[00:05] <jdstrand> hallyn: fyi, jjohansen1 is working through it over the weekend, but it won't be tonight
[00:06] <jdstrand> hallyn: last I heard he is striving for tomorrow midday (CDT)
[00:16] <jjohansen1> oh please, no.  I am striving for tonight, so it can be uploaded as soon as people come online tomorrow.  I believe it is actually working now, but I added new regressions tests and some are failing and I am debugging through it.
[00:39] <twb> Ballpark, what's the typical total throughput in gbps or tbps for a larga data center or IXP?
[00:51] <jdstrand> jjohansen1: I more meant striving for the package being available by midday, but I'll prepare/test/upload when you are ready
[00:52] <oApocalypse> hello i have a samba permission issue can anyone help
[00:53] <seekwill> What's the problem?
[00:53] <oApocalypse> well i have the availability to read but not write
[00:53] <seekwill> smbd.conf ? ls -la /path/to/folder ?
[00:55] <oApocalypse> the way i have it set in smb im utilizing username map also
[00:55] <oApocalypse> but ill tell u what i have set as far as the share
[00:55] <oApocalypse> security is set to user
[00:56] <oApocalypse> create mask = 0770 writable = yes valid user =@mygroup  guest ok = no read only = no
[00:56] <seekwill> writable = yes
[00:56] <seekwill> oh
[00:56] <seekwill> Pastebin works better :)
[00:57] <oApocalypse> when i rightclick in putty it pastes it over and over lol
[00:57] <oApocalypse> mind u im new so all i have for experience is reading
[00:57] <oApocalypse> i have a workgroup set also
[00:57] <oApocalypse> i tried to get to it from a non work group computer
[00:58] <oApocalypse> it asked me for credentials i entered valid group credentials and its still read only
[00:58] <oApocalypse> the device itself has permissions set at 0770 also
[00:58] <oApocalypse> (HD)
[00:59] <seekwill> What does ls -la /path/to/dir show?
[01:00] <oApocalypse> root group and 4096?
[01:00] <oApocalypse> drwxr-xr-x
[01:00] <oApocalypse> again not fluent in linux :P
[01:00] <oApocalypse> but i think thats 0770 lol
[01:00] <oApocalypse> group being the group that i set of course
[01:01] <seekwill> Copying and pasting the output of the command helps
[01:01] <oApocalypse> how to on putty
[01:01] <oApocalypse> it doesnt give optioin
[01:02] <seekwill> ?
[01:02] <seekwill> Give the option for what?
[01:02] <oApocalypse> # ls -la /Share
[01:02] <oApocalypse> total 8
[01:02] <oApocalypse> drwxr-xr-x  2 root torrent 4096 2012-03-25 13:21 .
[01:02] <oApocalypse> drwxr-xr-x 24 root root    4096 2012-03-25 13:17 ..
[01:02] <oApocalypse> -rw-r--r--  1 root torrent    0 2012-03-25 13:21 test
[01:02] <oApocalypse> #
[01:03] <seekwill> yeah
[01:03] <seekwill> Your torrent user group does not have permission to write
[01:03] <seekwill> It's not 770
[01:03] <seekwill> chmod -R 770 /Share
[01:04] <oApocalypse> ur the man
[01:04] <oApocalypse> ok where do u see this
[01:04] <seekwill> drwxr-xr-x  2 root torrent 4096 2012-03-25 13:21 .
[01:04] <seekwill> . is the current directory
[01:05] <seekwill> drwxr-xr- : 'd' it's a directory, 'rwx' read write execute for the owner (root in this case) 'r-x' read and execute (no write)
[01:05] <oApocalypse> nice
[01:05] <oApocalypse> heres another question
[01:05] <seekwill> Oh, you've used up your question allocation for the day
[01:05] <oApocalypse> why can i access this outside of the workgroup if i credentials
[01:05] <oApocalypse> LOL
[01:06] <oApocalypse> damn next time
[01:06] <seekwill> What do you mean?
[01:06] <oApocalypse> well i have a desktop off the workgroup
[01:07] <oApocalypse> and i was able to get onto it
[01:07] <oApocalypse> shouldnt me defining a workgroup only allow those on the same workgroup
[01:07] <seekwill> Workgroups aren't really security groups
[01:07] <seekwill> From my understanding at least
[01:08] <oApocalypse> no for instance i created a folder in this directory
[01:08] <seekwill> What directory?
[01:08] <oApocalypse> in /Share
[01:08] <oApocalypse> with one of my usernames
[01:08] <seekwill> I can't see your screen (not yet at least, give me a sec), so you need to be specific
[01:09] <oApocalypse> # ls -la /Share
[01:09] <oApocalypse> total 12
[01:09] <oApocalypse> drwxrwx---  3 root torrent 4096 2012-03-25 21:04 .
[01:09] <oApocalypse> drwxr-xr-x 24 root root    4096 2012-03-25 13:17 ..
[01:09] <oApocalypse> drwxr-xr-x  2 mike mike    4096 2012-03-25 21:04 New folder
[01:09] <oApocalypse> -rwxrwx---  1 root torrent    0 2012-03-25 13:21 test
[01:09] <oApocalypse> #
[01:10] <oApocalypse> does that mean new folder will be only accessable by mike?
[01:10] <seekwill> Mike is such a troublemaker...
[01:10] <oApocalypse> yea i hate him
[01:10] <oApocalypse> lol
[01:10] <seekwill> Pastebin your smbd config
[01:10] <seekwill> The part about your share
[01:11] <seekwill> But yeah, that's what it means
[01:11] <seekwill> I'm guessing you want to make this directory "public" for anyone who is in the torrent group?
[01:11] <oApocalypse> yes
[01:13] <a5m0> if i got one of the 12.04 beta downloads would it update itself to the final release at a later date?
[01:13] <oApocalypse> [Share]
[01:13] <oApocalypse>  comment = torrent drive
[01:13] <oApocalypse>  path = /Share
[01:13] <oApocalypse>  create mask = 0770
[01:13] <oApocalypse>  writeable = yes
[01:13] <oApocalypse>  valid user = @torrent
[01:13] <oApocalypse>  guest ok = no
[01:13] <oApocalypse>  read only = no
[01:14] <seekwill> Pastebin works great :)
[01:15] <shauno> a5m0: that's the plan, but there's no guarantees you might accumulate some 'cruft'. I prefer a clean install for production, but upgrading from beta does work
[01:15] <oApocalypse> lol i have to copy to clipboard then edit it so no its a pain in the ass
[01:15] <oApocalypse> whats pastebin
[01:17] <seekwill> oApocalypse: You paste all that to a webpage, and it gives you a URL. You paste just the URL here
[01:17] <seekwill> oApocalypse: Did you check this? http://www.samba.org/samba/docs/man/manpages-3/smb.conf.5.html for config options?
[01:18] <oApocalypse> what my question is  will things i put into a directory other than folders ahve that same read write bullshit
=== Jasonn is now known as juicy
[01:23] <locuse> hi.  i installed my 1st ubuntu-server 10.04LTS instance, for production use with Zimbra.  so far so good.  i want to stay as much as possible "on" the LTS release (+security updates) -- but would like to upgrade to the latest possible openssh (v5.9).  is that possible with the 10LTS release?  i.e. just one package, without derailing the 'rest' of the LTS release?
[01:32] <seekwill> locuse: Why do you want to?
[01:33] <locuse> seekwill: i want access to features in newer versions of openssh.  e.g., multiple AuthorizedKeyFiles + auto fallback ...
[01:33] <seekwill> I believe you can upgrade that without breaking everything else.
[01:35] <locuse> seekwill: at least worth a shot ... where's the "right place" in ubuntu-land (i'm used to opensuse, til now) for finding and installing a "latest" openssh package/repo that would install into 10LTS?
[01:37] <seekwill> locuse: I'm not sure on Ubuntu either, but I'd check 12.04 LTS to see if it comes with the version of openssh you want, then you won't have to go through this :)  If that fails, see if there's a precompiled package in a repo somewhere.
=== Arc_ is now known as a5m0
[01:38] <locuse> seekwill: atm, Zimbra is !supported on 12.04LTS, hence the issue ...
[01:38] <seekwill> ah
[01:38] <locuse> eventually, sure -- but not until "at least April", iiuc
[01:39] <seekwill> Can you live without that feature till then? :)
[01:40] <seekwill> If you don't find any other repos that you trust, I'd set up another 10.04LTS machine as your dev machine (you should have one already) and roll your own package, then deploy the .deb to your prod machine
[01:40] <locuse> seekwill: unfortunately, i suspect the 'old' ssh is a source of a problem.  if i have to, i'll build openssh in /usr/local and be done with it ...
[01:40] <locuse> haven't *ever* rolled a .deb ...
[01:41] <seekwill> I haven't either! :) I'm assuming it's the same as rolling an RPM though
[01:44] <locuse> seekwill: heh, apparently nothing is the same upon switching distros! ;-)
[01:47] <locuse> seekwill: closer -> https://launchpad.net/~jamon/+archive/backports.  it's v58, not v59 ...
[01:47] <oApocalypse> ty for help seekwill ive found a workaround
[01:48] <seekwill> locuse: doh!
[01:48] <seekwill> oApocalypse: Nice :)
[01:48] <oApocalypse> any idea why a 1tb drive would show up at 160gb networks share?
[01:49] <locuse> seekwill: is that a good "doh", or a bad "doh"?
[01:49] <hallyn> jdstrand: thanks for that info
[01:50] <twb> locuse: IIRC multiple authorized keys files are supported in lucid's openssh-server...
[01:51] <locuse> twb is 'lucid' 104LTS?
[01:51] <twb> lucid = 10.04 LTS, yes.
[01:52] <locuse> twb: tried it, didn't work -- complains about "garbage" in the config file if i add a second authkeyfile ...
[01:52] <twb> Well, I could be wrong
[01:53] <twb> I thought I tested it for monkeysphere
[01:53] <locuse> twb: as i understand it, didn't arrive til v5.9 -> http://lists.debian.org/debian-ssh/2011/09/msg00014.html
[01:53] <twb> Fair enough
[01:53] <locuse> the server i installed has v5.3
[01:53] <locuse> basically ancient
[01:54] <twb> I can confirm that
[01:59] <michaelh1> Hey, does anyone know who in Ubuntu creates the ARM UEC images http://uec-images.ubuntu.com/precise/current/ and the tool used?
[01:59] <locuse> hm, this looks like a path to a newer-openssh deb ... http://askubuntu.com/questions/100840/openssh-5-9p1-on-ubuntu-11-10
[02:00] <twb> michaelh1: try #ubuntu-arm
[02:01] <michaelh1> twb: will do
[02:01] <oApocalypse> why would a hd i shared looks like a 142gb hd
[02:02] <oApocalypse> does samba segment the drive automatically for different users
[02:04] <twb> uupdate -v 5.9p1 /root/packages/openssh/openssh-5.9p1.tar.gz
[02:04] <twb> locuse: would be better to cherry-pick the package from precise or whatever
[02:33] <hallyn> stgraber: I'm about to push an lxc that temporariliy does the temp proc mounting to fix your issue.  i'll leave it to you to do what's needed to get it through?
[02:34] <twb> hallyn: push it to all of ubuntu, or a staging/experimental area?
[02:35] <hallyn> all of ubuntu
[02:35] <twb> That sounds suboptimal for a temporary test...
[02:35] <hallyn> ican push it to staging area in the meantime...
