[00:01] <jkyle> what's your error?
[00:01] <qhartman> that's an excellent question, this erlang stuff is pretty inscrutable
[00:01] <qhartman> <- digging through logs now
[00:03] <qhartman> I thought at first that it miht be a hostname thing like https://bugs.launchpad.net/ubuntu/+source/rabbitmq-server/+bug/653405
[00:03] <qhartman> the symptoms are all the same
[00:04] <qhartman> but I haven't monkied with the hostname
[00:04] <qhartman> and resetting it to the apparently required "localhost" doesn't seem to help
[00:06] <qhartman> I guess it could still be the problem, EC2 name resolution is notoriously weird
[00:08] <qhartman> I'm sure erlang is neat and all, but having to learn yet another language just to figure out how to troubleshoot my chef-server installation is a bit annoying...
[00:12] <qhartman> interesting. If I nuke the existing (suppoedly empty) datadir it starts right up
[00:21] <qhartman> alright, well that seems to have gotten around the problem
[00:33] <jkyle> so, I keep getting this warning on preseed installs that say I must create a "Reserved BIOS boot area".
[00:33] <jkyle> but I cannot find a single reference on how to create this in a preseed file. it's only needed for gpt parts and the preseed also ignores specifying other partition lables, such as msdos
[00:35] <qhartman> is this useful: http://jasonschaefer.com/archives/190
[00:35] <qhartman> I haven't used preseeds in a long time, but if you could setup the partitioning as described there, it seems like you would be good to go
[00:37] <qhartman> I wouldn't be surprised though if the preseed stuff hasn't had the bits needed to fully support gpt put in yet
[00:48] <jkyle> yeah, that's manual though
[00:48] <jkyle> what I need to figure out is how to create a partition like that in partman
[00:48] <jkyle> or how to get partman to use a msdos label
[00:49] <delinquentme> if [ `python2.7 -V` == "python 2.7" ] ;then  echo "true"; fi
[00:49] <delinquentme> I want to test the command "python2.7 -V" to see if it returns "python 2.7"
[00:49] <delinquentme> how can I do this
[00:50] <jkyle> that looks fairly close
[00:59] <delinquentme> simpler; $V=`python2.7 -V`
[00:59] <delinquentme> How do I store the string which is returned on running that
[00:59] <delinquentme> in $V
[00:59] <delinquentme> right now it says = is not found
[01:00] <jkyle> v=`foo`, you ought to check out a beginners shell scripting guide
[01:03] <delinquentme> yeah but then echo $v
[01:04] <delinquentme> prints out nothing
[01:04] <patdk-lap> heh
[01:07] <delinquentme> sorry
[01:07] <delinquentme> its just if I read the manual for every programming language I ever wanted to use
[01:07] <delinquentme> well
[01:07] <delinquentme> yeah i think you know how that would end
[01:08] <patdk-lap> you would be me?
[01:08] <qhartman> try quoting the backticks
[01:09] <qhartman> if [ "`python2.7 -V`" == "python 2.7" ] ;then  echo "true"; fi
[01:09] <patdk-lap> c, c++, c#, several asm, fortran, forth, perl, php, lisp, pascal
[01:09] <patdk-lap> I think that list is getting shorter as I get older, memory loss :(
[01:09] <qhartman> though that will theoretically always be a 2.7.x binary though
[01:10] <delinquentme> smarter way to check if python 2.7 is installed?
[01:10] <patdk-lap> use whereis
[01:14] <delinquentme> no good as it gives output for both python2.6 and python2.7
[01:40] <Laice> anyone any good with ISPConfig?
[02:15] <delinquentme> if I'd like a variable which returns the particular OS name for any POSIX system...
[02:15] <delinquentme> what would that var be?
[02:18] <delinquentme> or what would i used to return "ubuntu 10.04"
[02:33] <ihashacks> well, in LSB Linux distributions you can use "lsb_release -d"
[02:35] <linocisco> how to install vbox guestaddtion on ubuntu ?
[02:43] <brainysmurf> Is it possible to have a folder in /var/www/ that symlinks to somewhere else and still be accessible to http?
[03:35] <brainysmurf> does netstat show you outgoing and incoming ports? or am i confused?
[03:40] <RamJett> brainysmurf: not sure if you are confused, but yeah netstat -a even better
[03:40] <brainysmurf> RamJett even better as in more info than just incoming and outgoing?
