=== wedgwood is now known as wedgwood_away
[00:52] <adam_g> jamespage, ive just consolidated all changes to $charm/hooks/lib/* and hopefully resolved all diffs safely and pushed to lp:~openstack-charmers.~openstack-charmers/openstack-charm-helpers/ha-helpers, going to sync across all ha-support branches that use them off
[01:29] <smoser> jcastro, oh. wow. wednesday.
[01:31] <jcastro> I know right
=== holstein_ is now known as holstein
=== ping__ is now known as rhyda
=== rhyda is now known as rhyda-kumisan
=== rhyda-kumisan is now known as klo
=== klo is now known as tukang-cukur-nai
=== tukang-cukur-nai is now known as tukang-sapu
[04:50] <spencer> Hey, anyone got experience with inotify or inotifywait to watch for file system changes?
[06:21] <sicness> Hello. By default dhclient runs with parametr "-1" that meen to be dead if dhcp server will not respone. How to make dhclient "PERSISTENT" at network start?
[06:31] <qman__> you could probably just set the max tries to something astronomical
[06:34] <sicness> qman, thx, I'l do that if wouldn't find right way.
[06:34] <sicness> Anybody know who does start  dhclient usually?
[07:00] <Syria> could someone please tell me how can I completly remove phpmyadmin and SQL ! after deleting the users from phpmyadmin by accedent I cannot log in anymore!
[07:15] <andol> Syria: Generally phpmyadmin doesn't know anything about your users, but merely relays the login to the mysql server.
[07:15] <Syria> andol:  So deleting the mysql-server should solve this problem?
[07:16] <Syria> and re-installing for sure.
[07:19] <andol> Syria: Well, there are a few more packages related, to might want to for all packages matching mysql-server*, as well as running the apt-get remove with the --purge flag. Depending on whatever the removal removes the /var/lib/mysql or not (depends between versions) you might manually need to remove that as well. That is assuming that you are still fine with the option of completly starting over with MySQL.
[07:20] <andol> (As I pointed you to yesterday, there are otherwise more surgical solutions to recreating your mysql users.)
[07:21] <Syria> andol:  Yeah I don't have any important information, so I have to start with "sudo apt-get remove --purge mysql-server" right?
[07:22] <Syria> andol:  The remove the files in " /var/lib/mysql " ?
[07:25] <andol> Syria: Well, mysql-server is really only an (empty) meta package, so as I said, you want to do the same with your other installed mysql-server* packages
[07:27] <Syria> andol:  I see.
[07:27] <Syria> andol:  Then it might be an easier thing that I try to add a "root" user !
=== Masshuu_ is now known as Masshuu
[08:35] <Masshuu> I want to get the contents from either a running screen or tmux. Issue is both require there be a terminal to attach to and i'm trying to do it programmatically where thats not an option.
[08:56] <Piego> hi! I have a problem. I have a linux box perfectly configured whose connection does not work. It is a virtual machine so it cant be a hardware problem. What can it be the problem? The ip address is correctly configured and route too. Dns is fine. The gateway works because other vms pass through it. Im lost :-/
[09:17] <nilli> I'm looking to install a php extension and the latest version I can find is a .deb for ubuntu 7 but my server is ubuntu 8.04.4. is that definitely a bad idea or should I just and see?
[09:18] <nilli> test and see*
[09:24] <apelles> hello.  can someone here help me setup email server - packages needed tweeks and stuff
[09:26] <apelles> is there special time i should visit this irc channel to get hel?
