[00:34] <compubomb> i have natty server installed on an aws instance, i noticed that mod_fastcgi is no longer there, anyone have any idea how to get it back?
[00:35] <twb> compubomb: might be called fcgi?
[00:35] <compubomb> twb: they are not the same..
[00:35] <twb> Oh, OK
[00:35] <smw> compubomb, install libapache2-mod-fastcgi
[00:35] <compubomb> my problem is i only know how to setup php-fpm with apache using mod_fastcgi not fcgi
[00:35] <twb> Well, there's no mod_fastcgi in Debian at this time, at any rate.
[00:36] <smw> There is always libapache2-mod-fcgid :-\
[00:36] <twb> smw: yeah, that one
[00:36] <twb> I think when I was forced to deploy PHP, on a lucid box, that was the only cgi-based interface that was available
[00:36] <compubomb> wy did they remove it?
[00:37] <twb> compubomb: I would guess because it was unmaintained; the changelogs probably say
[00:37] <compubomb> how do you configure php-fpm using fcgid
[00:37] <twb> I don't know, ask #php
[00:38] <twb> IMO if you are using PHP, you have lost.
[00:38] <compubomb> you are entitled to your opinion.
[00:40] <twb> I don't imagine it's any harder to point mod-fcgid at php5-fpm than it is to point mod-fastcgi at php5-fpm...
[01:35] <jetole> Hey guys. I added a new disk to one of our mail servers today for storing users mail and I am about to format it however I am looking to add some inode configuration to allow many many many mail files. Can anyone help me with how I should configure the filesystem for this?
[01:36] <twb> I wouldn't bother
[01:36] <jetole> how come?
[01:37] <twb> However mke2fs in /etc/mke2fs.conf has something like -T small or -T news
[01:37] <twb> jetole: because it's never been an issue for me
[01:38] <jetole> twb: do you mind if I ask how much space one of your mail servers use for mail storage and what file system you use?
[01:38] <twb> For example, look at this filesystem configured normally, storing only maildirs:http://paste.debian.net/135873/
[01:39] <jetole> one sec, want to take another look at /etc/mke2fs.conf
[01:39] <jetole> It looks like if I use -T news I will have one inode for each per block
[01:40] <twb> Goddamn kernel packages
[01:40] <jetole> at least on my system it shows the default blocksize as 4096 and news has inode_ratio set to 4096
[01:40] <twb> Recommends: grub-pc | grub | lilo (>= 19.1)
[01:40] <jetole> twb: did I understand that right
[01:40] <twb> ^^ gee thanks, I really need a kernel upgrade to try blowing away my working bootloader with a broken one
[01:41] <jetole> lilo. heh, that takes me back through nostalgia
[01:41]  * twb looks at pinning again to prevent same
[01:42] <jetole> hmmmm
[01:42] <jetole> I didn't realize I could view inodes used with df
[01:42] <jetole> which turns out not to be the problem I thought it was
[01:42] <twb> Now you do
[01:44] <jetole> my /home/vmail partition is using 43G out of 200G and has ocfs2 file system (though I was going to implement ext3 or 4 on the new disk) and I started receiving out of disk space errors today
[01:44] <jetole> twb: any idea what may have caused that?
[01:44] <jetole> oh and I am using 11M out of 50M inodes
[01:44] <twb> I know nothing of OCFS
[01:44] <twb> But my first guess would be per-user or -group quotas
[01:44] <twb> Try repquota -a
[01:45] <jetole> twb: fair enough but do you know of any common reasons that may happen on other FS's?
[01:45]  * jetole looks
[01:45] <twb> Otherwise df -h and df -hi, see if anything is full
[01:45] <jetole> I don't have a requota command
[01:45] <twb> *repquota
[01:45] <jetole> rep<tab><tab>: replace report-hw
[01:46] <jetole> both disk space and inode usage, both are 22%
[01:46] <twb> Obviously you'll need to use repquota on the host that has the filesystem mounted, not e.g. an NFS client, and probably needs to run as root.  But if it's not installed and there's no mention of quota in /proc/mounts, then I guess you don't have quotas set up
[01:46] <twb> If OCFS is anything like btrfs, it could also simply have gotten its knickers twisted about something else, filesystem-specific
[01:46] <jetole> yeah I don't hve quota installed
[01:46] <jetole> I did a apt-file search then used dpkg --get-selections to check
[01:47] <twb> That's great, whatever
[01:47] <jetole> twb: I was explaining how I found out what package it is and then how I... ah nevermind. you're right @ whatever
[01:48] <jetole> twb: it may have gotten it's knickers twisted about god knows what but worth checking the generic causes first since people who know ocfs2 may be harder to track down
[01:48] <jetole> twb: btw, ocfs2 is a clustered FS from oracle
[01:49] <jetole> a free one that's in the packages but that's what it does and who makes it
[01:49] <twb> Shrug
[01:49] <jetole> meh, in case you had cared
[01:50] <twb> FFS, does pinning even work AT ALL in ubuntu
[01:52] <jetole> twb: what are you talking about at pinning
[01:52] <twb> http://paste.debian.net/135875/ <-- fucked
[01:53] <jetole> hey watch your language
[01:53] <jetole> I personally don't give a fuck but everyone always says that
[01:53] <jetole> and yeah, don't know about pinning. read something about it a year or more ago but I haven't really needed to use it
[01:54] <twb> jetole: I don't expect *you* to, but I hoped someone else was awake.