[03:41] <brainysmurf> I'm trying to sort out if my connections are being refused due to something on my server or the firewall
[03:43] <RamJett> lots of LISTEN too .. -a is suppose to be (all) is a alias for --all
[03:48] <brainysmurf> i'm an idiot, what does listen mean :)
[03:48] <RamJett> netstat won't show any firewall rules. Do something like "iptables -vL" or "iptables -vvL" for even more info
[03:53] <linocisco> how to install vbox guestaddtion on ubuntu ?
[03:53] <linocisco> how to install vbox guestaddtion on ubuntu ?
[03:56] <JoeCoder> in ubuntu 10.04, I installed postfix-tls, but this package is missing in 12.04
[03:56] <JoeCoder> how do I figure out if packagse have been renamed?
[03:58] <RamJett> brainysmurf: LISTEN show ports and IP listening. Like if you have a local imap server running you would see something like this " tcp    ip.xx.xx.xx.xx:port       0.0.0.0:*   LISTEN .. but if iptables of firewall type rules are blocking it. It will still be LISTEN .. the rule would just block it on the INPUT
[03:59] <brainysmurf> RamJett thanks for that i'm currently trying to figure out what my firewall rule is just for smtp
[04:02] <RamJett> Who's router? Most many ISPs block 25 to end users. They want you to use their smtp
[04:03] <brainysmurf> RamJett I know but my sysadmin insists that isp nor school firewall isn't doing that; I need to prove it
[04:04] <JonEdney> Hey, anyone able to troubleshoot a mailserver issue?
[04:06] <RamJett> brainysmurf: So are you trying to have box run a smtp engine .. You should see "tcp        0      0 0.0.0.0:25              0.0.0.0:*               LISTEN" on the box
[04:08] <brainysmurf> RamJett: Yes, I am trying to run smtp engine and have that line on the box
[04:08] <brainysmurf> Mail comes in fine but mail does not go out
[04:08] <brainysmurf> postfix reports connection refused and telnet does the same
[04:10] <brainysmurf> Mail does not go out to machines outside of school, but does go out to machines inside of school
[04:13] <RamJett> brainysmurf: What box is refusing outside? It might be refusing for other reasons. Like IP reverse lookup. Not sure what your setup is.
[04:14] <brainysmurf> IP reverse lookup is setup: student.ssis-suzhou.net
[04:16] <brainysmurf> So far I havne't found a box that doesn't reject :)
[04:16] <RamJett> So YOUR postfix is reporting "connection refused" ?
[04:17] <brainysmurf> yes, mine
[04:17] <brainysmurf> as well as when I telnet
[04:17] <brainysmurf> "telnet: Unable to connect to remote host: Connection refused"
[04:18] <RamJett> telnet FROM outside to your postfix to outside smtp ?
[04:19] <RamJett> that is "or outside"
[04:19] <brainysmurf> that error happens from the server trying to connect to outside smtp
[04:21] <RamJett> Like from your smtp server box. If you do "telnet mail.ramnic.com 25" .. what do you get?
[04:21] <brainysmurf> "Trying 69.178.157.231...
[04:21] <brainysmurf> telnet: Unable to connect to remote host: Connection refused
[04:22] <RamJett> try one more time. I'll let you know if it hits
[04:22] <brainysmurf> okay i just tried
[04:22] <brainysmurf> and again
[04:23] <RamJett> It never gets to ramnic.com port 25. I would see it
[04:23] <brainysmurf> So it's a firewall issue
[04:23] <brainysmurf> ?
[04:24] <RamJett> so it is on the router outbound port 25 from your host. You know your host isn't blocking if it sends to local boxes
[04:25] <brainysmurf> Ah, that's a good point
[04:25] <brainysmurf> You're a king
[04:30] <RamJett> For sure not. I'm still hanging out here hoping some picks up on my question. Been waiting all day. I was told to be patient. So I'm trying.
[04:55] <ChmEarl> hurray - no more devel version: DISTRIB_DESCRIPTION="Ubuntu 12.04 LTS"
[04:58] <RoyK> erm - I thought release was tomorrow...
[04:59] <RoyK> ChmEarl: file a bug ;)
[04:59] <ChmEarl> LOL - I updated from mirrors.kernel.org too
[05:02] <___MAX> hi , i am using sakis3g for internet connection but after using apt-get auto remove it ask for chat dependency , so how can i get this dependency or package
[05:04] <___MAX> u suckkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
[05:22] <koolhead17> morning all
[07:29] <railsraider> anyone knows how to properlyy configure the server for redis so redis won't crash or get killed by oomkiller?
[07:38] <koolhead17> railsraider, trying will  #redis be helpful as well
[07:39] <railsraider> i tried no one answered in half an hour
[07:40] <rbasak> railsraider: I know nothing about redis. But I do know that you can mess at the kernel end with /proc/.../oom*. Maybe googling for these and redis would help? Although ideally redis would be configured to not need this tuning, and I don't know how to do that or if it's possible.