[09:29] <apelles> maybe i should leave my email address, and someone will get back to me: apellesnoel@gmail.com
[09:29] <nilli> that probably won't happen
[09:30] <maxb> apelles: I think your problem is that your question is too general
[09:31] <nilli> apelles: this article could help you get better help http://www.sabi.co.uk/Notes/linuxHelpAsk.html
[09:32] <maxb> Most people are happy to help on IRC with specific probles. However most people don't want to talk people through entire setup procedures
[09:32] <apelles> is there a server i can get help from? or a channel  to visit
[09:33] <apelles> nilli - article is great !  thank you
[09:33] <nilli> I think you'll have a hard time finding any one person to help you through every step, unless you ask the right questions or pay someone for it
[09:34] <maxb> Best start by reading some documentation, having a go, and asking questions about specific things which are unclear or which don't work
=== smb` is now known as smb
[09:41] <apelles> thank you - bye for now
=== dijit_ is now known as dijit
[10:05] <mark__> hello
[10:05] <mark__> i'm seeking help
[10:05] <mark__> i cannot use my apt-get to do anything
[10:06] <mark__> when i want to install anything apt tells me that the package exim4-base is broken
[10:06] <mark__> when i want to remove it (because i don't use it) it tells me it's broken and to reinstall it
[10:06] <mark__> when i reinstall it breaks
[10:06] <mark__> and when i want to remove it, it returns with an error from post-removal
[10:07] <mark__> can anyone shed some light on it?
[10:07] <mark__> i wanted to install postgresql-contrib-9.2 but i get errors about exim4-base
[10:07] <yolanda> hi, can you try with: dpkg -r <<name of package>> ?
[10:08] <mark__> i did. DPKG tells me that exim4-base should be reinstalled
[10:10] <mark__> is there a way to brutally remove exim4, so that apt forgets about it and will not try to do anything with exim4 while installing other packages?
[10:16] <yolanda> dpkg -r exim4-base doesn't do the trick?
[10:24] <mark__> tried that several times, but it tells me to first reinstall exim4-base
[10:24] <mark__> when i try apt-get install --reinstall i get errors
[10:27] <yolanda> and with --force flag?
[10:28] <mark__> apt-get install --reinstall --force exim4-base?
[10:29] <mark__> or dpkg -r --force exim4-base ?
[10:31] <yolanda> dpkg -r --force exim4-base, but not sure
[10:47] <Seveas> mark__, that's the 'dpkg is very confused' warning. Best to follow its advice
[10:47] <Seveas> can you pastebin the foll error that you get when reinstalling?
[10:48] <Seveas> s/foll/full
[10:59] <g0tcha> hey guys, really need some help, whenever i try to install something on my ubuntu server i get this: http://paste.ubuntu.com/5587550/
[11:00] <g0tcha> and the package doesnt get installed.. anyone know how to fix that by any chance?
[11:00] <Seveas> g0tcha, try removing redmin and ruby-rails-3.2
[11:01] <g0tcha> Seveas, i did, i still get the same error when i try to remove a package too
[11:02] <Seveas> g0tcha, pastebin the output of: dpkg -P redmin ruby-rails-3.2 (this will purge the packages completely from your system)
[11:04] <g0tcha> Seveas, here are the results: http://paste.ubuntu.com/5587557/
[11:04] <g0tcha> when i run apt-get install it gives another error
[11:04] <Seveas> g0tcha, typo in the command. redmin should be redmine
[11:06] <g0tcha> Seveas, cool.. i think it worked now!
[11:21] <g0tcha> Seveas, thanks alot, this fixed my issues no doubt.. i just have a question, what does "apt-get autoremove" does exactly?
[11:22] <g0tcha> last time when it suggested to run it, it was going to remove 380MB of data from the server
[11:23] <Seveas> autoremove removes packages that got installed as dependencies, where the dependent has already been removed manually
[11:24] <g0tcha> so its safe to run this from time to time to make sure the system is clean?
[11:24] <Seveas> so if you do 'apt-get install foo' and foo has bar as dependency, 'apt-get remove foo' won't remove bar but a subsequent 'apt-get autoremove' will.
[11:24] <g0tcha> it will Not remove something thats being used by something else, correct?