[02:35] <hallyn> it's not a temp test
[02:35] <hallyn> it's a temp fix until a dependency gets fixed
[02:35] <twb> ah, ok
[02:37] <stgraber> hallyn: ok, btw, jjohansen1 is planning an apparmor upload tomorrow morning
[02:37] <hallyn> oh i thought it was afternoon
[02:37] <hallyn> stgraber: it's up to you
[02:37] <hallyn> i'll happily wait
[02:37] <hallyn> less work and risk for me
[02:38] <hallyn> but this is working (finally.  two days of one little stupid mistake after another)
[02:38] <jjohansen1> hallyn: it was but I got some failures in new regression tests I added am working through them, I think most of them are actually in the tests but I want to make sure all the regressions are working
[02:38] <hallyn> jjohansen1: sorry i'm not following - so now it's going to morning or afternoon?
[02:39] <hallyn> mind you i'm not trying to rush you :)
[02:39] <jjohansen1> hallyn: I will have the upload waiting for when jdstrand comes on in the morning
[02:39] <locuse> twb seekwill thanks, will dig in precise.
[02:40] <stgraber> hallyn: can you push to a separate branch so we have it around just in case?
[02:40] <hallyn> jjohansen1: ok, thanks.  stgraber: so i'll wit?
[02:40] <hallyn> ok
[02:41] <stgraber> I'll be in #ubuntu-release all week, so I'll check for the apparmor upload and do the required nagging to get it in, making sure it fixes the problem here and then rebuild all the images
[02:41] <jjohansen1> stgraber: yeah I can setup another branch in a bit, I was just heading to dinner
[02:41] <stgraber> jjohansen1, hallyn: Thanks for tracking this one down and fixing it so quickly
[02:50] <hallyn> stgraber: lp:~serge-hallyn/ubuntu/precise/lxc/lxc-aa-proc
[02:55] <twb> grumble grumble direct root logins not enabled with keys, grumble
[02:56] <hallyn> denied from sshd_config ?
[02:56] <twb> correct
[02:57] <twb> just allowedusers cyber@<my class C> and *@<customer NAT class C>
[02:57] <hallyn> yeah i seem to recall that slowing me down when i was testing some remote libvirt stuff that required root
[02:57] <twb> The most important case is: ssh root@foo tar -C /foo | ssh root@bar tar -C / -x
[02:58] <twb> Because in that case both ends need to be root and you can't easily talk to stdin to escalate through e.g. sudo
[02:58] <twb> (Yes, you could use netcat, but that would send the tarball as cleartext over the networl.)
[02:59] <hallyn> so you netcat to a local port from whcih you forward through ssh, and ...
[02:59] <hallyn> j/k
[03:00] <hallyn> i think i'm done for the day.  it's been just a pathetic performance on my part.  i'm embarassed.  (or would be, if anyone'd been watching)
[03:00] <hallyn> gnight all
=== Arc_ is now known as a5m0
=== michaelh1 is now known as michaelh1|away
[05:37] <twb> I have a PHP weenie who is currently OOMing his allocated resources by running PHP lucene.  He suggests migrating to solr, but AFAICT that needs an entire tomcat stack, AND he'll still be running the rest of his app in PHP, so I'm not sure how this would be a win
[05:38] <twb> Am I right?  Would solr be a pointless memory hog?
[06:02] <imbrandon> twb: no solr is worth it
[06:03] <imbrandon> and the php solr api too that makes it more so
[06:03] <twb> This is not some ecommmerce site or anything, it's a flipping issue tracker for a ten-person company
[06:04] <twb> If it was up to me there the solution would be "STFU if you want body searches, use the IMAP backend"
[06:04] <twb> :-/
[06:04] <imbrandon> that sounds like it would need it more than a ecommerce site
[06:04] <imbrandon> :)
[06:04] <twb> I'm not happy about him using PHP either, but I was overruled for that one host
[06:05] <imbrandon> twb: heh i think you might be on the wrong end of the stick this time, what version of php and whats does your metrics like new relic say the memeory is ued by
[06:05] <imbrandon> rearely php casues it
[06:05] <twb> I was getting OOM reports from lucene via syslog
[06:05] <imbrandon> yea i hear that part
[06:06] <twb> Gimme  a minute and I'll pull them up for you, just got to deal with this (unrelated) security fuckup first...
[06:06] <imbrandon> but thats just the straw
[06:06] <imbrandon> that broke its back
[06:06] <imbrandon> you need the metrics
[06:06] <lifeless> twb: solr is a good thing, decouples your bits, will have less overhead on starting up php, lets you query it from other languages etc
[06:07] <twb> But it DOES mean I need a stupid jvm and tomcat, right?
[06:07] <imbrandon> well yea but if done properly it will be much better
[06:07] <twb> Will tomcat even fit into a 1GB RAM VM?
[06:07] <imbrandon> do you have a sysadmin handy ?
[06:07] <twb> I am the sysadmin
[06:07] <twb> It's my job to say "no" to the devs
[06:07] <lifeless> thats sad
[06:07] <imbrandon> ohhh you called a coder whinnie and tomcat stupid
[06:08] <imbrandon> those are very nice tools if used properly
[06:08] <imbrandon> solar and php both can be setup to not use a ton of ram
[06:08] <twb> imbrandon: all of "web app" space is stupid a priori, but tomcat is especially stupid because I am called in all the time to customer sites where it wants like 8GB of RAM just to do some trivial job for half-a-dozen users
[06:08] <imbrandon> try apc and memcache and then run newrelic to see where itss actually going
[06:09] <twb> this "newrelic" stuff is a property of what, php or lucene?
[06:09] <lifeless> twb: so use a lighter weight container, or get your instrumentation up and look at the data
[06:09] <twb> lifeless: they're LXC containers already
[06:09] <lifeless> newrelic is a company that runs a webstat instrumentation service
[06:09] <imbrandon> ties code metrics, it watches all kinda of code
[06:10] <imbrandon> ass see whats taking cpu ram disk io
[06:10] <lifeless> twb: nothing to do with LXC. Java container.
[06:10] <imbrandon> etc down to the functino
[06:10] <twb> lifeless: I don't even have java installed yet!
[06:10] <lifeless> I get that
[06:10] <lifeless> but you are ranting about something you don't have the basic terminology for.
[06:10] <lifeless> perhaps you should spend a little time learning :)
[06:10] <imbrandon> newrelic is totaly seperate, think of it like a tool, its not a part of the stack
[06:11] <twb> lifeless: so far it has been easier to install infrastructure that isn't bloated by default
[06:11] <lifeless> twb: thus my suggestion to use a different container
[06:11] <lifeless> anyhoo, I have to run, ciao
[06:11] <imbrandon> i dont know of a nother j2ee stack thats lighter than tomcat :) maybe try glasssfish ?
[06:11] <imbrandon> l8tr lifeless
[06:12] <twb> Does "j2ee" also imply it would have to be Sun java?
[06:12] <lifeless> imbrandon: spring
[06:12] <imbrandon> no
[06:12] <imbrandon> j2ee is a api for java
[06:12] <twb> Well at least there's that.  Oracle are also on my shit list
[06:12] <imbrandon> hum
[06:13] <imbrandon> lifeless: ahhh , mostly just deal with the java bridge in zend server
[06:13] <imbrandon> last few years
[06:13] <lifeless> imbrandon: also there is jboss
[06:13] <lifeless> and there is another one whose name escapes me just now, that jenkins uses by default
[06:14] <imbrandon> oh yea i actually forgot about that one, jboss is pretty sweet
[06:14] <imbrandon> thought jenkins was on tomcat default
[06:14] <imbrandon> maybe not though
[06:14] <imbrandon> i set it up and forgot about it :)
[06:15] <twb> lifeless: ah, you meant https://en.wikipedia.org/wiki/Web_container (right?)
[06:17] <imbrandon> btw : 06:07 < twb> It's my job to say "no" to the devs  <--- you need to do a little more reasearch on your job then, i dont know of anyone who's job it is to tell devs "no" except a CIO, DevOps or in your case sounds like Ops is there to facilitate the developer doing their job while protecting the companies intrest, even if you did have a propblem with a developers way of doing things its still above the ops paygrade to make the call 
[06:18] <twb> I can assign myself a different title if you prefer :-)
[06:20] <imbrandon> would no make a diffrence if that is how you look at the developers and you ARE in power over them, either A) you dont trust them and that should end the relationship right there, why have someone work under you you dont trust, or B you cant fire them and your overestimating your power over them
[06:21] <imbrandon> so its moot
[06:21] <twb> I don't trust the webdev because he's an idiot, but I don't have the power to sack him
[06:21] <imbrandon> ah then eat humble pie and do your job
[06:21] <twb> I have the power to say to management "I recommend option <B> because it is not retarded"
[06:21] <imbrandon> as in learn how to make a java stack
[06:22] <twb> That may well be option B, but if the webdev is only recommending solr because he read about it on reddit, it is my job to determine if there is a better solution
[06:22] <imbrandon> ok well this isnt productive, we could go on all day, if you have some more tech questions i'd love to help
[06:22] <twb> Where "better" doesn't necessarily mean "faster" but e.g. could mean more secure
[06:23] <imbrandon> and if you think solr is something from reddit and never heard of newrelic you sorely need to brush up on your ops skills
[06:23] <imbrandon> i understand exactly what it means , been at this almost 20 years
[06:25] <imbrandon> i do need to stop the back and fourth though, its disruptive in here, i'm happy to help technicly if you want recomendatinos on a good setup
[06:25] <imbrandon> but cant contiunue the back and fourth
[06:27] <twb> Shrug.  So far your suggestions have been to profile and find out where the problem actually like (agreed); and that solr is worth the overhead of learning to administer a java web stack (of which I'm still not convinced).
[06:27] <twb> If and when I establish that java really is needed / worth it, I'll be happy to discuss ricing it up
[06:27] <imbrandon> thats your call, i can tell you i implmented it this week , publicly, on a site with about 10.7 million hits a month
[06:28] <imbrandon> and its all documented on the whos and whys
[06:28] <imbrandon> if you want the papers and links
[06:29] <twb> I'm not interested in scaling to internet-sized systems
[06:29] <imbrandon> no but it sounds like you need to scale to one, and this was done on 2 machines, one web with 4gb ram total and one db mysql from oracle
[06:30] <imbrandon> thus it would still apply
[06:30] <imbrandon> honestly man, i';m not here to sell you on anything, i juist can tell you what my experince is and what i have to back that up, i have no need for you to beleaive me
[06:30] <twb> I'm not interested in infrastructure that would ALLOW me to scale to that size if it requires deploying infrastructure with which I'm not familiar.
[06:31] <imbrandon> take my adive or leave it , i'm not paid to be here :)
[06:31] <twb> Granted.
[06:32] <imbrandon> right but when your comfort zone dose not include a basic server like a java stack, or a solid php stack then i have shakey feeling your gonna find anything your familiar with to deploy
[06:32] <imbrandon> or you wouldnent be here
[06:32] <twb> Server ≠ web server
[06:33] <imbrandon> sure sure, semantics
[06:33] <imbrandon> you knew exactly what i mean
[06:33] <twb> What I babysit is mostly LDAP, postfix, dovecot, NFS, samba, etc.