[07:41] <railsraider> thanks rbasak: my swap is disabled this might be the cause for these issues
[07:41] <railsraider> is there a way to tell redis not to use the swap unless out of RAM?
[08:28] <Gallomimia> hi i'm trying to install a downloaded .deb but it is telling me a great many dependancies exist. it won't install these dependancies for me however. how do i make it do that? i asked in not-server channel and they said use software center. servers don't have that :)
[08:33] <Jeeves_> Gallomimia: Is it saying 'but is not going to be installed'
[08:33] <Jeeves_> Or is it saying 'cannot be installed'
[08:34] <greppy> Gallomimia: sudo apt-get -f install
[08:34] <greppy> Gallomimia: see if that grabs what is needed.
[08:34] <Jeeves_> greppy: That usually removes the package that you want to install
[08:35] <rbasak> Gallomimia: I would avoid what you're doing if at all possible. Not using a repository is a sure route to a broken system.
[08:35] <Gallomimia> Jeeves_: it's saying not
[08:35] <greppy> Jeeves_: when dealing with dependancies, I usually see that install all the dependcies, after a failed dpkg -i $foo.deb
[08:35] <Gallomimia> not going to be installed
[08:35] <Gallomimia> and not "can't
[08:35] <rbasak> (as well as a security risk)
[08:36] <Jeeves_> rbasak: That's BS :)
[08:36] <greppy> Gallomimia: pastebin the output?
[08:36] <Jeeves_> Gallomimia: ok, you can try -f install.
[08:36] <Gallomimia> i downloaded it from a rather trustworthy site with an SSL cert
[08:36] <Gallomimia> the package might even be avail on sources
[08:36] <Jeeves_> But also, installing the packages that it needs manually is fine too
[08:36] <Gallomimia> i dont know. it's folding at home fyi
[08:36] <rbasak> Does the trustworthy site not provide an apt repository?
[08:36] <Gallomimia> i don't even know
[08:36] <rbasak> Jeeves_: and what about security updates?
[08:37] <Gallomimia> security updates should be downloaded from anonymous
[08:37] <ogra_> rbasak, who is doing the release image testing for armadaxp ? there are no test results for the images on the isotracker at all
[08:37] <rbasak> ogra_: nfi. I didn't even know that the iso tracker covered it.
[08:37] <ogra_> (we usually dont release untested images, would be good if someone could test them ;) )
[08:38] <Gallomimia> http://pastebin.com/4kQZzsQv
[08:38] <Jeeves_> rbasak: You should monitor those, if you install stuff manually
[08:39] <Gallomimia> so i asked aptitude to install one of these things and it offered to remove 89 other packages
[08:40] <Gallomimia> great. i guess i just don't know jack about the finer points of apt
[08:41] <Jeeves_> Gallomimia: Probably, one of the packages creates that issue
[08:42] <rbasak> Gallomimia: sorry you're having trouble. It's not really expected for users to mess with .debs manually. That's what apt is for - it's supposed to be pointed at a repository and work everything out. If your third party doesn't provide one, you need to know about packaging internals (ie. dpkg).
[08:42] <Gallomimia> an issue that needs to be resolved by deleting 89 software packages? sure they're not needed for some other things we installed 2 weeks ago
[08:43] <Gallomimia> rbasak: i'm not too worried about the deeper packaging internals. it tells me the two packages in my pastebin are dependant, which are in turn dependant on a lot of other stuff. install those two and make the shell stop whining at me please.
[08:44] <rbasak> Gallomimia: try purging your third party package, which may make your package state consistent again. Then install the dependencies you know about using apt. Then dpkg -i the third party package again. The problem is that by using dpkg -i apt is missing the full story about what you want.
[08:45] <rbasak> ogra_: I can't find armadaxp in the iso tracker. Do you have a link?
[08:45] <Gallomimia> rbasak: i think all of this can be avoided. i think this package is in standard repositories
[08:45] <ogra_> rbasak, http://iso.qa.ubuntu.com/qatracker/milestones/214/builds in the netboot area
[08:48] <rbasak> ogra_: I can grab a board and run that test, but the step-by-step instructions are entirely wrong (panda-specific) so I'll have to be inventive. Which isn't a problem for me, but might be for the validity of any test.
[08:49] <ogra_> rbasak, well, someone of the arm server QA team should have added some proper tests (massimo) for that specific board, seems that didnt happen
[08:49] <Gallomimia> i.... think it's installed. but it appears i'm still sofa king lost.