[11:24] <g0tcha> right.. interesting
[11:24] <g0tcha> thanks for the info
[11:24] <Seveas> it's safe to run if you inspect the list of packages beforehand
[11:27] <freakynl> Hi, I'm running badblocks on 12 SAS disks connected to a LSI SAS controller. iostat shows they only do 7.3MB/s though, any ideas? 7.3 is really low, have system that's identical (hardware wise) that went way over 100MB/s (the first so much data (outter ring) anyways)
[11:56] <freakynl> Hmm dd does ~130MB/s (1mb bs), any guesses why badblocks would be so slow?
[11:56] <freakynl> The only thing I can come up with is that it's related to the USB stick it's running off which is connected to a USB 1.1 port (why would one wire the USB 2.0 the chipset provides if you can just wire 1.1 and forget about it)?
[12:13] <patdk-lap> freaknl, maybe you should actually show the iostat values?
[12:29] <ruben231> hi guys, i have installed apache2 on my ubuntu 10.04 LTS but problem when i reboot it wont start automatically i need to induce it with command before it starts /etc/init.d/apache2 restart <<<<<-------------------any idea how to solve this..?
[12:40] <Daviey> I want a biscuit
[12:40] <g0tcha> would changing the hostname of an ubuntu server break anything?
[12:41] <g0tcha> if its still a fresh install, just need to change the hostname set durring the installation
[12:41] <patdk-lap> ruben231, how exactly did you install apache2?
[12:42] <ruben231> patdk-lap: apt-get install apache2
[12:44] <ruben231> any idea..?
[12:54] <rbasak> jamespage: powerpc ftbfs fix: https://bugs.launchpad.net/ubuntu/+source/ipmitool/+bug/1074443/comments/11
[12:54] <uvirtbot> Launchpad bug 1074443 in dell-poweredge "Please upgrade ipmitool to 1.8.12 for 13.04" [Wishlist,In progress]
[12:56] <jamespage> rbasak, ta
[12:56] <jamespage> I'll review and upload
[12:56] <rbasak> Thanks!
[12:57] <jamespage> rbasak, as the resident non-x86 expert any ideas on the armhf failure here - https://launchpad.net/ubuntu/raring/+source/mongodb/1:2.2.3-0ubuntu1
[12:57]  * rbasak looks
[12:58] <jamespage> rbasak, I'm sure this worked OK for me locally but I think it may have been with 2.2.2 rather than 2.2.3
[12:58] <jamespage> my panda is being temporamental right now
[12:59] <rbasak> jamespage: looks like a preprocessor accident to me. I'll pull the sources and take a look.
[13:01] <Nul0>  Hello. I'm considering formatting my external drive as ZFS and using it as a backup storage system (mainly for its inbuilt deduplication). I'm running Ubuntu 12.04. Does this sound like a good idea? Any words of caution I should bear in mind?
[13:04] <jamespage> rbasak, ta
=== dayangkun_ is now known as dayangkun
[13:13] <zul> hallyn:  ill get 1.0.3 packaged today (i hope)
=== med_ is now known as medberry
=== med_out is now known as med_
=== med_ is now known as Guest44335
=== medberry is now known as med__
=== Guest44335 is now known as medberry
[13:22] <ruben231> hi guys, i have installed apache2 on my ubuntu 10.04 LTS but problem when i reboot it wont start automatically i need to induce it with command before it starts /etc/init.d/apache2 restart <<<<<-------------------any idea how to solve this..?
[13:35] <hallyn> zul: nonsense, you're busy today
[13:39] <zul> hallyn:  im always busy :)
=== cronus is now known as Guest82449
=== cronus_ is now known as cronus
[13:42]  * SpamapS wonders if the server team went out drinking on Google+ last night to make it a proper UDS
[13:43]  * jamespage is still getting over the hangover
[13:43] <ogra_> SpamapS, yeah, and they all ctached virtual ubuflu ... !
[13:44] <Daviey> i'm hanging out of bed
[13:47] <SpamapS> yeah I seem to have caught real ubuflu even w/o the flights and many nights of sleep dep and self-poisoning
[13:47] <ruben231> hi guys, i have installed apache2 on my ubuntu 10.04 LTS but problem when i reboot it wont start automatically i need to induce it with command before it starts /etc/init.d/apache2 restart <<<<<-------------------any idea how to solve this..?