[06:33] <imbrandon> ok
[06:34] <imbrandon> then grab your nearest web dude even if you dont trust him and have some faith
[06:34] <imbrandon> there is no replacement for expertise
[06:34] <twb> That would be you :-)
[06:35] <imbrandon> well i'm trying man but i'm getting shot down every corner by some 1999Linux users way of thinking
[06:35] <imbrandon> but i'mm willing to look past it
[06:35] <imbrandon> heh
[06:35] <twb> It is fair to say that IMO CLI beats GUI beats web UI :-)
[06:35] <imbrandon> and dont get me wrong , i started using linux with rh 4 arround 96ish :)
[06:36] <imbrandon> i'm in irrssi and screen now
[06:36] <imbrandon> that dont mean i cant keep up with the latest trends of ohh say the last decade
[06:36] <imbrandon> this stuff isnt new man
[06:36] <imbrandon> its proven tech
[06:36] <imbrandon> i'm not saying go run nginx
[06:36] <imbrandon> :)
[06:37] <twb> I've done that
[06:37] <imbrandon> as have i, take a gander at brandonholtsclaw.com , just posed about it yesterday
[06:37] <imbrandon> :)
[06:37] <imbrandon> i'm gonna grab a soda, when i get back lets start over and see if we cant get you workin here
[06:38] <imbrandon> back in ~5 min
[06:40] <twb> btw newrelic appears to be newrelic.com, and AFAICT that's not DFSG-free so it would be an automatic fail here
[06:40] <imbrandon> its a service, what does dfsgfree have to do with it
[06:41] <twb> The actual ITS, which has been written by generations of work experience students, is http://allocpsa.net (darcs get --lazy http://dev.allocpsa.net )
[06:41] <imbrandon> and no theirs or any other metric provider that is not you looking over syslogs wont be either
[06:42] <twb> I'm currently using logcheck and collectd
[06:42] <imbrandon> rember your paiying for their time, ever herd the joke about the man that hit the furnis and charged $1000, the customer balked and the man said its $1 for the hammer and $999 for knowing where to hit it
[06:43] <imbrandon> twb: right but you cant tell me where the memory was going
[06:43] <imbrandon> e.g. able to intrreprit them
[06:43] <twb> If you just mean which process, sure, I could do that
[06:43] <imbrandon> no
[06:43] <imbrandon> i mean what function in what context
[06:43] <imbrandon> and how man calls to it
[06:43] <imbrandon> and was it cacheed
[06:44] <imbrandon> and if so where in apc or memcache
[06:44] <imbrandon> or redis
[06:44] <imbrandon> or ......
[06:44] <imbrandon> yea lots more to it than just the process , and i'm not the expert
[06:44] <twb> Right, that is the kind of problem I leave to the webdev because it's happening inside his web stack
[06:44] <twb> s/leave /normally leave/
[06:44] <imbrandon> thats not a web dev problem though , you are the one with the tools, like newrelic
[06:45] <imbrandon> you job dont stop at the stack man, thats the bigest downfall most sysops make
[06:45] <twb> well, normally profiling within the app would require loading magic inside the app, not just reading syslog
[06:46] <imbrandon> no it requires loading magic on the server to intercept the calls at the cpu
[06:46] <twb> And/or compiling the app (or the interpreter) with CFLAGS += -g and shoving it through gdb
[06:46] <imbrandon> in the app does no good
[06:46] <imbrandon> cuz then you modify the app
[06:46] <imbrandon> no and no
[06:46] <imbrandon> thats the way it was done a few 10 years ago or so
[06:46] <imbrandon> VT is all new now
[06:47] <twb> VT as in Intel VT-x?
[06:47] <imbrandon> yea thats what i was refering to
[06:47] <imbrandon> crap phone one sec man
[06:47] <twb> I'm already using that to isolate containers from one another
[06:47] <imbrandon> sorry
[06:47] <twb> np
[06:47] <imbrandon> i was getting at intercepting the calls there, you want to do it in a way that you DONT modify the code
[06:48] <imbrandon> e.g. with a newrelic type module
[06:48] <imbrandon> if not them similar
[06:48] <imbrandon> brb
[06:48] <twb> How is that different from strace and/or gdb ?
[06:48] <imbrandon> because your modifyting the code when recompiling it diffrent
[06:48] <twb> Or are you talking about instrumenting the PHP interpreter?  I'm a little confused.
[06:49] <imbrandon> yes
[06:49] <imbrandon> and yes
[06:49] <imbrandon> and really the whole server not just the php runtime, it hasent been an intrepiter since php 3
[06:49] <imbrandon> we're are 5.4 now
[06:50] <twb> If you're instrumenting the PHP runtime, you're inherently changing the runtime.  I grant you, that you're not changing the PHP app, i.e. the code written in PHP stays the same.
[06:51] <imbrandon> right but without the code changes, i am not enough of an expert to shoot down your every retort on this subjkect but i do konw if you research it a bit
[06:51] <imbrandon> it IS the way to go
[06:51] <imbrandon> and stop saying php, i mean EVERY PROCESS
[06:52] <imbrandon> mysql , top, w, whatever is runnining on the box
[06:52] <imbrandon> all of it at the cpu level
[06:52] <twb> So what, you're saying run that instead of kvm?
[06:52] <imbrandon> well sorta, depends on the kvm you are refering to but bsicly yes
[06:52] <imbrandon> thats what new relic does
[06:53] <imbrandon> than then chomps all the info and gives it to you in a way you can actually make actions on
[06:53] <twb> I'm surprised that gives meaningful information
[06:53] <imbrandon> like let me finish i'm still on the phone but when i'm done i'll show you my new relic console
[06:53] <twb> Sure sure
[06:53] <imbrandon> and show you the actionalble data
[06:55] <imbrandon> and why an approach like this is so much more than just sysog etc, in other words you wouldne tbe here, you would be a) making a java server cuz he was undenyably right,m or shoveing what exact part of the code is broken in his face, but this time with solid proof
[06:55] <imbrandon> not a fillled log
[06:55] <imbrandon> that could have filled with anything then boke the camels back on that
[06:55] <imbrandon> or hell it could be a php bug its self and not his code
[06:55] <imbrandon> thats what we'll beable to see at a glance
[06:56] <imbrandon> thats HARD to do with sysslog
[06:56] <twb> Granted.
[07:01] <imbrandon> back, ok give me sec to get the newrelic up
[07:02] <twb> btw, the actual (unhelpful) error that sparked this is:
[07:02] <twb> [Fri Jul 01 12:53:09 2011] [error] [client 203.7.155.8] PHP Fatal error:  Allowed memory size of 268435456 bytes exhausted (tried to allocate 80 bytes) in /var/www/alloc-stage/zend/Zend/Search/Lucene.php on line 1136, referer: https://alloc-stage.cyber.com.au/task/taskList.php?
[07:03] <imbrandon> yea all that means is the setting in the php.ini is too low
[07:03] <imbrandon> ok got a browser up ?
[07:03] <twb> Well, the web dev said we should 1) increase the cap; and 2) migrate from PHP lucene to solr
[07:04] <imbrandon> yup, hes right on both
[07:04] <imbrandon> on most prod boxes i set it about 512m
[07:04] <imbrandon> and then never think twice about not using solar
[07:04] <twb> That feels like giving up
[07:05] <imbrandon> solr is likeapche of search
[07:05] <twb> Well, it *is* apache solr/lucene :-)
[07:05] <imbrandon> twb: you got to loose that feeling man, it only will drive ya nuts, i am not saying give in to eveyr little whim or 8gb for 3 users
[07:06] <imbrandon> but both of those are VERY VERY reasonalble requests
[07:06] <imbrandon> like so reasonable i'm suprise YOU dident sugest them to him
[07:06] <twb> Well, he said that php lucene is loading the entire lucene db into memory in ever php process, and I went "WTF GTFO"
[07:07] <imbrandon> well only about half right
[07:07] <imbrandon> but yea
[07:07] <imbrandon> thats now the the zf1 ( what your using ther e) works
[07:07] <twb> You don't give in to design like that, you put your foot down and say "go back and do it properly"
[07:07] <imbrandon> zf2 and php 5.3 does a much better job and 5.4 even more so
[07:08] <twb> php5.3 here
[07:08] <imbrandon> twb: thats just it there is no proberly unless its from anther stack, that IS properly
[07:08] <twb> He's using a convenience copy of lucene so I can't easily see what version
[07:08] <imbrandon> and thats when i tell a sysop that tells me know to start packing his desk cuz as soon as i'm off the phone with his boss thats what would happen
[07:08] <imbrandon> no*
[07:09] <imbrandon> exactly twb and he is trying to get you to load the industry best practice of solr
[07:09] <imbrandon> that you scoff at
[07:09] <imbrandon> so what is he SUPOSSE tyo do
[07:09] <twb> imbrandon: well, the "right" way would be a separate process that PHP talks to via a socket or port, and it just runs continuously and responds to requests.
[07:09] <imbrandon> code it in asm >
[07:09] <twb> Which I grant you is AIUI what solr does, I just have a knee-jerk reaction to deploying java to SOLVE memory issues
[07:10] <imbrandon> php dont work like that, thats what things like solar and nodjs loadn hals are for
[07:10] <imbrandon> exactly what they are for
[07:11] <twb> loadn hals?
[07:11] <imbrandon> twb: this gut feeling and kneejerk reaction is what is kiling you , not the dev, have you actually profiled modern j2ee stacks ?
[07:11] <twb> my experience with j2ee stacks is limited to customers asking me to increase their mail or web server's allocation from 8GB to 16GB because "8GB isn't enough [for ten users]"
[07:12] <twb> Oh and asking me to restart it when it crashes
[07:12] <imbrandon> that may have been the case of java is a hog long ago, but vb6 was shittly long ago, and so was php, now php runs 2 of the 3 most traffic websites on the planet, with that kinda of backing you can bet your not the first sysabmin to have this issue, you just need to fiund who solved it best
[07:13] <twb> AFAICT the "solutions" for PHP, at least, are to put a great big honking caching reverse proxy in front of it
[07:13] <imbrandon> twb: the more you talk them more i would be scared to work at your company, it soulds like youve been at the same desk since 1999 and not left per se and learned anything new
[07:13] <imbrandon> no its to run FPM actually
[07:13] <imbrandon> and a little reverse microcache maybe
[07:13] <imbrandon> maybe
[07:14] <twb> Is that the fancy new stuff that byte-compiles PHP to an IR?
[07:14] <imbrandon> no
[07:14] <imbrandon> and php has been mytecode compiled since zend engine 1
[07:14] <imbrandon> e.g. php 4.0.0
[07:14] <imbrandon> its now "fancy nerw stuff"
[07:14] <twb> Oh *that* fpm
[07:15] <imbrandon> php 4.0.0 came out what 12 years ago
[07:15] <imbrandon> i konw i keep saying it man but your your own worst enemey here , i almost need to checdk if i'm just being trolled
[07:15] <uvirtbot> New bug: #965022 in openssh (main) "ssh crashed with SIGSEGV" [Undecided,New] https://launchpad.net/bugs/965022
[07:16] <imbrandon> twb: fpm a.k.a fast page module, a.k.a fastcgi
[07:16] <imbrandon> not a new tech
[07:16] <imbrandon> maybe a new name and a new way to implment it
[07:16] <imbrandon> but been arround ages
[07:17] <imbrandon> and if you need a reverse proxy its probably apache killing you not the php
[07:17] <imbrandon> as demonstrated in that blogpost
[07:17] <imbrandon> , ok i have my passwd on newrelic set to a temp one if you wanna poike arround a little
[07:18] <twb> msg me the credentials or something?
[07:18] <imbrandon> its running on my blog that i just deployed 2 days ago and not been optimized yet, so lots of nice little errors and such to see
[07:18] <imbrandon> sure
[07:18] <imbrandon> one sec
[07:19] <twb> I log in from newrelic.com, or from your site?
[07:19] <imbrandon> newrelic.com
[07:20] <twb> hm, might as well use a GUI browser...