[08:50]  * ogra_ proposes the arm server QA team should have a beer with the release manager in SF :) 
[08:50] <ogra_> to plan for Q
[08:57] <lynxman-> morning o/
[09:05] <railsraider> i need to enable swap again on a machine that it was disabled how do i do that on 10.04 LTS?
[09:07] <ogra_> make sure you have a line for it in fstab ...
[09:07] <ogra_> then: sudo swapon -a
[09:07] <lynxman-> jamespage: ping, about bug #986973 after talking with upstream they reckon it's not worth fixing as of yet and recommend to wait to facter 2.0.0, but can't change the status to "won't fix" with my permissions...
[09:11] <railsraider> i have this line ogra_ : /dev/sda2       none        swap    sw          0 0
[09:11]  * koolhead17 pokes lynxman- :)
[09:11] <railsraider> when i try to swap on it fails:
[09:11] <railsraider> swapon: /dev/sda2: stat failed: No such file or directory
[09:11] <ogra_> you should instead use the UUID there
[09:11] <railsraider> how do i find the uuid?
[09:12] <ogra_> use: sudo blkid /path/to/device
[09:12] <ogra_> ogra@horus:~$ grep swap /etc/fstab
[09:12] <ogra_> UUID=2634be76-0cc2-443f-a629-4604a5d09b6d	none	swap	defaults	0	0
[09:12] <ogra_> thats how such an entry should look like
[09:12] <ogra_> (with the right UUID indeed)
[09:13] <railsraider> i don't have that line
[09:13] <railsraider> i miss the uuid
[09:13] <railsraider> how do i create that?
[09:13] <greppy> railsraider: blkid
[09:13] <railsraider> ok
[09:15] <railsraider> i did sudo blkid /dev/sda2
[09:15] <railsraider> there's no uuid in /etc/fstab
[09:15] <railsraider> what am in doing wriong
[09:16] <greppy> blkid /dev/sda2 would just give you the UUID for that device.
[09:16] <greppy> you then need to edit /etc/fstab
[09:16] <jamespage> lynxman-, so its not really that much of an issue?
[09:16] <lynxman-> jamespage: it really looks like, in my opinion not worth going through a SRU
[09:17]  * lynxman- waves koolhead17 
[09:17] <jamespage> lynxman-, OK - I'll do the bits then...
[09:17] <lynxman-> jamespage: thanks
[09:17] <koolhead17> lynxman-, 1 day to go. :D
[09:19] <jamespage> lynxman-, guess thats probably been in the codebase for a while :-)
[09:19] <railsraider> greppy: that command didn't output anything
[09:20] <greppy> railsraider: what does 'ls -l /dev/disk/by-uuid/' give you?  pastebin it?
[09:20] <lynxman-> jamespage: it's a marginal problem as well, which can carry some regression issues, that's why it has been targeted for 2.0.0
[09:21] <jamespage> lynxman-, makes sense
[09:22] <railsraider> greppy: http://pastebin.com/haqU0Kjb
[09:23] <greppy> railsraider: you don't have a /dev/sda2
[09:29] <Daviey> ivoks: around?
[09:30] <railsraider> greppy: i disabled swap on that machine long time ago and now i need it back
[09:31] <railsraider> maybe its in /proc?
[09:31] <greppy> railsraider: check your disk, make sure that it has a swap partition.
[09:37] <railsraider> greppy:  how do i add the swap manually ?
[09:37] <railsraider> i have sdc1 which is 16GB
[09:37] <railsraider> i have the  uuid from it
[09:37] <greppy> railsraider: you want to use sdc1 as your swap?
[09:37] <railsraider> yes
[09:38] <railsraider> i think that used to be the old swap
[09:38] <greppy> railsraider: mine looks like this
[09:38] <greppy> UUID=0ae26ab8-0a8c-47c5-aa8a-6582eec84a1d none            swap    sw              0       0
[09:38] <greppy> change the UUID=... to match your UUID for sdc1
[09:38] <railsraider> is this in /etc/fstab?
[09:39] <greppy> yes
[09:40] <ivoks> Daviey: yes
[09:41] <Daviey> ivoks: is bug 927594 still accurate ?
[09:41] <ivoks> Daviey: i don't think so...