[13:48] <GridCube> hi, what is the alternative to use >> sysctl net.inet.tcp.msl=2500 << in ubuntu, given that there is no /proc/sys/net/inet directory?
[13:49] <SpamapS> ruben231: you should see something in the console/boot messages about why it is failing.
[13:49] <SpamapS> ruben231: any reason you're not using 12.04 ?
[13:49] <SpamapS> GridCube: is it possible that setting has been deprecated?
[13:50] <jamespage> rbasak, uploaded
[13:50] <rbasak> jamespage: thank you!
[13:50] <GridCube> SpamapS, and what would be an alternative to that?
[13:50] <rbasak> jamespage: I'm reproducing your mongodb ftbfs now
[13:50] <jamespage> rbasak, thanks for looking at that - it does take a v long time to build on arm btw
[13:51] <GridCube> http://serverfault.com/questions/129950/can-i-make-tcp-ip-session-to-run-less-than-60-seconds this is from where im taking it
[13:52] <GridCube> its there something like tcp.nolocaltimewait?
[13:54] <rbasak> GridCube: looks like tunables are in /proc/sys/net/ipv4 now. I don't see msl though
[13:55] <rbasak> GridCube: I don't think this is an Ubuntu-specific thing. I think the change probably applies to anyone running a more recent Linux kernel. That might help your Googling perhaps?
[13:55] <GridCube> :( i've been searching for hours
[13:56] <GridCube> well, 'bout two
[13:56] <rbasak> You could find the Linux git changelog message associated with the change. The commit message might list something useful.
[13:56] <rbasak> (eg. what replaces it, or why they removed it)
[13:56] <jamespage> rbasak, w00t - https://launchpad.net/ubuntu/raring/+source/ipmitool/1.8.12-0ubuntu2
[13:56] <hdd1> rbasak: but then the msl is hardcoreded in some sys/*.h library?
[13:57] <rbasak> hdd1: they removed the tunable sysctl and hardcoded it?
[13:57] <rbasak> If they did, presumably they had a reason and the commit message would be where to start looking
[13:58] <hdd1> rbasak: thanks we will keep looking
[13:58] <GridCube> where would one find such commit message?
[13:59] <rbasak> I would start by going to a kernel that had the sysctl and finding the source code responsible. Then go to the current kernel and ask git for changes only to that file
[13:59] <rbasak> jamespage: \o/
[13:59] <rbasak> jamespage: this mongodb build is indeed taking a while. I'll leave it going :)
[13:59]  * rbasak didn't run it inside a screen though :-/
[14:00] <SpamapS> GridCube: why do you want to make tcp sessions run less than 60 seconds?
[14:00] <GridCube> SpamapS, to many conections that saturate the server
[14:03] <SpamapS> hm
[14:03] <hdd1> SpamapS: as GridCube say, there are too many tcp connections and that "overbloat" the server with timewait
[14:04] <hdd1> SpamapS: we try with tcp_recycle and reuse, but got the problem that sometimes when an old conection is reused,this old connection can send again a new package and all got miss up
[14:04] <hdd1> SpamapS: but this can only happend with the WAN connections, so the perfect value that solves all is nolocaltimewait
[14:05] <SpamapS> interesting
[14:07] <nxvl> Daviey: c'mon!
[14:07] <hdd1> SpamapS: yup. We saw some forums where they set the net.inet.tcp.nolocaltimewait=1 , but we try this and we get the unknow key error
[14:07] <nxvl> Daviey: don't make that face again!
[14:08] <hdd1> SpamapS: i know that in freebsd that value stills on (its really usefull in servers) but dont know what append in ubuntu-server
[14:08] <g0tcha> when trying to bring a network interface up using 'ifup eth1' and it says "RTNETLINK answers: File exists" .. does this mean the IP is being used?
[14:31] <eutheria> hello, i was wondering, is there something like landscape for my LAN?