[07:20] <imbrandon> you'll want the app that says brandonholtsclaw.com zerox@linode
[07:20] <imbrandon> uh yea
[07:22] <twb> It's listing dev/live/test no brandonholtsclaw
[07:22] <imbrandon> hit the dropdown
[07:22] <imbrandon> in the top right
[07:22] <imbrandon> it needs to say xerox@linode
[07:22] <imbrandon> near the logout, prbably says enews pro or something now
[07:23] <twb> got it
[07:23] <imbrandon> kk, poke arround in there i actually need to take another call
[07:23] <imbrandon> trusting a stranger here
[07:23] <imbrandon> heh dont blow it up
[07:24] <twb> It's struggling a bit, probably either the .au / .us link, or because I'm on an ARM netbook with zero video acceleration
[07:24] <imbrandon> possibly, but the datapoints are the real key
[07:24] <twb> mm
[07:24] <imbrandon> notice all the toip greay bar
[07:25] <imbrandon> there it lots of metiric and i only have some loaded
[07:25] <imbrandon> but see it break down to the fucntin calll level
[07:25] <imbrandon> and then what inside the functino even
[07:26] <twb> OK so it can profile PHP files, modules and modules' SQL queries
[07:26] <imbrandon> deeper than file
[07:26] <imbrandon> but yea
[07:26] <imbrandon> much more detailed than syslog
[07:27] <imbrandon> with no code changes
[07:27] <imbrandon> also note on the fron page
[07:27] <imbrandon> i'm getting what about 2 hits per minute on my blog
[07:27] <imbrandon> basicly idle
[07:27] <twb> Where it says memcache, is that memcached?
[07:27] <imbrandon> and its sitting at 455mb ram un optimized with solar
[07:27] <imbrandon> yes
[07:28] <imbrandon> memcache is the api ( now part of mysql from oracle ) that memcached implments
[07:29] <imbrandon> but the new interface to the mysql ndb server long with handlersocket is msuch nicer now, it will be mainstream in another 3 to 6 months
[07:30] <twb> FWIW I doubt that's instrumenting at the VT/KVM layer, I expect that's more more likely to be instrumenting the PHP runtime plus collecting stats from memcached and friends, and maybe sniffing packets
[07:30] <imbrandon> sureand it can
[07:30] <twb> Pulling that data out by watching x86 instructions would be a real bitch
[07:30] <imbrandon> like i said i dont have most of it loaded
[07:30] <twb> OK
[07:30] <imbrandon> see all the empty blanks
[07:31] <imbrandon> like all of background tasks is empty
[07:31] <imbrandon> i only have the raw basics going as i'm in the mid of a redesign
[07:31] <imbrandon> on that blog, thus only one post re-imported too
[07:31] <imbrandon> :)
[07:31] <twb> Oh "background" tab doesn't fit on my screen
[07:31] <twb> I can just see "ba..."
[07:31] <imbrandon> nother 24 or 48 hours and it will be back to norm
[07:32] <imbrandon> ahh well its empty anyhow
[07:32] <imbrandon> heh
[07:33] <imbrandon> anyhow youve seen the tools and know my opinon, i can give you sites to back it up but in reality ANYWHERE you look its best practice so just pick somewhere
[07:33] <imbrandon> but in the end he is right this time
[07:33] <imbrandon> maybe not every time
[07:33] <twb> Well fair enough
[07:33] <imbrandon> but solr and 512m in the php,.ini is the way to go
[07:33] <twb> Definitely not every time -- we had to explain to him that double-ledger accounting was a good idea, and using floating-point for financials wasn't :-)
[07:34] <imbrandon> heh sure, he is a coder not an accountant :) i'd expect that
[07:34] <imbrandon> i;d be sacred if he did know
[07:34] <imbrandon> well maybe  ot but you get my drift
[07:34] <twb> Sorry, I forgot to mention this ITS also manages all the money stuff :-(
[07:35] <imbrandon> sure but he surely isnt the only one, if there isnt code reviews by other developers and its money on the line that just dumb, fire the ceo
[07:35] <twb> But yeah, double ledger accounting has been best practice since the 17th century, so I would expect anyone doing financial ANYTHING to know about it
[07:35] <imbrandon> NOW
[07:35] <twb> imbrandon: there's only one dev
[07:36] <imbrandon> thats ok for a blog, i thought you said this was a business that meney was at stake
[07:36] <twb> This code is kept around instead of migrating to RT because the CEO (who also owns the company) thinks it's awesome :-/
[07:36] <imbrandon> if you cant hirer the help outsource , thats just nuts, your ceo should be put in jail for being liable
[07:36] <imbrandon> as well as anyone who follows him
[07:37] <twb> AIUI it's not a public company so it's 100% legal to run it into the ground by mismanagement
[07:37] <imbrandon> yea let me give all my money to this 16 year old to go alone and make a midnight night deplosit
[07:37] <twb> Not that this has happened yet... it's a 25yo company
[07:37] <imbrandon> that sounds smart
[07:37] <twb> imbrandon: that's pretty much what happens, only he's twenty-five or so now
[07:38] <twb> He doesn't come into the office because we will yell at him
[07:38] <imbrandon> he is dumb for not quiting and the ceo needs shot
[07:38] <imbrandon> yell at him sounds like he is savin yalls ass
[07:38] <twb> Well he also took the ITS off and made his own company so now he's a contractor!
[07:39] <imbrandon> nice , 8x the monney is always better
[07:39] <twb> So we say "ITS should do this" and he goes "but that might break things for all my other (imaginary) customers"
[07:39] <twb> Of course he has probably joined this channel by now and is collecting incriminating evidence...
[07:39] <imbrandon> well it is only a request unless you are willing to walk
[07:39] <imbrandon> but it dont sound like it
[07:40] <twb> CEO isn't and he gets veto :-
[07:40] <twb> :-/
[07:40] <imbrandon> heh the cannel is loged
[07:40] <imbrandon> all #ubuntu cannels are public logged
[07:40] <twb> Yeah I don't really care
[07:40] <imbrandon> i know, was just letting ya know
[07:40] <imbrandon> :)
[07:41] <imbrandon> okies i got to get some real work done, you good on the newtelic or wanna poke some more ?
[07:42] <twb> Nah, I'm done.  Thanks.
[07:42] <imbrandon> kk, i'm round if ya need soemthing , might need to say my name so irssi hilights me
[07:42] <imbrandon> :)
[07:42] <twb> No worries
[07:43] <imbrandon> last thing though i have no affilatin with new relic btw, if you dont use them gram something similar
[07:43] <imbrandon> they are very trusted in the industry though
[07:43] <imbrandon> if you grab reviews
[07:43] <twb> home time for me too
[07:43] <imbrandon> l8tr
[07:56] <jamespage> morning all
[07:56] <bencer_> morning jamespage
[07:57] <jamespage> morning bencer_
[07:57] <jamespage> good weekend?
[07:57] <bencer_> seems that we made all: both packages and doc
[07:57] <bencer_> yup, a friend from frankfurt visited me, and the bbq in the park season has started
[07:57] <bencer_> according to the german standards :)
[07:58] <jamespage> +1
[07:58] <jamespage> sounds good
[07:58] <imbrandon> elllo ello
[07:58] <bencer_> i've one questions, are the bugs filled against the old ebox src packages automatically deleted or do we have to close/move them?
[07:58] <jamespage> bencer_, I was going to get something added to the release notes for precise as well
[07:59] <bencer_> that would be cool
[07:59] <jamespage> bencer_, the archive-admins will close them once the work has been completed
[07:59] <jamespage> bencer_, no further action required from our perspective
[07:59] <bencer_> just got removed now
[07:59] <bencer_> 2012-03-26 07:11:56 INFO    Removed-by: Martin Pitt
[07:59] <bencer_> 2012-03-26 07:11:56 INFO    Comment: superseded by zentyal-*, LP #957109
[07:59] <bencer_> 2012-03-26 07:11:57 INFO    15 packages successfully removed.
[07:59] <uvirtbot> Launchpad bug 957109 in ebox "Please remove libebox, ebox and ebox-* packages from the precise archive" [Undecided,Fix released] https://launchpad.net/bugs/957109
[08:00] <jamespage> bencer_, ah - so I see
[08:00] <bencer_> i was going to prepare and send a couple of blueprints we want to propose for q
=== bencer_ is now known as bencer
[08:01] <jamespage> bencer, sounds good
[08:02] <bencer> me and jacalvo we will be in the uds :)
[08:03] <jasonmsp> hey all.  What is a good spam-filter to use with 10.04 and a postfix/dovecot setup?
[08:08] <greppy> jasonmsp: I like amavisd-new + spamassassin + clamav with sanesecurity signatures.
[08:08] <greppy> jasonmsp: http://www200.pair.com/mecham/spam/ can be made to fit ubuntu without too much headache.
[08:14] <jasonmsp_> timed out...  Anyone have a suggestion for a spam filter?
[08:14] <jasonmsp_> postfix/dovecot on 10.04
[08:15] <jamespage> bencer, excellent! look forward to meeting you in person
[08:15] <bencer> jamespage: yes. do you know when are people starting to submit blueprints for q?
[08:16] <bencer> maybe its a bit too early?
[08:16] <jamespage> bencer: most people are still focussed on getting precise finished
[08:16] <jamespage> blueprints normally get raised/approved in the 2 weeks or so before UDS
[08:17] <jamespage> but if you have time now it won't hurt to raise them!
=== Lcawte|Away is now known as Lcawte
[08:17] <bencer> ok, i will
[08:17] <bencer> probably today, as i've already some drafts
[08:20] <greppy> jasonmsp_: that url timed out?  worked for me.
[08:27] <jasonmsp> greppy..  no my connection sucks and I keep dropping out.  Wanted to make sure I didn't miss a response.
[08:28] <jasonmsp> im looking for a spam filter for 10.04 dovecot/postfix setup.
[08:28] <greppy> jasonmsp: http://www200.pair.com/mecham/spam/
[08:28] <greppy> jasonmsp: that can be made to fit ubuntu 10.04 without too much pain.
[08:32] <jasonmsp> greppy: thx..  I prefer quality over complexity.  Do you think thats the best solution?
[08:32] <yolanda2> hi, good morning
[08:33] <yolanda2> any one can help with a dbconfig problem?
[08:33] <yolanda2> i always asks for password confirmation all the time
[08:38] <soren> yolanda2: Which package?
[08:39] <yolanda2> hi soren, i'm doing a new package for openerp, a new one that i'm building
[08:40] <soren> yolanda2: Ah.
[08:40] <soren> yolanda2: And it asks for password on every upgrade?
[08:42] <greppy> jasonmsp: it's a pretty comprehensive solution.  There isn't really a single "do this one thing and it will fix your spam issues" solution.
[08:42] <yolanda2> soren, every time i even try to install the package. Only ask for server, and password confirmation
[08:42] <greppy> jasonmsp: at least, not without paying someone else to do it for you :)
[08:42] <yolanda2> it's quite strange. It doesn't ask for a password, so password confirmation never works, of course, and blocks the install
[08:46] <soren> yolanda2: You could try enabling debconf debugging to see if that offers any hints.
[08:46] <yolanda2> hi soren, new with that, how can i do it?
[08:46] <soren> Set DEBCONF_DEBUG=developer before doing dpkg -i.
[08:46] <yolanda2> ok
[08:46] <yolanda2> thx
[08:46] <soren> Either like so:
[08:46] <soren> export DEBCONF_DEBUG=developer
[08:46] <soren> ..and then run dpkg.
[08:46] <soren> Or:
[08:47] <soren> (better)
[08:47] <soren> sudo DEBCONF_DEBUG=developer dpkg -i blah.deb
[08:47] <soren> It might only really be helpful if you know at least a bit about how debconf works.
[08:54] <yolanda2> soren, a little, but i'm doing the first package with it
[09:02] <lynxman> morning o/
[09:04] <jamespage> morning lynxman
[09:04] <lynxman> jamespage: good morning sir :)
[09:05] <jamespage> nice weekend?