[09:42] <Daviey> ivoks:  it looks fixed, but wanted to check
[09:42] <ivoks> Daviey: i need to check it, give a minute
[09:42] <Daviey> $ ls -alh /etc/openstack-dashboard/local_settings.py
[09:42] <Daviey> -rw-r--r-- 1 root root 4.2K Apr 23 09:28 /etc/openstack-dashboard/local_settings.py
[09:42] <ivoks> how about /usr/share/openstack-dashboard/local/local_settings.py
[09:43] <Daviey> dammit, no
[09:43] <ivoks> but that's not a problem anymore, i believe
[09:43] <ivoks> cloud-live works without that link :)
[09:44] <ivoks> let's install it :)
[09:44] <Daviey> local_settings.py -> /etc/openstack-dashboard/local_settings.py .. aha!
[09:45] <Daviey> /usr/share/openstack-dashboard/openstack_dashboard/local/local_settings.py .. it is fixed!
[09:45] <Daviey> thanks
[09:45] <ivoks> ok then
[09:46] <Daviey> ivoks: are you doing the memcache fix in cloud-live
[09:46] <Daviey> ?
[09:46] <ivoks> sed -i "s|CACHE_BACKEND.*|CACHE_BACKEND = 'memcached://127.0.0.1:11211/'|g" /etc/openstack-dashboard/local_settings.py
[09:46] <ivoks> yes
[09:47] <railsraider> thanks greppy !
[09:50] <greppy> railsraider: so that worked for you?
[09:51] <railsraider> i think so
[09:52] <greppy> cool
[10:05] <Daviey> ivoks: looking to put http://pb.daviey.com/qKej/ in precise.. happy with that?
[10:16] <ivoks> Daviey: well, if you must... i'll update cloud-live then
[10:21] <samuel_> hello all, whoever helped me yesterday - thanks!
[10:21] <samuel_> i have a query regarding the organisation of users and groups however
[10:21] <samuel_> i have inherited a server (ie not setup by myself) and it appears the group permissions are a bit of a mess
[10:22] <samuel_> it seems that when users create files or folders in the shared area's on the server, other users cannot read or edit them, as each users has his/her own group which is the same as the username
[10:23] <samuel_> is this right?? i have to keep changing the ownership to something more neutral so that filesharing can resume, its very annoying! i have read through the documentation but i'm not 100% sure on the best way to resolve this, can anyone offer advice? thanks in advance!
[10:26] <ivoks> samuel_: if this is samba, look at force user and force group
[10:27] <ivoks> samuel_: if it's NFS or some other native linux protocol, look at acl
[10:27] <ivoks> acl and umask
[10:27] <Daviey> ivoks: it won't break cloud-live, as it'll override existing values :)
[10:28] <rbasak> samuel_: I think you want the sgid bit. Take a look at http://en.wikipedia.org/wiki/Setuid#setuid_and_setgid_on_directories
[10:28] <Daviey> ivoks: really.. carrying what should be patches in cloud-live isn't ideal, right? :)
[10:28] <samuel_> hi, sorry i did not give specifics, i'm using ubuntu server 10.04, sharing files with windows and linux users. it appears that each user belongs to a group that is the same as the username, and when they create a file, it is owned by them, and other users are unable to view them
[10:29] <ivoks> Daviey: i don't consider it a patch... it's more usecase scenario
[10:29] <samuel_> while each user is a member of a common group, the common group is not used for ownership when files are created
[10:30] <samuel_> i will check out the link posted and read up on acl, thanks for throwing a bone! :D
[10:30] <rbasak> samuel_: yeah the sgid bit will do what you need
[10:30] <ivoks> samuel_: acl (getfacl/setfacl) and force user and force group in samba are answers to your questions
[10:30] <ivoks> i can just assume you are talking about samba
[10:30] <rbasak> ivoks: that would work, but sgid is simpler and the more universal to this problem.
[10:30] <rbasak> *more universal answer
[10:31] <ivoks> ah, right...
[10:31] <Daviey> ivoks: well the current horizon seems to log people out very regular, which is sucky.. we need to use memcache.
[10:31] <samuel_> well the files are being added and edited using either nautilus or windows explorer (in the case of the windows users)
[10:31] <ivoks> samuel_: that's irrelevant... what's relevant is what service is sharing those files
[10:34] <samuel_> ok, thanks for the advice
[10:34] <samuel_> its samba fyi
[10:35] <ivoks> then you can use sgid, as rbasak told you, or force user and force group in smb.conf; it really depends on what you want
[10:45] <samuel_> ok i will look into both and decide which is best, thanks!
[11:12] <linocisco> alamar, hi
[11:20] <robin> Hello I need to bridge eth0 and eth1 to allow internet on server2 through server 1, I know how to do that but I have to assign 4 ips to server 1 and 6 to server 2,  Do I alias the bridges and repeat?