[14:35] <SpamapS> eutheria: yes you can get an on-premise landscape
[14:35] <eutheria> i bet that is expensive
=== wedgwood_away is now known as wedgwood
[15:08] <g0tcha> heya guys, i installed LAMP on ubuntu server and the home dir is /var/www, if my username is 'gotcha', how can i give permission to 'gotcha' to be able to edit /var/www ?
[15:33] <Nafallo> SpamapS: hey. I have a question for you, if you're around :-)
=== fire is now known as nerd
[15:45] <CPrompt^> greetings.  For some reason apache will not run on my ubuntu server unless I am logged in.
[15:45] <CPrompt^> where can i start to look for a cause?
[15:45] <CPrompt^> i have not changed anything except that I pointed the main web directory to a directory in my home folder
[15:45] <CPrompt^> i have done this in the past and it never requried me to log in
[15:46] <Pici> CPrompt^: is your home encrypted?
[15:46] <Nafallo> CPrompt^: ecryptfs?
[15:48] <CPrompt^> ah...that is a good question.  I just installed did a fresh install a few days ago so honestly I can't remember if I told it to encrypt the home directory or not.
[15:50] <CPrompt^> oh man
[15:50] <CPrompt^> yeah.  you are correct
[15:50] <CPrompt^> it is set to encryptfs.  So...since I made a bonehead move, is there a way to allow it run without being logged in?
[15:50] <Nafallo> not without running with ecryptfs.
[15:50] <CPrompt^> bummer
[15:50] <Nafallo> alt. changing directory
[15:51] <CPrompt^> so...if I create a new user will it automatically use ecryptfs?
[15:51] <Nafallo> my tests say no
[15:51] <CPrompt^> yeah that's an easy test.
[15:51] <Nafallo> but then, do you need ecryptfs on a server? :-)
[15:52] <CPrompt^> well...this is just for my personal use so to me? no, I don't
[15:54] <Nafallo> CPrompt^: https://help.ubuntu.com/community/EncryptedPrivateDirectory#How_to_Remove_an_Encrypted_Private_Directory_Setup
[15:56] <CPrompt^> oh nice.  I'll give that a go.  Need to back a backup and try it out
[15:56] <CPrompt^> thanks Nafallo
[15:56] <Nafallo> no worries :-)
[15:56] <Nafallo> glad I could help
[15:56] <thane_> Hi, I'm trying to configure an virtual Ubuntu server as a router. IP forwarding is on, interfaces are running, routing table looks ok, other machines can ping the server, but other machines cannot ping eachother using the server. Any tips on how to troubleshoot this?
[15:56] <CPrompt^> i'm glad you could too :)
[16:09] <thane_> Hi, I'm trying to configure an virtual Ubuntu server as a router. IP forwarding is on, interfaces are running, routing table looks ok, other machines can ping the server, but other machines cannot ping eachother using the server. Any tips on how to troubleshoot this?
[16:12] <Nafallo> !ask | thane_
[16:12] <ubottu> thane_: Please don't ask to ask a question, simply ask the question (all on ONE line and in the channel, so that others can read and follow it easily). If anyone knows the answer they will most likely reply. :-) See also !patience
[16:12] <Nafallo> bah, that might not be the one for repetative questions :-P
[16:14] <GridCube> !patience | should be
[16:14] <ubottu> should be: Don't feel ignored and repeat your question quickly; if nobody knows your answer, nobody will answer you. While you wait, try searching https://help.ubuntu.com or http://ubuntuforums.org or http://askubuntu.com/
[16:26] <jcastro> robbiew: I'll fire up the plenary like 10 minutes early if you want to prepare
[16:27] <robbiew> jcastro: cool, thx
[16:53] <rbasak> smoser: could you please comment on bug 1069570? What's our current answer, and is the patch anywhere?