[09:05] <lynxman> jamespage: it was indeed, first picnic of the season
[09:06] <lynxman> jamespage: and yours?
[09:06] <jamespage> lynxman, very nice thankyou! lots of time enjoying the nice weather for a change!
[09:07] <lynxman> jamespage: lovely :D
[09:08] <soren> yolanda2: Any luck?
[09:09] <soren> yolanda2: Otherwise, I recommend you share the code somewhere. Otherwise, it's quite hard to guess what's going on.
[09:10] <yolanda2> sorry, soren, meeting now
[09:11] <yolanda2> i have the code public, let me show you later
[09:14] <soren> yolanda2: no worries.
[09:30] <linocisco> ubuntu server sucks.
[09:34]  * soren bites
[09:34] <greppy> linocisco: you are free to use something else.
[09:34] <soren> linocisco: Are you going to qualify that somehow?
[09:35] <linocisco> soren, I want to setup loadbalancer like pfsense using ubuntu. but no easy way
[09:36] <soren> linocisco: I wonder if that could be phrased as a constructive question...
[09:37] <linocisco> soren, I love ubuntu alot. but whenever I asked on forums or IRC about seting up dual WAN router to failover+loadbalancing proxy server using two ISP links which has two different upstream proxy
[09:37] <linocisco> it is not possible. like impossible
[09:38] <soren> And that's easy as pie with this pfsense thing you speak of?
[09:39] <linocisco> soren, I have not finished my setup using pfsense. but it is more closer
[09:43] <soren> linocisco: Doesn't sound unachievable at all. haproxy + policy routing + something that monitors connectivity and tears down connections as needed and retries.
[09:43] <soren> linocisco: Just complex if you've never used either before.
[09:49] <linocisco> soren, yes. even ubuntu geeks said it is difficult
[09:49] <linocisco> soren,  there is no predefined how to guide
[09:50] <soren> linocisco: Predefined how to guides for every single scenario doesn't scale.
[09:50] <soren> Simple as that.
[09:51] <soren> You're just going to have ti learn the tools (or find someone else to do it all for you).
[09:52] <linocisco> soren, ti =?
[09:54] <soren> "to"
[09:54] <soren> You're just going to have to learn the tools (or find someone else to do it all for you).
[10:02] <yolanda2> hi soren, debug showed some problems and now it asks for password, i'm having some other problems now but i'm trying to solve them
[10:26] <yolanda2> hi soren, just another question. Now my package fails because i don't have postgresql and postgresql-client installed. If i use dbconfig, and choose postgres, isn't that dependency handled automatically? do i have to add a dependency for postgres anyway?
=== shift_ is now known as sectionme
[10:49] <soren> yolanda2: Yes.
[10:50] <soren> yolanda2: Not all packages work with all db types, so a dependency from dbconfig-common to all the various client tools would be wrong.
[10:50] <soren> yolanda2: Add a dependency on the relevant tools for the databases your package supports.
[10:51] <yolanda2> so dependency for postgres and postgresql is needed? and what about if you choose to don't config postgres? the tool gives you that option, postgres will be installed anyway?
[10:51] <Daviey> You can actually tell dbconfig what databases it should support.
[10:52] <Daviey> Unless you have a foo-database package, i'd make it a recommends... as i don't always want to run the database on my app server
[10:52] <Daviey> If you do make a foo-database package, make it a recommends of the app serverpackage
[10:52] <Daviey> I might have repeated what soren said, sorry.
[10:53]  * ogra_ prefers the bar-database packages ... they have a promis of beer in them 
[10:53] <yolanda2> Daviey, i have a recommend now
[10:53] <ogra_> *promise
[10:53] <yolanda2> but then after supplying the password, i have this error
[10:53] <yolanda2>  │ An error occurred while installing the database:                                                                                                        │
[10:53] <yolanda2>  │                                                                                                                                                         │
[10:53] <yolanda2>  │ psql: could not connect to server: No such file or directory Is the server running locally and accepting connections on Unix domain socket              │
[10:53] <yolanda2>  │ "/var/run/postgresql/.s.PGSQL.5432"?
[10:53] <yolanda2> as Postgres isn't installed
[10:54] <Daviey> yolanda2: postgres isn't running
[10:54] <yolanda2> Daviey, i just added as a recommends, so it shouldn't be installed by default, but i thought that dbconfig was handling that
[10:54] <Daviey> you probably need to check it's running, and if not, run it from postinst
[10:55] <Daviey> i think dbconfig SHOULD.. but it doesn't it seems
[10:55] <yolanda2> Daviey, can i show you my debian files? perhaps i'm doing something wrong
[10:56] <soren> yolanda2: There's a config option to tell dbconfig whether to ask for remote hsots by default.
[10:56] <soren> yolanda2: dbc_remote_questions_default or something.
[10:56] <yolanda2> dbc_remote_questions_default=true
[10:57] <soren> yolanda2: If that's not set to "true", it won't ask for remote hosts by default (unless you've lowered your debconf priority threhold)
[10:57] <soren> yolanda2: Where do you set that?
[10:57] <Daviey> yolanda2: I assume you are seeing the issue apt apt-get time, rather that throught the installer?
[10:58] <yolanda2> let me show you the files, just a moment
[11:00] <yolanda2> it's pushed here: sftp://yolanda@chinstrap.canonical.com//home/yolanda/openerp-package-test
[11:00] <yolanda2> it's my first contact with dbconfig, i followed documentation and some samples, but perhaps something is wrong
[11:01] <soren> It's been a while since I've had chinstrap access. :)
[11:01] <soren> Daviey: It's all yours :)
[11:01] <Daviey> yolanda2: fancy scp'ing to people.canonical.com:~/public_html/ ?  Then everyone can see it :)
[11:02] <yolanda2> ok
[11:09] <yolanda2> sorry, Daviey, i'm trying an scp to people.canonical.com and gives me access denied
[11:13] <soren> yolanda2: Perhaps scp isn't permitted?
[11:13] <soren> yolanda2: (It's different from sftp)
[11:13] <koolhead11> yolanda2: what are you installing via apt-get?
[11:14] <yolanda2> koolhead11, no, using a locally generated debian file
[11:14] <yolanda2> Daviey, i did a bzr push: bzr push sftp://people.canonical.com//home/yolanda/public_html/openerp-package-testing/
[11:14] <yolanda2> can you check it?
[11:16] <uvirtbot> New bug: #965138 in juju "Juju concatenates the full API endpoint URL when it gets absolute paths from the MAAS api." [Undecided,New] https://launchpad.net/bugs/965138
[11:16] <soren> yolanda2: Looking at your config.. It might be correct, but it's certainly different from how I usually do it.
[11:17] <yolanda2> soren, if you can give me some guidance is good, i'm new to that so perhaps i didn't look at the right samples
[11:17] <soren> yolanda2: I tend to source the various files, set the dbc_* variables and then dbc_go.
[11:18] <yolanda2> soren, i do more or less like that, what difference you see?
[11:18] <soren> yolanda2: I've admittedly never done dbtype-specific stuff.
[11:18] <soren> I've always had a generic package and set dbc_dbtypes
[11:19] <soren> yolanda2: You check for the existence of the pgsql file, then you set the dbc_* vars, then you source the config and run dbc_go
[11:19] <soren> Oh, hang on.
[11:19] <soren> You source the config and run dbc_go in a subshell.
[11:20] <yolanda2> soren, if you can show me some of your code it can be good
[11:20] <soren> The dbc_vars won't propagate to dbc properly.
[11:21] <Daviey> $ bzr branch http://people.canonical.com/~yolanda/openerp-package-testing/
[11:21] <soren> yolanda2: http://paste.ubuntu.com/900274/
[11:21] <soren> yolanda2: vs
[11:21] <soren> yolanda2: http://paste.ubuntu.com/900275/
[11:21] <soren> (first is yours, second is what I'd usually do.
[11:22] <soren> Getting rid of the subshell would possibly solve everything, though.
[11:22] <yolanda2> thanks soren, i'll try
[11:22] <soren> It certainly won't work without it.
[11:22] <soren> Mind you, I've only looked at the config.
[11:23] <soren> yolanda2: In your postinst, you're calling "getent passwrd | grep". Please don't do that.
[11:23] <Daviey> yolanda2: does this help, http://bazaar.launchpad.net/~maas-maintainers/maas/packaging/view/head:/debian/maas.config
[11:23] <Daviey> yolanda2: lazily generates a random password.
[11:23] <soren> yolanda2: If there's an ldap user backend with thousands of users, that's going to dreadful.
[11:23] <Daviey> that is a .config file
[11:23] <soren> yolanda2: Do "getent passwd <username>"
[11:24] <yolanda2> soren, ok, that sentence came by default in openerp initial package so i just left like it was, but i'm seeing that the initial package wasn't very good
[11:24] <soren> yolanda2: I also don't see the postinst actually consuming the info from dbconfig, but perhaps that's not done yet?
[11:25] <yolanda2> should be there
[11:25] <yolanda2> let me see,
[11:25]  * soren stares
[11:25] <soren> Nope, don't see it
[11:25] <soren> You're also a bit heavy on the calls to dbc_go.
[11:26] <yolanda2> seems that i pushed wrong version
[11:26] <soren> Three of them, (one being in a subshell)
[11:26] <Daviey> yolanda2: When you are ready with the package, i might have a tester lined up.
[11:27] <yolanda2> soren, Daviey, pushed right version to people.canonical.com now, sorry
[11:28] <soren> yolanda2: You don't need to set all the dbc_* vars in the postinst.
[11:29] <Daviey> yolanda2: Does openerp need a shell?
[11:29] <soren> I still don't see the postinst doing anything with the stuff that dbconfi provides.
[11:29] <yolanda2> perhaps something is wrong then, it's the right code now
[11:30] <yolanda2> that is in the postinst
[11:30] <yolanda2> # source dbconfig-common shell library, and call the hook function
[11:30] <yolanda2> if [ -f /usr/share/dbconfig-common/dpkg/postinst ]; then
[11:30] <yolanda2>     . /usr/share/dbconfig-common/dpkg/postinst
[11:30] <yolanda2>     dbc_dbtypes="pgsql"
[11:30] <yolanda2>     dbc_authmethod_user="password"
[11:30] <yolanda2>     dbc_dbuser="openerp"
[11:30] <yolanda2>     dbc_remote_questions_default=true
[11:30] <yolanda2>     dbc_go openerp $@
[11:30] <yolanda2> fi
[11:30] <yolanda2> isn't right?
[11:30] <soren> You don't need to set all those vars in the postinst.
[11:30] <soren> ...but as is, the postinst will just create the user.
[11:30] <soren> It won't shove the config into openerp.
[11:30] <soren> Perhaps it's not needed?
[11:31] <soren> I'm just so used to mysql where I always need to pass the password on to the app.
[11:31] <soren> ..but I guess with postgres that's not necessarily necessary.
[11:31] <soren> What with its ident auth and whatnot.
[11:31] <yolanda2> soren, well, i need to put the username and password in a openerp.conf file
[11:32] <yolanda2> can be done with md5 or with ident, we are trying md5
[11:32] <soren> yolanda2: Ok.
[11:32] <soren> yolanda2: That's what I mean then :)
[11:32] <soren> yolanda2: I don't see the code to do that, but if that's still not done, that's clearly why :)
[11:33] <yolanda2> soren, i still haven't done it, i'm in the process :)
[11:33] <soren> Ok, my servers are done deploying, I'll go and poke at them for a while.