[11:26] <rbasak> robin: just add multiple addresses to the bridge interface. That should work fine.
[11:26] <robin> rbasak how do I do that ?
[11:27] <robin> rbasak as you can only add one address per iface... so do I alias and make like 4 bridges ? that doesnt seem right
[11:28] <rbasak> robin: just add "up ip addr add 1.2.3.4 dev br0" to your br0 line for each extra IP.
[11:28] <rbasak> robin: that's what I do anyway. I don't know if there's a better way.
[11:29] <rbasak> robin: no need for extra bridges
[11:29] <rbasak> *to your br0 stanza
[11:37] <robin> rbasak thanks for helping what is the dev tag for ?  pastie of my interfaces http://pastebin.com/cMXkJvpT
[11:39] <novato> hola
[12:00] <sanderj> Hi. Do anyone know about if it's possible to use ssd disks as cache for harddrives?
[12:12] <rbasak> sanderj: you might be interested in http://bcache.evilpiepirate.org/ although it looks like it's not production-ready - you'd need to be a developer to try it.
[12:15] <sanderj> rbasak, thanks :-)
[12:33] <zul> good morning
[12:34] <lynxman-> zul: morning! :)
[13:14] <samuel_> thanks to all that helped me earlier! i used the setgid just for the shared area and its working fine now
[13:17] <rbasak> samuel_: \o/
[13:19] <samuel_> altering the samba.conf is not something i wanted to do on a server that is in constant use, but i will remember that for future
[14:02] <mattt> on a stock install (say 11.04), is there anything outside of openssh-server's post install that creates ssh host keys?
[14:58] <halvors> I have some network issues here.
[14:58] <halvors> How do i release an ipv6 address?
[15:11] <pmatulis> j launchpad-ops
[15:28] <claude2> does anyone here understand why nagios-nrpe-server wants to install mysql?
[15:33] <Pici> claude2: it shouldn't... let me take a look here
[15:34] <claude2> mysql-common comes up in the additional packages
[15:44] <SpamapS> claude2: mysql-common is needed for libmysqlclient
[15:44] <SpamapS> claude2: it needs /etc/mysql/my.cnf
[15:44] <SpamapS> Pici: ^^
[15:45] <claude2> ohh to check the status of a mysql server?
[15:45] <SpamapS> precisely
[15:45] <Pici> claude2: Aye, I was just going to say that nagios-plugins-standard recommends libmysqlclient16
[15:45] <claude2> ohh ok. well that's not the end of the world i guess
[15:46] <claude2> thanks guys!
[15:46] <Pici> claude2: You could choose not to install recommends if you prefer.
[15:46] <SpamapS> claude2: consider not installing with recommends on servers.. sometimes it can be a bit overkill :)
[15:46] <claude2> ohh interesting. there is a lot in there, and we are trying to keep these machines minimal
[15:46] <claude2> ill give that a try
[15:48]  * Pici hits apt-repends with a stick a few times
[16:00] <three18ti> anyone got MAAS working?  I keep checking the wiki but it just says "when we have a working avahi live cd".
[16:21] <rockets> Somebody please kill me: https://gist.github.com/2491036
[16:36] <Pici> bugs :(
[16:42] <jamespage> odd - somethings catching up
[16:50] <kirkland> SpamapS: ping
[16:50] <kirkland> SpamapS: when developing an upstart job / init script, and wanting to try and maintain compatibility of the package with both Debian and Ubuntu
[16:51] <kirkland> SpamapS: are there any best practices?
[16:51] <kirkland> SpamapS: ie, do I just need to write both, an upstart job and a sysvinit script?
[16:51] <SpamapS> kirkland: unfortunately, yes. :-P
[16:52] <SpamapS> kirkland: even as Debian grows upstart support, you'll still have to support the LCD which is sysvinit
[16:52] <kirkland> SpamapS: okay, and what about the debian/rules, to decide whether to install the upstart or sysvinit job?
[16:53] <SpamapS> kirkland: we were goign to try and fix dh_installinit in wheezy to support having both in a package dir, and just installing them happily to coexist... but that didn't happen...
[16:53] <SpamapS> kirkland: so for now, you just have to do conditional "is this ubuntu, yes, cp debian/upstart.in debian/upstart" logic
[16:53] <kirkland> SpamapS: yuck
[16:53] <SpamapS> I know
[16:54] <kirkland> SpamapS: okay, is there a package example you can suggest?
[16:54] <SpamapS> I think we could probably build that into dh_installinit .. actually thats not a bad idea.
[16:54] <kirkland> SpamapS: yeah, ideally, at dpkg --install time
[16:54] <SpamapS> kirkland: hrm.. none off the top of my head.