[16:53] <uvirtbot> Launchpad bug 1069570 in isc-dhcp "1 MAC Address, two IPs - DNS is "out of sync" with DHCP leases databases, I think..." [Undecided,Confirmed] https://launchpad.net/bugs/1069570
[16:54] <hallyn> smoser: fwiw, http://thread.gmane.org/gmane.comp.emulators.qemu/198455/focus=198477    mjt brought up the memory size patch on qemu m-l
[16:54] <hallyn> smoser: i particularly love the second email - basically a previous patch had been nacked bc it didnt' also handle numa (feh)
[16:55] <smoser> rbasak, the patch is in the branch there.
[16:56] <rbasak> smoser: ah yes of course - thanks. I missed that.
[16:56] <smoser> i'll put a comment in though
=== wedgwood is now known as wedgwood_away
=== wedgwood_away is now known as wedgwood
[17:16] <smoser> rbasak, i just re-based raring branch there.
[17:16] <rbasak> smoser: thanks!
[17:16] <smoser> stgraber, you've recently touched isc-dhcp...
[17:16] <smoser> so i'm wondering on your feelings on https://bugs.launchpad.net/ubuntu/+source/maas/+bug/1069570
[17:16] <uvirtbot> Launchpad bug 1069570 in isc-dhcp "1 MAC Address, two IPs - DNS is "out of sync" with DHCP leases databases, I think..." [Undecided,Confirmed]
[17:17] <stgraber> smoser: I've seen a few emails from LP this morning about that bug. I'm currently busy preparing my plenary but I'm scanning through isc-dhcp bugs fairly regularly, so I'll get to it
[17:17] <smoser> read comment 34 and see patch in my branches.
[17:17] <smoser> stgraber, no hurry.
[17:17] <smoser> thanks.
=== wedgwood is now known as wedgwood_away
[17:35] <rbasak> smoser: thanks for the update on that bug!
[17:39] <adam_g> jamespage, is this fix still appropriate for the charms? it was proposed to into nova-compute but i'd like commit the same to the other relevant places.  https://code.launchpad.net/~gandelman-a/charms/precise/cinder/lp1099900/+merge/151812
=== wedgwood_away is now known as wedgwood
[17:39] <jamespage> adam_g, yes it is
[17:40] <adam_g> jamespage, thanks
[17:41] <jamespage> adam_g, man do we need to migrate to python :-(
[17:41] <jamespage> maybe raise that at the charms session for post HA work
[17:41] <adam_g> jamespage, yeah...
=== med__ is now known as medberry
[17:58] <SpamapS> Nafallo: I'm around now, wassup?
[18:00] <jamespage> adam_g, I added a comment to https://code.launchpad.net/~gandelman-a/charms/precise/glance/lp1099900/+merge/151813
[18:00] <jamespage> re the get_ip function
[18:00] <jamespage> and some feedback jimbaker gave me on the ceph charms which made sense to me
[18:01] <adam_g> jamespage, ah,thanks
[18:01] <jamespage> adam_g, with regards you comment on https://code.launchpad.net/~gandelman-a/charms/precise/hacluster/ocf/+merge/151345
[18:02] <jamespage> would the packages ocf's not overwrite the charm ones?
[18:03] <adam_g> jamespage, yeah, thats what i was getting at (sorry, it was 6AM:) if the ceph-resource-agents package does not provide the rbd OCF, it will fallback to what is installed by the charm
[18:05] <jamespage> adam_g, ah - right I think I see now
=== blitzkrieg3 is now known as jmleddy
=== medberry is now known as med_
=== NomadJim_ is now known as NomadJim
[19:46] <paco1> hello masters!
[19:49] <paco1> i have a little with my bind servers. i have 2 servers, master and slave with TSIG and it works fine. if i incorpore a second slave bind server with the same TSIG key, i have that error: zone 172.in-addr.arpa/IN: refresh: failure trying master 172.31.250.1#53 (source 0.0.0.0#0): tsig indicates error
[19:50] <paco1> i generated the TSIG key like that: dnssec−keygen −r /dev/random −a HMAC−MD5 −b 512 −n HOST name
[19:57] <paco1> -n HOST: is it just for one slave server?