[11:33]  * soren wanders off
[11:47] <Daviey> yolanda2: Are you looking at  a 6.1 snapshot, rather than 6.0 release?
[11:47] <yolanda2> Daviey, yes, the package is for 6.1
[11:50] <Daviey> yolanda2: when is 6.1 final due?
[11:50] <yolanda2> Daviey, openerp 6.1 is already out
[11:51] <Daviey> yolanda2: then why does it look like you are doing a snapshot?
[11:53] <yolanda2> Daviey, because they still delivery nightly builds from time to time, do you mean that?
[11:54] <Daviey> yolanda2: Do they have a 6.1 release that can be downloaded?
[11:55] <yolanda2> Daviey, yes
[11:55] <yolanda2> let me send you to the page
[11:55] <yolanda2> http://nightly.openerp.com/6.1/releases/openerp-6.1-1.tar.gz
[11:56] <Daviey> ah, cool - http://nightly.openerp.com/6.1/releases/openerp-6.1-1.tar.gz
[11:56] <Daviey> thanks
[12:08] <yolanda2> np
[12:11] <uvirtbot> New bug: #965188 in mcollective (universe) "Mcollective is launched more than once" [Medium,Invalid] https://launchpad.net/bugs/965188
=== smb` is now known as smb
[12:43] <smoser> zul, i just opened https://bugs.launchpad.net/nova/+bug/965225
[12:43] <uvirtbot> Launchpad bug 965225 in nova "EC2 TerminateInstances api call hangs and returns 500 when given bad data (ip address)" [Undecided,New]
[12:43] <zul> smoser:  cool thanks (i think ;)
[12:44] <zul> smoser: you can use an ip address to terminate an instance?
[12:49] <smoser> no
[12:49] <smoser> its just bogus data.
[12:52] <zul> smoser: ah i see
=== bladernr_afk is now known as bladernr_
[13:13] <zul> Daviey:  can you review please? http://paste.ubuntu.com/900421/
[13:16] <Daviey> zul: looking
[13:17] <Daviey> zul: how is keystone looking for MIR btw?
[13:17] <zul> Daviey: good...we have to fix up the swift testsuite and then it should be ok
[13:18] <Daviey> fix-ubuntu-tests.patch i'd call it, tests-disable-external-deps.patch .. but just a taste thing.
[13:19] <Daviey> --#log_config = /etc/keystone/logging.conf
[13:19] <Daviey> +-#log_config = ./etc/logging.conf.sample
[13:19] <Daviey> ?
[13:19] <Daviey> patching a commented field?
[13:20] <Daviey> -	python setup.py build_sphinx || true
[13:20] <Daviey> +	python setup.py build_sphinx
[13:20] <Daviey> Does it currently fail to generate doc's?
[13:22] <Daviey> Seems an odd fix for, * debian/rules: Fix doc builds + clean (LP: #956019)
[13:23] <zul> Daviey: not sure adam_g made those modifications for the log_config stuff
[13:23] <zul> smoser: fixed locally
=== Lcawte is now known as Lcawte|Away
[13:36] <koolhead11> so finally i can see step by step Essex guide for precise http://www.hastexo.com/resources/docs/installing-openstack-essex-4-ubuntu-1204-precise-pangolin
[13:50] <yolanda2> hi Daviey, i got some success in creating openerp user now. However, seems that is also creating a database, is there any way to create only the openerp user, but without any database? this is going to be created by openerp interface
[13:56] <uvirtbot> New bug: #965283 in openvswitch (universe) "include openvswitch documentation / package openvswitch-docs" [Undecided,New] https://launchpad.net/bugs/965283
[14:04] <ttx> zul: looks like a packaging todo list: http://www.hastexo.com/resources/docs/installing-openstack-essex-4-ubuntu-1204-precise-pangolin
[14:07] <zul> ttx: cool ill have to read it
[14:07] <ttx> zul: there are a few weird things, but overrall, it's not bad
[14:08] <zul> ttx:yeah
[14:08] <koolhead11> ttx: :P
[14:08] <ttx> koolhead11: weird things include copying an old volume.py file to solve a bug, losing all other improvements to the file
[14:09] <koolhead11> ttx: Madkiss is the guy who wrote it
[14:09] <koolhead11> he is in #openstack channel
[14:09] <ttx> ack
[14:17] <raydog45000> 	Hello all, I am trying to set up an ubuntu server, I checked the disc for defects and tested memory, after I choose a host name it fails at 43% of starting up the partition, how can I get it to complete?
[14:18] <raydog45000> anyone here?
[14:19] <greppy> raydog45000: nope, we're all just idling.
[14:19] <greppy> if someone can help, they will :)
[14:19] <raydog45000> m'k
[14:33] <smoser> zul, did you run test on that change?
[14:33] <zul> smoser: yeah
[14:33] <smoser> i'm just looking at the tests in ova/tests/api/ec2/test_ec2_validate.py, and would have thought you'd break one.
[14:33] <smoser> maybe you should add one there.
[14:33] <zul> smoser: ack
[14:34] <smoser> and i commented in the review a bout a little thing to fix while you're there.
[14:38] <smoser> hallyn, http://www.mail-archive.com/openstack@lists.launchpad.net/msg09018.html
[14:38] <smoser> i'm not the only one.
[14:39] <hallyn> without opening the link, i think i saw the email
[14:39] <smoser> althoguht his post (and my recent test) does not show regression in performance due to guest (lucid versus precise), which is what i originally thought.
[14:39] <smoser> hallyn, yeah, you can't trust me enough to open links
[14:40] <hallyn> i'm afraid high prio bugs will continue to take priority over the performance tests though
[14:40] <hallyn> at leaset, the ones in precise
[14:40] <hallyn> still,
[14:40] <hallyn> most of your perf loss was still due to https right?
[14:41] <smoser> did you do this ?
[14:41] <smoser> http://s3hh.wordpress.com/2012/01/11/kvm-performance-measurements-my-plan/
[14:41] <hallyn> heck whatever is causing that may also be responsible for the other perf loss
[14:41] <hallyn> smoser: no, that's what i'm saying i'll do as soon as high prio precise bugs are all cleared or waiting on something
[14:41] <smoser> by "do this", i mean, the blog says "I intend to do". did you do ?
[14:41] <hallyn> no.  i continue to intend to do this
[14:42] <smoser> hallyn, well, on e thing not listed there, is a regression test.
[14:42] <smoser> its fine/good to turn all the knobs on precise
[14:42] <hallyn> right
[14:42] <smoser> but if we got 100M/s in lucid, and 20M/s in precise, we need to know that.
[14:43] <hallyn> right, so i'm going to first do a basic test (with precisely one configuration) to compare oneiric to precise
[14:43] <hallyn> you think i should do lucid instead of oneiric?
=== DLobanov_ is now known as DLobanov
[14:51] <Adri2000> zul: are there plans to merge openstack packages with debian's? if yes, before or after precise?
[14:51] <uvirtbot> New bug: #965356 in nova (main) "unnecessary dep: nova-api -> nova-cert" [Undecided,New] https://launchpad.net/bugs/965356
[14:51] <zul> Adri2000: the process has already started it should be completed after precise
[14:55] <Maleko> can someone explain why ubuntu automatically re-add/set the highlighted route as def gw
[14:56] <Maleko> http://pastebin.com/imadQJuB
[14:56] <addictedboy> Trying to set firefox custom server using their guide, but getting this error message while restarting
[14:57] <addictedboy> the apache
[14:57] <addictedboy> apache2: bad group name sync
[14:57] <addictedboy> Action 'configtest' failed.
[14:57] <addictedboy> The Apache error log may have more information.
[14:57] <addictedboy>    ...fail!
[14:57] <Adri2000> zul: ok. another question: do you or anyone knows if ~ubuntu-server-dev is open to non-coredevs motus? or: can I directly contribute to openstack packaging without applying for coredev?
[14:58] <zul> Adri2000: yes you dont have to ubuntu-server-dev/core-dev to contribute to openstack packaging you just have to do merge proposals
[14:59] <Adri2000> I know about merge proposals, that's why I said "directly contribute" :)
[15:02] <zul> Adri2000: right we accept merge proposals but you need either ubuntu-server-dev or core-dev to merge directly into the branches
[15:04] <Adri2000> ok
[15:05] <SpamapS> Adri2000: ubuntu-server-dev is approved via the developer membership board
[15:06] <SpamapS> Adri2000: it covers quite a few things outside openstack, and thus, one needs to show a wider breadth of contribution to be granted those rights.
[15:06] <Adri2000> I see
[15:07] <SpamapS> Adri2000: we'd b *more* than happy to help you achieve it. Another option is to just apply for per package upload rights for the packages you are most interested in.
[15:08] <SpamapS> Adri2000: for that you'd just need to show a sustained contribution to the packaging of those things over about the length of one release cycle (so 6 months)
[15:08] <SpamapS> Adri2000: in order to show that.. merge proposals. :)
[15:10] <Adri2000> actually I'm not looking specifically for more upload rights, but rather a convenient (i.e. commit+push :)) way to contribute some fixes to the packaging. just like it's possible in some debian teams, where their git/whatever repo is open to a larger team than just the team of debian developers
[15:11] <uvirtbot> New bug: #237164 in xorg "kvm needs to correctly simulate a proper monitor" [Medium,Invalid] https://launchpad.net/bugs/237164
=== jvdz_ is now known as jvdz
[15:15] <csotelo_> hello people I have a problem on creating a user on postgres after install it
[15:15] <csotelo_> createuser: could not connect to database postgres: could not connect to server: No such file or directory
[15:15] <csotelo_>     Is the server running locally and accepting
[15:15] <csotelo_>     connections on Unix domain socket "/tmp/.s.PGSQL.5432"?
[15:17] <KM0201> csotelo_: i have no idea on that.. patdk-wk may know, as i recall, he is pretty knowledgeable.
[15:18] <patdk-wk> hmm?
[15:19] <Adri2000> zul: may I suggest to rename /etc/apache2/conf.d/dashboard.conf in horizon to something more explicit? (like openstack-dashboard.conf) - tell me if you want a bug report
[15:19] <zul> Adri2000: bug report with a merge proposal would be good :)
[15:20] <Adri2000> ehhhh - that's such a small change that I feel a merge proposal is more work for everyone...
[15:20]  * Adri2000 files a bug at least
[15:23] <csotelo_> thanks KM0201
[15:23] <csotelo_> patdk-wk, please.. have you read my problem
[15:23] <KM0201> csotelo_: he may not have saw it, as i think he came in after you posted it
[15:23]  * patdk-wk sees no question
[15:23] <patdk-wk> had to reboot for kernel update
[15:27] <csotelo_> patdk-wk, my quetsion is why postgres dont permit me to create a new user
[15:28] <csotelo_> because when I do a createuser -SRDp myuser as postgres user, then I get the error
[15:28] <patdk-wk> dunno about postgres
[15:28] <patdk-wk> most likely cause you didn't authenicate to postgres as an admin
[15:28] <csotelo_> ?
[15:28] <csotelo_> patdk-wk, why I have done is install a postgres on a new machine
[15:30] <csotelo_> I usually done something like sudo apt-get install postgresql postgresql-client
[15:30] <csotelo_> and the
[15:30] <csotelo_> sudo -s and then su postgres
[15:30] <csotelo_> and filelly createuser -SDRp myuser
[15:30] <csotelo_> finally*
[15:30] <csotelo_> and normally it work
[15:30] <csotelo_> I don the same and get that error on a new machinne
[15:32] <SpamapS> Adri2000: so bzr packaging branches are open to a wider team.. commit, push, and one more command 'bzr lp-propose'
[15:36] <koolhead11> !postgres
[15:36] <koolhead11> !pgsql
[15:37] <koolhead11> csotelo_: https://help.ubuntu.com/10.04/serverguide/C/postgresql.html  see if this helps
[15:39] <caribou> hallyn: Regarding bug 965231, would a copy of the VM I'm using for testing help ?