[16:54] <SpamapS> zookeeper maybe
[16:54] <zul> nova
[16:54] <SpamapS> nova is in sync with Debian?!
[16:55] <kirkland> SpamapS: it would figure out if upstart is on the system
[16:55] <kirkland> SpamapS: and if so, do upstart install
[16:55] <kirkland> SpamapS: if not, do legacy install
[16:55] <SpamapS> kirkland: you'd have to use triggers for that so it fixes things when you install upstart after the jobs, but yeah, that would work.
[16:55] <kirkland> SpamapS: right
[16:55] <zul> SpamapS: most of it
[16:56] <zul> SpamapS: but we use the same upstart/init script logic you were talking about
[16:56] <kirkland> zul: cool, thanks
[16:56] <kirkland> zul: I'll have a look
[16:59] <SpamapS> zookeeper also does it very nicely
[16:59] <SpamapS> and is in sync w/ Debian
[16:59]  * SpamapS ^5's jamespage
[17:00] <SpamapS>     if dpkg-vendor --derives-from ubuntu ; then \
[17:00] <SpamapS>         cp debian/zookeeperd.upstart.in debian/zookeeperd.upstart ; \
[17:00] <SpamapS> pretty simple
[17:02]  * koolhead17 wants to jump from his chair out of frustration :(
[17:09] <kirkland> SpamapS: that's nice, cool, thanks
[17:10] <kirkland> SpamapS: one more packaging question for you, this time python
[17:10] <kirkland> SpamapS: I have a package that works well with python2.6 and above, but it's not very happy at all when python2.5 is installed on the system
[17:10] <kirkland> SpamapS: any other packages having this issue?  what are packagers doing about it?
[17:11] <kirkland> SpamapS: just using a conflicts << 2.6?
[17:13] <SpamapS> kirkland: errr.. why would it matter what was on the system, what matters is what is set to default, no?
[17:14] <kirkland> SpamapS: well that's what I thought, but in practice, I'm seeing some strange behavior
[17:14] <hallyn> stgraber: would you agree that lxc-init should be in /bin or /sbin, not /usr/lib/lxc ?
[17:14] <kirkland> SpamapS: tell me this, can I require one to be the default, within my package?
[17:15] <hallyn> stgarber: it's an issue if we switch lxc to be multiarched (as debian does)
[17:15] <SpamapS> kirkland: I don't think so, but thats beyond my knowledge. You can just explicitly call the interpreter with the highest version available though.
[17:15] <SpamapS> kirkland: nothing from 2.5 is going to interfere if you run 'python2.6'
[17:18] <kirkland> I think my mom told me I needed to attend more "canonical maas"
[17:19] <kirkland> she makes me feel guilty about it too
[18:39] <cjs226> i have a server that seems to be operating fine, has a minimal load average (0.10, 0.16, 0.22), low CPU, has ~52MB RAM free but is using 1.5GB of swap (out of 5GB).  why is it swapping so much when it has freem RAM?  should this be a concern?
[18:56] <patdk-lap> cjs226, that is hardly enough info
[18:57] <patdk-lap> and I wouldn't call, 52mb enough space to fit 1.5gb of data
[18:59] <cjs226> patdk-lap: thx for replying.  i realize 52mb < 1.5gb.  I'm assuming if the system was truly swapping to the point of using ~30% of swap space, 1) it'd be noticeably slow and 2) the load average would be relevant
[18:59] <cjs226> and using all of it's RAM
[19:02] <patdk-lap> load average wouldn't matter
[19:02] <RamJett> What's the protocol here for re-asking a question? How long should I wait to ask again?
[19:02] <patdk-lap> loadaverage is a poor method to determin in a system is overloaded
[19:03] <patdk-lap> the whole idea of swap is, it swaps out things that aren't used
[19:03] <patdk-lap> if it starts swapping out things that are used, then you will defently notice it going slow
[19:04] <patdk-lap> can you post the output of "free -m"
[19:05] <cjs226>              total       used       free     shared    buffers     cached
[19:05] <cjs226> Mem:          7466       7418         48          0          6         93
[19:05] <cjs226> -/+ buffers/cache:       7317        148
[19:05] <cjs226> Swap:         5119       1503       3616
[19:06] <patdk-lap> hmm, 148megs free
[19:06] <patdk-lap> depends on your disk workload, but that is way too tight for my normal workloads
[19:06] <patdk-lap> I normally need a nice 2gigs in cached
[19:09] <cjs226> so, you normally run with 2GB free?