[19:59] <rbasak> jamespage: I couldn't reproduce the ftbfs in mongodb. My build succeeded using sbuild. There is only one build dependency selection difference, which is a missing libsqlite3-0 on the buildd, but I don't see how it could be related. I'll dig more tomorrow.
[20:11] <smoser> hallyn, or anyone...
[20:11] <smoser> how can i know how many blocks are on a block device
[20:11] <smoser> hoping not to parse /proc/partitions
[20:12] <RoyK> smoser: /proc/partitions is pretty easy
[20:13] <RoyK> it's just 1kB blocks
[20:13] <smoser> RoyK, if you happen to know the kernel name
[20:14] <smoser> which, yes, is fairly easy to find out given udev's nicely created /dev/block/MAJOR:MINOR
[20:14] <smoser> but i'm just hoping to depend on just about as little as possible.
[20:16] <smoser> RoyK, (sorry if i sound ungrateful, thanks for your input)
[20:17] <RoyK> smoser: np
[20:17] <smoser> i think i might have to end up parsing /proc/partitions
[20:17] <smoser> but its a PITA for testing
[20:18] <smoser> (since not only am i needing to read that, i'm wanting to invoke 'update' and see it change)
[20:19] <astor> what is *roughly* the minimum image/install size for a server that is set up to be part of a juju-cluster? before any services are installed?
[20:20] <sarnold> astor: I think "hundreds of megabytes" isn't too far wrong
[20:21] <astor> sarnold: ok, thanks.
[20:23] <astor> second question: The faq states: "Currently each service unit is deployed to a separate machine that...".  I assume that jojo can create a separate VM on a single machine for each service? True or not?
[20:23] <sarnold> astor: it's complicated...
[20:24] <sarnold> astor: in 'local' mode, juju can deploy services in LXC containers -- but that won't ever leave the one machine
[20:25] <sarnold> astor: you can also start a pile of VMs yourself and have them run openstack or MAAS and deploy to those VMs, but juju doesn't provide much help in that case, as I understand it
[20:25] <hallyn> smoser: uh, sudo fdisk -l?
[20:25] <hallyn> you probably mean from c code...
[20:26] <sarnold> hallyn: I thought about suggesting that but figured (a) it only reports partitions tables, not actual device data (b) it doesn't grok gpt partition tables either :(  ...
[20:26] <smoser> well, no. i meant from python.
[20:26] <smoser> i was just wondering if hter ewas something easy.
[20:26] <smoser> ie, an explicit size attribute back from stat or something.
[20:26] <soren> smoser: blockdev --getsz?
[20:27] <astor> sarnold: if I spin up a dozen VMs that have the "hundreds of megabytes" image that is juju-ready, then I can somehow register these with juju, and juju can deploy to them, but not create new ones (or make changes to them).  Is that close to the truth?
[20:27] <smoser> that would seem to be it, soren.
[20:28] <hallyn> cool
[20:28] <soren> Yeah. That's why I suggested it. :)
[20:28] <soren> Happy UDS, by the way.
[20:28] <soren> Or is it merry UDS? I always forget.
[20:29] <astor> another q: does MAAS only cover the concept of Linux+LXC, and not Xen+Linux?
[20:29] <sarnold> astor: it ought to be able to make changes as needed; just not create new ones itself
[20:30] <astor> sarnold: ok, thanks. good to know.
[20:30] <sarnold> astor: well, my suggesting MAAS on top of kvm or xen is an abuse of maas -- which is intended for running on bare metal
[20:31] <smoser> http://unix.stackexchange.com/questions/52215/determine-the-size-of-a-block-device gives lots of answers
[20:31] <astor> sarnold: I understand. I'm looking for a small TCB.  OpenStack is Xen + lots of services, while MAAS is without Xen, so there's something missing in between.
[20:32] <soren> smoser: If you want to avoid shelling out (even though that's a perfectly reasonable thing to do, but I know many frown upon it), the ioctl you want is BLKGETSIZE64.
[20:32] <smoser> soren, right.