[15:39] <uvirtbot> Launchpad bug 965231 in qemu-kvm "Virtual Machine is not terminated when doing "shutdown" in VM with Oneiric/Precise" [High,Confirmed] https://launchpad.net/bugs/965231
[15:39] <Adri2000> zul: ok, I must admit that it's a bit more complicated: I forgot the handling of the rename in maint scripts :), so well, I'll go ahead and propose a branch
[15:40] <Adri2000> SpamapS: right, I'll give it a try ^
[15:40] <zul> cool
[15:40] <hallyn> caribou: itmight...  i know shutdown was working ofr me last week, but i can't recall if that was with the precise package, or with upstream
[15:40] <hallyn> so i just need to get my test box up and test.
[15:40] <caribou> hallyn: it's a very basic Oneiric-server install nothing particular
[15:40] <koolhead11> zul: is diablo/essex and precise/essex updated simultaneously ?
[15:41] <hallyn> caribou: but is acpid installed?
[15:41] <zul> koolhead11:  diablo/essex?
[15:41] <caribou> hallyn: now it is & it still hangs after "System halted"
[15:41] <SpamapS> Adri2000: renames are easier these days with dpkg-maintscript-helper (though that makes things less backportable since it is fairly new)
[15:41] <koolhead11> zul: oops. i meant oneiric/essex srry
[15:41] <caribou> hallyn: I must be missing something obvious otherwise it would have been noted before
[15:41] <hallyn> ok.  sounds like a definate bug :)
[15:42] <hallyn> no no, it may be broken in precise and fixed upstream.  let me test (sometime today)
[15:42] <caribou> hallyn: why would I be the only one seeing this, it started on Oneiric
[15:42] <caribou> hallyn: no rush, it' end of day here in France
[15:42] <hallyn> caribou: I don't know that many people use 'virsh shutdown'
[15:43] <caribou> ah ok
[15:43] <caribou> hallyn: thanks for looking at this, see you tomorrow
[15:43] <hallyn> caribou: np, ttyl
=== jvdz_ is now known as jvdz
[15:45] <zul> koolhead11: no it isnt
[15:46] <koolhead11> zul: okey. i thought so :)
[15:48] <Adri2000> SpamapS: yep, found http://raphaelhertzog.com/2010/10/14/correctly-renaming-a-conffile-in-debian-package-maintainer-scripts/ ; it says it works since 1.15.7.2 which is there since maverick, so that will work I think :)
[15:53] <jjohansen1> hallyn, stgraber: so we have a problem, turns out there was a kernel bug in addition to the userspace issue.  We need both to fix the issue.  This will require a kernel freeze exception from the release team
[15:54] <hallyn> jjohansen1: thanks for the info.  stgraber: soudns like we should go with the temp-proc-mount fix then
[15:54] <hallyn> jjohansen1: quick related q,
[15:54] <jjohansen1> hallyn: shoot
[15:54] <hallyn> jjohansen1: is unconfined allwoed to request transition to all domains?
[15:54] <jjohansen1> hallyn: yes
[15:54] <hallyn> ok
[15:54] <hallyn> right now lxc-start enters the usr.bin.lxc-start comain, but lxc-execute doesn't
[15:54] <hallyn> but both request entring the container's profile
[15:55] <hallyn> so i think that's what we want right
[15:55] <hallyn> thanks
[15:55] <jjohansen1> sounds right
[15:57] <stgraber> hallyn: yeah, the temporary fix sounds good then, feel free to upload and I'll poke #ubuntu-release to have it accepted
=== jcastro_ is now known as jcastro
[16:02] <hallyn> stgraber: lxc_0.7.5-3ubuntu45_source.changes dput'ed
[16:06] <stgraber> hallyn: thanks
[16:07] <hallyn> stgraber: jjohansen: I'm goign to commit some changes to the server guide lxc section to fill in the apparmor bits.  I'll post you to the merge request if you care to review.  (Though it shouldn't be contraversial)
[16:25] <kantlivelong> my grubmenu seems to be missing... how can i rebuild the grub conf? im on ubuntu 11.10 server x64
[16:26] <RoyK> update-grub ?
[16:27] <kantlivelong> oh lol :P
[16:27] <stgraber> hallyn: lxc accepted
[16:28] <hallyn> stgraber: cool.
[16:28] <hallyn> I'm going to do a merge proposal now for https://code.launchpad.net/~serge-hallyn/serverguide/lxc-aa/
[16:28] <hallyn> (and then consider myself done with lxc server guide for real)
[16:35] <zul> adam_g: https://launchpadlibrarian.net/98406449/buildlog_ubuntu-precise-i386.swift_1.4.7-0ubuntu3.2_BUILDING.txt.gz
[16:46] <adam_g> zul: cool
[16:51] <uvirtbot> New bug: #965476 in lxc (universe) "Can't install Precise guest on Oneiric host" [Undecided,New] https://launchpad.net/bugs/965476
[16:51] <uvirtbot> New bug: #965478 in image-store-proxy (universe) "Depends: eucalyptus-common but it is not installable" [Undecided,New] https://launchpad.net/bugs/965478
[16:52] <adam_g> zul: is that in the archive or a PPA somewhere?
[16:52] <zul> ppa
=== Arc_ is now known as a5m0
[17:18] <adam_g> zul: is keystone rc1 going to go in today?
[17:18] <zul> adam_g: i wanted to get daviey to review but i think he had a question about the logging.conf
[17:18] <zul> Daviey: ^^^
[17:19] <Daviey> no, the real question was about the docs.. the 'fix' is || true.. which seems to be a regression?
[17:19] <adam_g> huh?
[17:19] <Daviey> the logging.conf was more of a question, as why we are patching a commented field.
[17:20] <Daviey> 14:20 < Daviey> -Ipython setup.py build_sphinx || true
[17:20] <Daviey> 14:20 < Daviey> +Ipython setup.py build_sphinx
[17:20] <Daviey> oh, wait - i'm looking inverted
[17:20] <Daviey> *doh*
[17:20] <adam_g>  bzr+ssh://bazaar.launchpad.net/~openstack-ubuntu-testing/keystone/precise-essex-proposed/
[17:20] <adam_g> ^ i fixed the doc issue in that branch, which should be merged with the upload
[17:20] <uvirtbot> adam_g: Error: "i" is not a valid command.
[17:21] <zul> adam_g: yeah i got that locally
[17:21] <adam_g> zul: what do you mean
[17:21] <Daviey> adam_g: why are we doing:
[17:21] <Daviey> 14:19 < Daviey> --#log_config = /etc/keystone/logging.conf
[17:21] <Daviey> 14:19 < Daviey> +-#log_config = ./etc/logging.conf.sample
[17:22] <adam_g> -#log_config = ./etc/logging.conf.sample
[17:22] <adam_g> +log_config = /etc/keystone/logging.conf
[17:22] <adam_g> ...in -proposed
[17:23] <adam_g> http://paste.ubuntu.com/900800/
[17:24] <zul> oops...hold on
[17:24] <adam_g> i spent a while getting that package in better shape last week in preparation for rc1 landing
[17:25] <adam_g> Daviey: which branch were you looking at?
[17:25] <Daviey> adam_g: from zul's debdiff
[17:26] <adam_g> zul: is that debdiff a merge of -proposed into ubuntu-server-dev, or some other??
[17:26] <zul> bad merge
[17:26] <adam_g> hm, ok
[17:26] <uvirtbot> New bug: #965507 in juju "Juju uses http to contact uec-images.ubuntu.com" [High,In progress] https://launchpad.net/bugs/965507
[17:28] <zul> adam_g: the local branch here has the doc build changes
[17:28] <adam_g> zul: the local branch where?
[17:28] <zul> adam_g: on my pc
[17:29] <Daviey> adam_g: branch it from zul's PC, and we are gold!
[17:29] <adam_g> zul: why isnt it in -proposed? i wouldn't have spent time fixing it if were, and now we'd have a clean merge :)
[17:29] <zul> adam_g: it has the testsuite fixes that i was working on lemme, push back
[17:30] <smoser> zul,  fyi, you had smokestack pep8 errors in your https://review.openstack.org/#change,5809,patchset=1
[17:31] <zul> smoser: son of..
[17:32] <zul> adam_g: https://code.launchpad.net/~openstack-ubuntu-testing/keystone/precise-essex-proposed
[17:32] <zul> ^^^ that is what is going to be uploaded today
[17:32] <uvirtbot> zul: Error: "^^" is not a valid command.
[17:32] <zul> shut up uvirtbot
[17:33] <zul> adam_g: s/python-swift/swift/g though
[17:35] <adam_g> zul: ok
[17:36] <adam_g> so thats a different branch than the one Daviey was looking at?
[17:36] <zul> adam_g: he was looking at a debdiff
[17:39] <adam_g> zul: okay, well, that branch LGTM then. curious to know why you need the entire swift package as a builddepends, tho
[17:39] <zul> adam_g: yeah that should be just python-swift
[17:40] <zul> adam_g: so im going to change that to python-swift and then upload it
[17:42] <adam_g> zul: before you do, push change to -proposed and lets run it thru the CI build job once
[17:42] <zul> adam_g: k
[17:42] <zul> done
[17:42] <adam_g> zul: cool
[17:43] <zul> adam_g: swift is still failing in the ci though, 1.4.9 needs some dependencies for the testsuite
[17:43] <adam_g> zul: also the quantum package is still uninstallable, should i try to fix it up?
[17:43] <Adri2000> zul: I don't know if you get notified about this: https://code.launchpad.net/~adri2000/ubuntu/precise/horizon/rename-apache-config-file/+merge/99366
[17:44] <zul> adam_g: should be fixed i think
[17:44] <zul> Adri2000: i might have been i just havent gotten around to it yet
[17:44] <Adri2000> ok no problem
[17:46] <zul> adam_g: if you want to kick one off go ahead
[17:49] <adam_g> zul: about quantum http://paste.ubuntu.com/900840/
[17:50] <zul> adam_g: grr...i thought i fixed that part can you have a look
[17:50] <adam_g> yeah
[17:50] <adam_g> also strange that 'quantum-server' pulls in no other dependencies
[17:51] <lynxman> zul, adam_g: by any chance do any of you guys remember how to recover the admin credentials out of a cobbler setup?
[17:51] <zul> lynxman: re-install cobbler?