[19:09] <patdk-lap> no
[19:09] <patdk-lap> I normally run with 2gigs in cached
[19:09] <cjs226> this is a worker system running on EC2 that we run to exhaust a que of work then is terminated.  we're wanting to bleed it dry.
[19:10] <patdk-lap> but cached is just free memory that is being used to speed up the disk
[19:10] <patdk-lap> well, if you don't have any wait i/o %, as long as that stays low, it's ok
[19:10] <patdk-lap> if it's getting high, you might want some more cached ram
[19:11] <patdk-lap> it's something you will have to play with, is you usage i/o heavy?
[19:11] <patdk-lap> if so, you might be able to cut the time in half, by giving it more memory, and saving money overall
[19:11] <cjs226> no, we try to stay in ram
[19:11] <patdk-lap> or if it's purely more cpu workload, then it won't matter
[19:11] <cjs226> iowait, according to iostat, stays around #%
[19:12] <cjs226> er, 3%
[19:12] <patdk-lap> then your likely just cpu bound, or limited by exernal stuff, like tcp connections and wait time
[19:13] <cjs226> well, i don't see any slowness at all, I'm just concerned about the amount of swap being used
[19:13] <patdk-lap> so?
[19:14] <patdk-lap> as long as it's not slowing you down
[19:14] <patdk-lap> it sounds like your application loaded a crapload of stuff, but doesn't use it often
[19:14] <patdk-lap> like if it's running mysql, I would say you made your memory buffers too large
[19:14] <patdk-lap> it's keeping idle, rarely used info, in ram, when it could just read it from disk again, as needed
[19:15] <patdk-lap> but in this case, it's reading from disk swap as needed
[19:15] <patdk-lap> no real difference
[19:15] <cjs226> that's what I'm thinking as well, I was just wondering if i should be concerned given the amount of swap being used.  thx for the info
[19:15] <patdk-lap> now if you where ALWAYS accessing swap, then you DO need more ram
[19:15] <cjs226> right
[19:15] <cjs226> and if so, I'd assume my iowait would be much higher, correct?
[19:16] <cjs226> well, with conventional storage
[19:16] <patdk-lap> well, you would notice pauses
[19:16] <patdk-lap> use vmstat and look at the swapin/out columns
[19:16] <cjs226> procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
[19:16] <cjs226>  r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa
[19:16] <cjs226> 14  0 1544900  55712   5380  79332   19   33   273   215  447  275 10  3 84  3
[19:16] <cjs226> clif@mail-sched-02:/var/log/spanning$ vmstat
[19:16] <cjs226> procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
[19:16] <cjs226>  r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa
[19:16] <cjs226>  0  0 1544900  55844   5388  79360   19   33   273   215  447  275 10  3 84  3
[19:16] <cjs226> clif@mail-sched-02:/var/log/spanning$ vmstat
[19:16] <cjs226> procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
[19:16] <cjs226>  r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa
[19:16] <cjs226>  0  0 1544900  55968   5388  79360   19   33   273   215  447  275 10  3 84  3
[19:16] <cjs226> clif@mail-sched-02:/var/log/spanning$ vmstat
[19:16] <cjs226> procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
[19:16] <cjs226>  r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa
[19:16] <cjs226>  0  0 1544900  55968   5388  79360   19   33   273   215  447  275 10  3 84  3
[19:16] <cjs226> clif@mail-sched-02:/var/log/spanning$ vmstat
[19:16] <cjs226> procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
[19:16] <cjs226>  r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa
[19:16] <cjs226>  0  0 1544900  55968   5388  79360   19   33   273   215  447  275 10  3 84  3
[19:17] <guntbert> !pastebin | cjs226
[19:18] <cjs226> doh, sorry 'bout that
[19:18] <cjs226> http://paste.ubuntu.com/946182/
[19:18] <cjs226> i just don't see the system using swap in a detrimental way
[19:19] <cjs226> even though it's using a good deal of it
[19:22] <patdk-lap> well, that is what swap is suppost to be for
[19:22] <patdk-lap> I never have workloads like that :(
[19:23] <cjs226> thx for your help
[20:39] <Psi-Jack> Anyone here ever linked up an ubuntu server up to a FreeIPA server? I'm trying to figure out how plausible it is, as I already have the FreeIPA server up and running now.
[21:44] <jkyle> I'm having an issue where sometimes the grub menu timeout is ignored and clients "stick" at the grub boot menu
[21:44] <jkyle> 12.04
[23:26] <jkyle> I have an issue where grub sticks at the boot menu on the first reboot after a pxe install
[23:26] <jkyle> never timing out