak, +1 [09:56] jamespage: thinking of updating doc for grizzly :) [09:56] koolhead17, sounds like a plan! [09:57] Anywhere I did need to touch though, rather than writing new arm-specific low level stuff (kuser or arm assembly) I used gcc builtins. I'm more confident using them [09:58] The bits I needed to touch were where the return value semantics of jon's patch didn't exactly match the intel assembly [09:58] rbasak, excitingly mongodb 2.4.0 just got released [09:58] jamespage: yeah. Bit annoyed about that :) [09:58] jamespage: I take it we're going to be stuck on 2.2 for raring due to FF/ [09:58] just considering whether we go for a FFe [09:58] ? [09:58] well [09:58] I think it might be worth an FFe [09:58] I actually might push for it [09:58] Most mongo users will want 2.4 I think [09:59] I think they might consider 2.2 too old to be worth using [09:59] the default JS engine is now v8 rather than the obsolete embedded spidermonkey engine [09:59] And we're only a few days late [09:59] rbasak, ack [09:59] jamespage: so let me work on it with this repo and once grizzly releases i will change the repo and try it on stable main cloud archive [10:00] adam_g, zul: something is wonky with the glance package build in the lab [10:00] it takes neally 2 hours [10:00] something dies at the end of the build and it hits a timeout in sbuild [10:01] ttx, hows grizzly looking generally? [10:01] jamespage: everything looks under control. not so many showstoppers [10:02] I have a harder time following all the projects to get a more precise idea, so i have to rely on the PTLs reports [10:02] buit they look on top of incoming bugs and on the right track to fix all RC ones [10:20] rbasak, OK - I uploaded your arm fixes + a couple of minor tweaks of my own [10:20] looking at 2.4 now [10:33] Daviey, hmm [10:33] http://paste.ubuntu.com/5630788/ [10:33] Thanks jamespage! [10:33] http://paste.ubuntu.com/5630789/ [10:34] the code that references openssl in the mongodb codebase is actually apache-2.0 licensed [10:34] but that code is used by AGPL licensed code [10:40] ttx, great - I've been recycling my trunk testing environment 3-4 times a day and I've only found a few minor niggles [10:40] (that cover all of openstack aside from ceilometer - working on getting that incorporated as well) [10:41] * jamespage melts his laptop building mongodb 2.4 [11:48] zul, yolanda: https://code.launchpad.net/~james-page/horizon/grizzly-rc1/+merge/154340 [11:52] jamespage: how do you re-compress the static js? [11:52] zul, see the README in debian/ [11:53] jamespage: ack [11:53] zul, its a dumb MP as its so large [11:53] due to the asset regeneration [11:53] jamespage: yeah anywyas +1 [11:54] oh, i was just approving it also :) [11:55] zul, are you OK to pickup glance and swift rc's? [11:56] jamespage: always [11:56] zul, ta [11:56] zul, I'd already done the work for horizon last week so thought I would help out with that one [11:56] jamespage: no worries thanks [12:35] hey I'm trying to copy/move some data and with dd if=/dev/zero of=/dump.dd I get 350M/s write, but with cp, mv or dd from /input.dd to /output.dd I always get 99% io and only 5-10 M/s read and write [12:36] any idea how to optimize it? [12:36] hi what's the best way to keep SSH sessions to an Ubuntu server alive? [12:36] TCPKeepAlive yes is set on the server [12:40] sw, heh? [12:41] by default, sessions are always alive [12:41] sounds more like a nat issue [12:41] patdk-wk: SSH connections just stop responding if left idle for a couple of minutes [12:41] are you behind nat? [12:41] patdk-wk: no [12:41] then a tcp connection, even without any keepalives, will last forever [12:42] keepalives are only useful for two things, detecting dead connections, or keeping nat from timing out [12:42] that'd odd, have a handful of SSH connections that we leave open, no issues with any of our Debian or Fedora boxes [12:42] and since, ClientAliveInterval defaults to 0, tcpkeepalive=yes won't do anything [12:42] we have 6 Ubuntu servers, it happens on them all [12:42] at the same location as the others, same connection etc. [12:43] sw, you have a firewall on the ubuntu server? [12:43] patdk-wk: no === Pici` is now known as Pici [12:43] this doesn't happen on any of my ubuntu servers [12:43] sw: on the client in ~/.ssh/config insert "ServerAliveInterval 60" [12:43] in fact, I can hibernate my laptop, disconnect it [12:43] and then turn it on the next day, and all my ssh sessions resume without issue, no drops [12:44] keepalive would have killed those [12:44] sw: the best thing to do is to run tcpdump to capture all relevant network traffic and then have someone who understands TCP to interpret it. That will pin down the cause. [12:44] mardraum: I'll give that a shot, thanks [12:44] normally you will need a tcpdump from both sides [12:45] sw: I usually need it behind certain cisco stateful "firewalls" (eg CBAC) [12:45] patdk-wk: sure, but a tcpdump on the server side will be able to either eliminate the server or implicate it. [12:49] jamespage/yolanda: https://code.launchpad.net/~zulcss/glance/rc1/+merge/154355 [12:51] easy one [12:51] zul, +1 [12:51] jamespage: thanks [13:00] jamespage/yolanda: im just doing a local build before uploading it to raring [13:00] zul, ack [13:11] zul, Daviey, Any news on the review/sponsoring frontier? [13:11] smb: xen? [13:11] zul, yup [13:11] smb: will do so tonight just almost caught up [13:12] s/tonight/this morningg [13:12] zul, Ok, just asking since Daviey has been asking for a debdiff yesterday. Not sure *exactly* why [13:25] what state are maas nodes supposed to be in, when available but offline ? [13:27] yolanda, I'm testing a potential upgrade to mongodb 2.4.0 for raring - if I stick it in a ppa can you check it out with ceilometer for me? [13:28] jamespage/yolanda: https://code.launchpad.net/~zulcss/swift/rc1/+merge/154366 [13:30] jamespage, sure [13:31] yolanda, thanks [13:33] yolanda, ppa:james-page/mongodb2.2 [13:34] just uploading - may take a few to build [13:34] ok, let me try it [13:34] yolanda, that will be on raring btw [13:34] my openstack is running precise :( [13:34] can it be possible to test that on precise? [13:45] smb: looking now [13:46] zul, ack === wedgwood_away is now known as wedgwood === frankban_ is now known as frankban [14:17] i am stuck, can someone please help me ? I need to get past juju bootstrap: 229 INFO Bootstrapping environment 'maas' (origin: distro type: maas) 2013-03-20 ,880 ERROR No matching node is available. [14:18] i have 7 nodes in state: commissoning, but juju refuses to bootstrap properly [14:26] Daviey: omg, that SRU of maas into precise has been sitting there since January? [14:28] jcastro: welcome to the dark side [14:28] hi === racedo` is now known as racedo === CaptainBirdseye is now known as unreal [16:03] jamespage: nova MP cut [16:04] There once was a man named Daviey [16:04] who needed to do me a favey [16:04] he should fix the DNS [16:04] because it's a mess [16:05] and then our website would be savey [16:05] who's dns? [16:05] ahh yours [16:05] yea [16:05] tgm4883: what's the url? [16:05] histo, basically, http://mythbuntu.org doesn't redirect to http://www.mythbuntu.org [16:05] it's a google site, so it's a DNS fix [16:06] tgm4883: ahh need a @ [16:06] yea [16:12] If that doesn't work, I also have a haiku [16:15] ttx, ack [16:25] 3 [17:10] rbasak: hadn't seen it, just answered [17:10] kirkland: thanks! [17:11] rbasak: np [17:11] rbasak: thanks for the heads up [17:14] kirkland: you can subscribe to tags apparently. Maybe helpful for you? === matsubara is now known as matsubara-lunch [17:39] After the installer asks if I want to activate the "ATA RAID Controller" it takes me to the partition manager. There I only have 3 options, one for iSCSI, one to undo changes, and one to write the changes to disk, as well as a blank line which can be selected but does nothing. [17:40] I've tried the "Undo Changes" option but I go to a purple screen and nothing seems to be happening. [17:42] DavidScherer: a fakeraid controller probably got activated, have you setup raid in your bios ? [17:43] DavidScherer: if not, you probably want to reboot again and choose to not activate ata raid controller, and just use the disks directly (or like setup linux software raid - mdadm) [17:53] xnox: I used the Lenovo Thinkserver easy setup to setup the RAID [17:55] DavidScherer: no idea what that is. I generally find mdadm software raid most reliable, alternatively with hardware raid the drives should appear correctly. If this Lenovo thing is "fakeraid" e.g. chipsets on the motherboard, I'd advice against using it. [17:59] xnox: I'm pretty sure this is hardware raid. [18:00] ok. Sorry, I don't have any knowledge of Thinkservers, maybe somebody else on here does. [18:00] Otherwise http://askubuntu.com is a good place to ask a question about Thinkservers support status / experience of other people. [18:01] adam_g/jamespage: https://code.launchpad.net/~zulcss/nova/rc1/+merge/154453 [18:31] Is there any difference on the server cd vs the regular cd other than a text based installer? [18:33] no gui installed also [18:35] server is more designed to be used without a monitor attached and any changes that need to be done are done with putty [18:42] adam_g: ping https://code.launchpad.net/~zulcss/nova/rc1/+merge/154453 [18:51] hai i am installing apache 2.4.4 on linux and got the following error any help?? "checking for APR... configure: WARNING: APR version 1.4.0 or later is required, found 1.3.8" === matsubara-lunch is now known as matsubara [18:52] any help?? [18:53] hai all i am installing apache 2.4.4 on linux and got the following error any help?? "checking for APR... configure: WARNING: APR version 1.4.0 or later is required, found 1.3.8" [18:53] Newbee: 'apr' is 'apache runtime'. if you're rolling your own apache, you'll probably also need to roll your own apr too. [18:53] how to upgrade apr from command line === JanC_ is now known as JanC === Jikan is now known as Jikai === Jikai is now known as Jikan === Jikan is now known as Jikai === Jikai is now known as Jikan [19:49] smoser: ping [19:49] hey [19:54] kirkland, ^ [20:15] so what would be needed to setup a server to use smarthost if it was originally set to not be smarthost enabled ? === Jikan is now known as Jikai [20:59] jamespage, zul http://people.canonical.com/~agandelman/folsom/nova-2012.2.1+stable-20121212-a99a802e-0ubuntu1.4~cloud0/ [21:00] adam_g, +1 [21:00] adam_g, upload it now and I'll swish it through first thing tomorrow :-) [21:00] adam_g, whats happening with 2012.2.3? [21:00] jamespage, k [21:01] jamespage, NFI what was uploaded ~1 month ago is now totally obsolete. === Jikai is now known as Jikan [21:03] jamespage, i'm happy to put together another batch of uploads, but i worry they'll just sit in queue for another month until they've been trumped by a bunch of security updates [21:03] adam_g, so I see they got accepted 20/02 [21:03] right? [21:04] into proposed that is [21:05] jamespage, one sec. doing to many things at once ((just pushed that nova update to the grizzly-staging ppa :\ ) [21:05] adam_g, oppps [21:09] jamespage, so: originally pushed 2012.2.3 (all projects) into queue. ~2 weeks later they got accepted to proposed. around same day they were accepted, a bunch of security updates were released together, which got released as patches onto 2012.2.1 [21:09] jamespage, i began rebasing packages as new uploads into the queue to include the security fixes.. those are still pending [21:11] jamespage, and have been for some time, since then more security updates come. those in queue (and in proposed) lack some of the security updates, i think. [21:13] adam_g, looks like cinder is pretty much the only thing not superceded by security updates [21:13] * jamespage sighs [21:15] jamespage, yeah.. :| i'd much prefer updates to get rolled out coordinated across all projects. [21:15] adam_g, +1 [21:16] it gets tested upstream like that after all [21:16] adam_g, I think we need to agree some sort of cadence cycle with the SRU team so we don't keep getting bumped like this [21:17] adam_g, lemme talk to Daviey about it tommorow and see what we might cook up [21:17] jamespage, the bottleneck currently is getting things approved into -proposed, whereas last cycle it was getting things into queue. i guess we've made some progress :) [21:18] adam_g, OK - leave it with me [21:20] I'm not sure the cinder in -proposed has the patch that is in -security. please verify [21:31] jdstrand, correct [21:42] hello! [21:42] I've just setup a VM with ubuntu-server 12.10 [21:43] Does anyone know how to install php5.3.3 instead of the current version? [21:43] I'm trying to replicate a production system [21:44] apt-cache showpkg php5 only shows 5.4.6-1ubuntu1.2 and 5.4.6-1ubuntu1 [21:44] IgorNym: there's a pile of other versions in the repository: http://us.archive.ubuntu.com/ubuntu/pool/main/p/php5/ [21:44] .. though I don't see a 5.3.3 [21:44] Cool! 5.3.x would be fine. [21:45] Noob question - how to install from there, once I've chosen my .deb? [21:45] IgorNym: you can wget the package and dpkg -i to install -- it will leave all the dependency resolution to you, that way, which will be .. painful. [21:46] well, I guess there are worse things than life without apt. but still. :) [21:46] IgorNym: this page will give you a feeling of which version to pick for the version you're trying to replicate: https://launchpad.net/ubuntu/+source/php5 [21:47] Cool I'll give that a go [22:07] zul, jamespage this too http://people.canonical.com/~agandelman/folsom/keystone-2012.2.1-0ubuntu1.3~cloud0/ [22:07] adam_g: +1 [22:09] success! [22:09] just needed the right php5-common package from that page. [22:09] Thank sarnold [22:09] *Thanks [22:09] IgorNym: woot :) === cronus__ is now known as cronus [22:49] utlemming: ping [22:50] utlemming: what is http://cloud-images.ubuntu.com/quantal/current/tool-version-info.txt ? [22:51] kirkland: it just shows the changes made in the build system. Its legacy and honestly should just die [22:51] utlemming: okay, cool [22:51] utlemming: next question ... so I'm downloading quantal-server-cloudimg-amd64-disk1.img [22:52] utlemming: I'm going to mount it, edit /etc/overlayroot.conf, unmount, and register my own ami [22:52] utlemming: the best way to do this is ec2-bundle-image? [22:53] kirkland: if you want instance store, yes, that is how I'd do it [22:53] utlemming: for the root disk? [22:53] kirkland: yes [22:53] utlemming: okay -- and if I just want ebs? [22:54] utlemming: i don't want to actually boot the system, which is why I can't use the AWS web interface [22:54] kirkland: install qemu-util, mount the qcow2 image, dd it over, snapshot and register [22:54] utlemming: booting it peturbs the image [22:55] utlemming: "dd it over" meaning, dd it to an ebs volume? [22:55] kirkland: i.e. modprobe nbd; qemu-nbd -c /dev/nbd0 ; dd if=/dev/nbd0 of= bs=1M [22:55] kirkland: right, the ebs volume [22:55] utlemming: oh, interesting [22:56] kirkland: the qcow2 is a complete system with grub2 [22:56] utlemming: k, lemme give that a shot [23:02] utlemming: okay, so is /dev/nbd0 mountable? [23:02] kirkland: /dev/nbd0p1 is [23:02] utlemming: aha [23:04] utlemming: when I create the volume in the AWS ebs interface, what's the right "size"? [23:04] utlemming: I tried 8GiB and it came out a little short, like 7-something === Ursinha-afk is now known as Ursinha === wedgwood is now known as wedgwood_away [23:16] utlemming: hmm, okay, did all of that, created snapshot, created image, tried launching... doesn't appear bootable [23:16] utlemming: did I need to do anything after the dd, and before the snapshot? [23:16] utlemming: I selected Kernel ID: aki-825ea7eb [23:20] utlemming: ah, hang on, I think I know what's wrong [23:20] utlemming: I forgot to add an ephemeral disk, which I need for the overlay to work [23:22] utlemming: hmm, that's not working either [23:22] utlemming: this is strange [23:24] utlemming: http://paste.ubuntu.com/5632736/ [23:34] Hello [23:34] When I try install something on my VPS, I get the following errors: [23:34] locale: Cannot set LC_CTYPE to default locale: No such file or directory [23:34] locale: Cannot set LC_MESSAGES to default locale: No such file or directory [23:34] locale: Cannot set LC_ALL to default locale: No such file or directory [23:35] utlemming: yeah, I'm at a loss at this point [23:35] I did: `sudo locale-gen en_US en_US.UTF-8` and `dpkg-reconfigure locales` but that didn't fix it [23:39] Extreme: if you run 'locale' do you see that your locale settings are only en_US or en_US.UTF-8 ? [23:43] one sec [23:43] This is the output for "locale", http://pastie.org/private/j8xbyltkcoqa57l0wmwsrw [23:44] utlemming: yeah, for some reason, the image created from that snapshot is just not bootable [23:45] Extreme: do you have the en_IN and en_IN.utf8 locales generated? (in the output of locale -a) [23:45] sarnold: I think no, http://pastie.org/private/binvqjdm7z4t0u4x0fl8sw [23:47] Extreme: you'll either have to set your locale to en_US or build the en_IN locales.. [23:48] sarnold: how to do that? Can you point me to an article or something? [23:54] Extreme: edit /etc/default/locale and/or execute "locale-gen en_IN en_IN.UTF-8", respectively [23:56] kerframil: I couldn't find anything in /etc/default/locale file. (I opened with `nano /etc/default/locale) [23:56] Extreme: in older releases, it's /etc/environment. more info: https://help.ubuntu.com/community/Locale [23:57] hello [23:57] I am having 12.04 [23:58] Extreme: I don't know then. perhaps LANG was defined by some other means in your current environment. on my 12.04 system, it's sourced from /etc/default/locale. [23:59] Extreme: in any case, LANG is what needs to be set to make a locale effective, and locale-gen can be used to generate missing locales that you may wish to use. [23:59] I just want to turn off the error messages while installing stuff. Is that possible?