[20:32] <smoser> but i think i might just read /sys/class/block/<devname>/size
[20:32] <smoser> although i'm not certain that is in 512 byte blocks (always).
[20:33] <fluvvell> Does anyone have a happy tale to tell about grub installing on a new GPT partitioned drive? I've got one that wont boot without some help
[20:33] <sarnold> astor: you may wish to ask in #juju if there's a better provider for your goals :)
[20:33] <SpamapS> smoser: ioctl(3, BLKGETSIZE, 0x7fff215a5f60)    = 0
[20:33] <sarnold> astor: (everything I've got is second-hand at best :)
[20:33] <astor> sarnold: ok, thanks.
[20:34] <SpamapS> smoser: unfortunately, its not clear whether python has a way to know what BLKGETSIZE is
[20:34] <soren> SpamapS: ..and hope it fits in an int.
[20:36] <SpamapS> yeah, fcntl.ioctl looks really low level and not very python-programmer friendly
[20:36] <SpamapS> http://stackoverflow.com/questions/2773604/query-size-of-block-device-file-in-python
[20:36] <SpamapS> smoser: ^^ might work
[20:38] <smoser> hm... thats interesting.
[20:40] <smoser> SpamapS, that seems pretty nice.
[20:42] <soren> smoser: You can rely on /sys/block/blah/size to results consistent with --getsz (i.e. both in sector counts).
[20:42] <soren> smoser: fwiw
[20:42] <hallyn> heh, that (seek to end) is what i like to do for files, but some filesystems don't support it
[20:42] <hallyn> (esp some proc files)
[20:42] <soren> smoser: They are implemented pretty much the same in the kernel.
[20:43] <smoser> hallyn, well, here i know the file is a block device
[20:45] <hallyn> smoser: I know, i chuckled bc i didn't think to do it for a blockdev, whereas it would be myd efault for a file
[20:45] <hallyn> i was bein gsilly
[20:45] <hallyn> ttyl :)
=== JonnyNomad_ is now known as JonnyNomad
[21:16] <hallyn> stgraber: see my reply to Christian about merge requests.   we should perhaps formalize some workflow.  You and I work differently so I'm not sure what would be preferred in the end...
[21:17] <hallyn> (several ppl submitting patches have github trees, so we can at any rate make this easier than hand-applying patches from email)
[21:21] <stgraber> hallyn: I tend to prefer processing the patches with git am. I also tend to review the ones we get through github, but it's usually more painful as I need to click through the whole site to get meaningful diffs, then cherry-pick and edit the commits to ack them
[21:22] <hallyn> stgraber: agreed, i don't want pull requests
[21:22] <hallyn> git fetch + get cp or get rebase is nice
[21:22] <stgraber> hallyn: so it's nice to have a github branch to be able to easily see the end result and try some test builds, but for the actual review, I prefer individual patches to the ML
[21:22] <hallyn> (especially since email patches tend to be line-wrapped)
[21:22] <hallyn> stgraber: so it sounds like we're agreed,
[21:22] <hallyn> to my surprise :)
[21:23] <hallyn> see the process i recommended in email and lemme knwo if you'd change anything
[21:23] <stgraber> the only pull requests I deal with is the ones I send to Daniel for releases because sending a few hundred e-mails to him seems rather pointless when he realistically won't look at them all anyway
[21:24] <hallyn> yup
[21:26] <hallyn> hm, i was hoping that with vUDS going on, ppa build farm would be under-utilized, and builds woudl go through quicker :)
[21:26] <hallyn> alas, no
[21:27] <stgraber> hallyn: if you give me a URL I can bump the score for you
[21:28] <hallyn> stgraber: qemu build in https://launchpad.net/~serge-hallyn/+archive/virt
[21:29] <hallyn> i'm hoping it'll give a more stable qemu-arm-static for builders
[21:30] <stgraber> hallyn: bumped
[21:31] <hallyn> thanks!
[21:45] <hallyn> well that was underwhelming
[23:53] <Dataoriginal> hello
=== wedgwood is now known as wedgwood_away