[17:51] <adam_g> id like to get quantum installable and try to solicit (beg for) some testing from the openstack list
[17:51] <lynxman> zul: it's a shared setup :/
[17:52] <adam_g> lynxman: i thought they were stored in a config file in plain text, but i could be wwrong. i know i had to reset it once, but forget how i did exactly
[17:52] <adam_g> roaksoax: ^
[17:53] <lynxman> adam_g: hmm thanks :)
[17:55] <lynxman> adam_g: found the file and resetted, thanks
[17:55] <lynxman> adam_g: it's /etc/cobbler/users.digest
[17:55] <adam_g> np
[17:55] <adam_g> zul: that keystone branch looks good-to-go
[17:59] <zul> adam_g: uploaded now just need to do keystone
[17:59] <zul> er...keysoneclient
[18:01] <uvirtbot> New bug: #965523 in mysql-dfsg-5.0 (universe) "mysql 5.5.22, 5.1.62, 5.0.96 security update tracking bug" [Undecided,Confirmed] https://launchpad.net/bugs/965523
[18:06] <zul> Daviey/adam_g: keystone/python-keystoneclient is in the queue
[18:07] <Daviey> zul: i see it, waiting for the diff
[18:07] <zul> Daviey: k
=== koolhead17|away is now known as koolhead17|dead
[18:09] <adam_g> zul: whats the story with the quantum packaging branch? i was trying to figure out the deal when you were on vacation but couldn't :)
[18:09] <zul> adam_g: should be up to date...ill take a look this afternoon
[18:09] <adam_g> zul: the ubuntu branch is not related to the ubuntu-server-dev or openstack-ubuntu-testing branch
[18:10] <zul> adam_g: yeah lemme poke it
[18:20] <roaksoax> lynxman: dpkg-reconfigure -plow cobbler
[18:30] <zul> adam_g: i should have the console tick stuff ready tomorrow
[18:31] <adam_g> zul: party time
[18:31] <zul> adam_g: i got rid of the tick interval though so it will check everytick count
[18:32] <adam_g> zul: https://bugs.launchpad.net/bugs/956019  whats the correct way to fix this? it needs a new binary package, but dont want to repeat last week
[18:32] <uvirtbot> Launchpad bug 956019 in keystone "keystone doesn't install the copyright file" [High,New]
[18:32] <adam_g> ooops
[18:32] <adam_g> https://bugs.launchpad.net/bugs/965356
[18:32] <uvirtbot> Launchpad bug 965356 in nova "unnecessary dep: nova-api -> nova-cert" [Undecided,Confirmed]
[18:32] <adam_g> oh, nvm
[18:32] <adam_g> nothing new required there
[18:33] <zul> yeah you got me confused for a minute
[18:34] <lynxman> roaksoax: thanks, didn't want to do that since its a shared orchestra server :)
[18:35] <roaksoax> lynxman: :)
=== YankDownUnder_ is now known as YankDownUnder
=== shadeslayer_ is now known as shadeslayer
[19:41] <ha1dfo> hi all. I'm trying to install grub2 to a highly customized ubuntu based linux running on ramdisk, but it grub-probe says: "cannot find a device for /" What can I do to make it work?
[20:02] <kantlivelong> is there a proper place to set readahead values?
[20:05] <smoser> utlemming, could you go through SRU for bug 948461
[20:05] <uvirtbot> Launchpad bug 948461 in apt "apt-get hashsum/size mismatch because s3 mirrors don't support http pipelining correctly" [High,Confirmed] https://launchpad.net/bugs/948461
=== michaelh1|away is now known as michaelh1
[20:06] <smoser> i'd say it would be sufficient to : boot instance , ssh instance, enable proposed, apt-get update, apt-get intsall cloud-init, show apt-setting was modified.
[20:06] <utlemming> smoser: I've already verified that one friday...using the same tests
[20:06] <utlemming> smoser: see comment #30
[20:07] <WeissLehrer> i have a problem on my samba configuration, for some reason the user-configuration for windows users using pdc are not saved ( http://pastebin.com/gXvvjC5u )
[20:07] <smoser> you need to change 'verification-needed' to 'verification-done'
[20:09] <smoser> utlemming, ^
[20:09] <utlemming> smoser: done
[20:15] <uvirtbot> New bug: #965654 in openssh (main) "/var/log/auth spam every 2 mins: sshd[*]: Connection closed by 127.0.0.1 [preauth]" [Undecided,New] https://launchpad.net/bugs/965654
[20:26] <uvirtbot> New bug: #965663 in openssh (main) "ssh-copy-id doesn't call restorecon on SELinux enabled destination hosts" [Undecided,New] https://launchpad.net/bugs/965663
[20:27]  * cwillu_at_work stabs an anonymous ubuntu developer at random
[20:28] <guntbert> ouch, you got me instead!
[20:30] <cwillu_at_work> guntbert, that means you need to update btrfs-progs in the repo :p
[20:32] <cwillu_at_work> !info btrfs-tools precise
[20:32] <ubottu> btrfs-tools (source: btrfs-tools): Checksumming Copy on Write Filesystem utilities. In component main, is optional. Version 0.19+20100601-3ubuntu3 (precise), package size 794 kB, installed size 1724 kB
[20:32] <cwillu_at_work> that's only about 2 years out of date :p
[20:32] <guntbert> cwillu_at_work: will do, ... uuhmm          what was the key again?
[20:32] <Nafallo> it's been the same since forever...
[20:32]  * guntbert stops the bantering immediately
[20:32] <Nafallo> like, several releases.
[20:33] <cwillu_at_work> Nafallo, I know
[20:33] <Nafallo> I had luck with the version in debian sid though.
[20:33] <cwillu_at_work> I just (light-heartedly) wish that they wouldn't say they support btrfs when the support is ancient, broken and deprecated by upstream :p
[20:33] <Nafallo> that's at least from last autumn, and includes scrub
[20:35] <cwillu_at_work> restripe support in 3.3 is nice
[20:40] <Aison> I suddenly have got big problems with setting up vlans on bonding network devices
[20:40] <Nafallo> cwillu_at_work: any news on working btrfsck?
[20:40] <Aison> all my machines with bonding and vlan are no longer reachable after reboot
[20:40] <cwillu_at_work> dangerdonteveruse branch
[20:40] <cwillu_at_work> fsck has been around and working fine for years
[20:40] <Aison> this setup worked for a long time now
[20:40] <cwillu_at_work> it's a repair utility that's been missing
[20:41] <Nafallo> cwillu_at_work: hmm. yeah. I don't really count an fsck as working without the ability to actually fix problems ;-)
[20:41] <cwillu_at_work> Nafallo, that's not what fsck is for
[20:42] <Nafallo> all the other fscks seems to do that stuff. just saying.
[20:42] <cwillu_at_work> no, they don't :p
[20:42] <cwillu_at_work> some of them do
[20:43] <Nafallo> if you say so. the ones I've used have been able to fix things, except for btrfs :-)
[20:43] <Nafallo> and btrfs is planning to implement that support, so...
[20:43] <cwillu_at_work> http://linux.die.net/man/8/fsck.xfs
[20:43] <Nafallo> *shrugs*
[20:43] <cwillu_at_work> "fsck.xfs - do nothing, successfully"
[20:44] <Nafallo> anyway. I don't care. I want something that can fix broken btrfs. whatever that is called ;-)
[20:44] <cwillu_at_work> btrfs can (mostly) fix broken btrfs :p
[20:45] <Nafallo> that "mostly" in there isn't convincing me :-)
[20:45] <cwillu_at_work> (coincidentally enough, btrfsck can fix it to about the same extent :p)
[20:45] <Nafallo> haha
[20:45] <cwillu_at_work> Nafallo, fsck.ext3 is just a mostly too you know
[20:46] <Nafallo> ehrm
[20:46] <cwillu_at_work> you may be confusing e2fsck's journal playback with its repair mode
[20:46] <Nafallo> all btrfsck have done for me is to tell me there are problems. and telling me that have taken it forever, to the point where I've told fstab to not use it ;-)
[20:46] <cwillu_at_work> Nafallo, it never should have been linked to fsck.btrfs
[20:47] <cwillu_at_work> (xfs ships a no-op fsck.xfs for precisely the same reason)
[20:48] <Nafallo> bah. I better go do the things I was supposed to do today before the day is completely gone.
[20:48] <Nafallo> later
[20:49] <Nafallo> well, I'll save implementing a spam-filter for tomorrow I think. I let thunderbird do it's thing now and tomorrow ;-)
[20:49] <Nafallo> ...morning
=== Arc_ is now known as a5m0
[21:29] <Jammie> Hi, I'm having trouble with eth0 not reconnecting after reboot (same as the problem here http://askubuntu.com/questions/73431/when-ubuntu-server-restarts-eth0-doesnt-come-back-up)
=== adam_g_ is now known as adam_g
[21:38] <Jammie> anyone have any ideas/suggestions?
[21:41] <ha1dfo> Jammie, after reboot, ifconfig -a does show anything?
[21:41] <ha1dfo> before config (if you have the case when udev messes up ifaces)
[21:42] <Jammie> no, on first reboot ifconfig -a is blank
[21:42] <Jammie> but after running sudo /etc/init.d/networking start it all worls
[21:42] <Jammie> *works
[21:47] <a5m0> how can i uninstall dhcpd on ubuntu server 12.04 i keep getting "
[21:47] <a5m0> Virtual packages like 'dhcp' can't be removed
[21:47] <a5m0> "
=== jvdz_ is now known as jvdz
[22:11] <uvirtbot> New bug: #952397 in juju "Juju is forcing *ALL* charms in a repo to be perfect to be able to deploy *any*" [Critical,Fix released] https://launchpad.net/bugs/952397
[22:21] <Aison> what's bad with this network configuration?
[22:21] <Aison> http://pastebin.com/y9Az7uDG
[22:22] <Aison> after reboot, I have to ifconfig eth0 up
[22:22] <Aison> ifconfig eth1 up
[22:22] <Aison> ifenslave bond0 eth0
[22:22] <Aison> ifenslave bond0 eth1
[22:22] <Aison> vconfig add bond0 10
[22:22] <Aison> etc....
[22:22] <Aison> myself
[22:27] <JanC> Aison: "manual"?
[22:28] <Aison> yes, this configuration worked for over two years now
[22:29] <Aison> but after some update, it no longer works
[22:33] <Jammie> did you upgrade to 11.10?
[22:35] <Aison> it's 11.10, yes
[22:35] <Aison> but the upgrade is some time back
[22:35] <Aison> it did not start with 11.10
[22:35] <Aison> the problems did not start with the upgrade to 11.10
[22:36] <Aison> they started for about 3 weeks ago
[22:46] <Jammie> ah, I thought it may be the same problem I'm having
=== bladernr_ is now known as bladernr_afk
[22:51] <theq_> Hello, some services can't connect using localhost or 127.0.0.1 or 127.0.0.2 address, I have fixed issues in mysql and apache by setting IP in config, but it doesn't work in postfix, courier, ejabberd, bind - I just can't connect to these from that pc which we are talking about, from any pc from network which see this PC I can connect. Where is the issue?
[22:51] <theq_> it comes after upgrade to 11.10
[22:51] <theq_> before it was working fine
[22:57] <locuse> hi.  i'm modifying my remote status-check scripts to use upstart rather than chkconfig.  "service --status-all" provides no run-level info as "chkconfig --list --all" did.  i *do* see chkconfig pkg is available for Ubuntu.  is it still used/useful for the runlevel info? or is chkconfig considered deprecated, and there's an equivalent, more-informative "service ..." cmd?
[22:59] <Daviey> locuse: have a sample output of what you were using before, chkconfig?
[23:01] <locuse> Daviey: hi.  sure -> http://pastebin.com/d5L97JJM.
[23:05] <locuse> Daviey: looking at the upstart output, iiuc, it shows the *actual* running status of each service ... i'm simply looking for the per-runlevel toggle status.  i.e., what each service is set to do @ startup
[23:08] <Daviey> locuse: Yes, i see the difference.. nothing at the top of my head, but the person to ask is jodh in 9 hours from now.  Sorry.
[23:08] <locuse> Daviey: heh, long after after my bedtime. np :-)
[23:09] <locuse> Daviey: upstart in an ubuntu constrcut, not JUST for ubuntu-server, right?
[23:09] <locuse> oops. s/in an/is an/
[23:09] <Daviey> locuse: right.
[23:10] <locuse> ok, i'll check in "the big room" as well.
[23:17] <locuse> Daviey: just fyi, stumbled on : http://www.linuxpowerup.com/en/Chkconfig-is-good-but-for-Ubuntu-sysv-rc-conf-is-a-better-alternative-for-enable-and-disable-services-at-startup