[01:54] <jetole> twb: don't supposed you can dpkg -i thispkg.deb for the version you need then echo thispkg hold | dpkg --set-selections
[01:54] <jetole> would that work?
[01:55] <twb> jetole: as I said, I need grub to NEVER EVER EVER be installed, and the kernel packages are doing their best to break that.
[01:55] <jetole> oh
[01:55] <twb> Pinning it to -1 should be the trivial solution.
[01:55] <jetole> again, can't you place a hold on the package?
[01:55] <twb> You can't hold non-installed packages
[01:57] <jetole> yep. guess you're right
[01:58] <jetole> however you can hold uninstalled packages if that helps
[02:01] <twb> Doesn't help
[02:01] <twb> As at lucid, anyway
[02:02] <twb> http://paste.debian.net/135876/
[02:09] <jetole> twb: it works well. it's aptitude that doesn't: http://pastebin.com/E65sNwHA
[02:09]  * jetole goes for a smoke, bbiab
[02:18] <jetole> twb: did you notice that aptitude didn't actually place a hold on the package?
[02:18] <twb> I didn't
[02:18] <jetole> 22:09 < jetole> twb: it works well. it's aptitude that doesn't: http://pastebin.com/E65sNwHA
[02:18] <twb> I saw that, yes
[02:19] <jetole> aptitude didn't place a hold on the package in that paste but dpkg --set-selections did
[02:22] <jetole> twb: every file needs a inode and needs only one inode. right?
[02:23] <twb> Every object in the filesystem requires one inode
[02:23] <twb> e.g. a directory has an inode
[02:23] <twb> Hard links share one inode between all their copies, IIRC
[02:24] <jetole> also, do you know a quick way I can see how many files exist on a file system? I tried `find /home/vmail -type f | wc -l` | wc -l but this is taking forever
[02:24] <twb> I can't think of a case where one object would need more than one inode
[02:24] <twb> df -i
[02:25] <jetole> damn. 11072746
[02:26] <jetole> I've decided to go with the new disk and ext4 and using -T news it only created 250k inodes
[02:27] <twb> That's an order of magnitude more than you are using
[02:27] <jetole> I'm using 11 million according to df -hi
[02:28] <twb> Oh sorry, misread
[02:28] <jetole> I know of the top of my head one user has 25k files in his inbox plus he has a bunch of other maildir's / folders
[02:28] <twb> Fiddle with -i or so, I guess
[02:29] <jetole> yeah I'm looking into it
[02:29] <twb> It does seem odd that you're hitting this, though
[02:30] <jetole> it's a mail server for our entire corporation as well as other domains outside our company
[02:31] <twb> Yes, but the number of inodes is proportional to the filesystem's overall size
[02:31] <jetole> 43GB of mail
[02:31] <jetole> both the current disk and the new one are 200GB
[02:32] <jetole> and if I understand everything right, 250k inodes means we can only store 250k files and directories. Is that right?
[02:32] <twb> Yes
[02:34] <twb> Hmm, that's odd
[02:34] <jetole> I'm checking my workstation and my home machine and I see different specs on inode counts so going to play with mkfs and the options and see what I can come up with
[02:35] <twb> http://paste.debian.net/135878/
[02:35] <jetole> I don't suppose the fact that the mail server is Ubuntu server 9.04 would have anything to do with it?
[02:35] <twb> Doubt it
[02:36] <jetole> I'll figure it out
[02:37] <jetole> did you have any luck using echo grub-pc hold | dpkg --set-selections then running aptitude?
[02:38] <jetole> hmmm, changing from ext4 to ext3 I can see mkfs is using a much higher inode count in mkfs and I think ext4 was still pretty new @ 9.04
[02:38] <twb> I didn't do that because I want pinning to work
[02:39] <twb> I didn't think ext4 was available as at 9.04
[02:39] <jetole> oh. so even if a hold works you still want to see it work with pinning?
[02:40] <jetole> yeah I don't think it was on a fresh install and may have been through a upgrade. or maybe it was but it wasn't even an option to format / and other partitions with it during install
[02:40] <jetole> but yeah, point being ext4 is new under 9.04
[02:42] <jetole> oh wtf. df -hi /dev/vda1 still showed only 249k inodes after mkfs was done but I used tune2fs -l like you pasted and it shows the 52 million that mkfs said
[02:42] <jetole> I'm betting it was the same shit for ext4 too and df was the app shitting bricks. not the FS
[02:43] <twb> I dunno man, all I really know is I've never actually run out of inodes in normal usage
[02:44] <jetole> yeah I don't think I have either but it was saying that only 249k existed and I know I am already using more then that
[02:53] <rhb5d> Question: how do I shut down Ubuntu server from the login prompt?
[02:54] <rhb5d> I forgot my username and password, and I'm running it on VMWare, and I just want to turn it off but can't seem to figure out how
[02:55] <rhb5d> I can restart it through VMWare's settings, but can't shut it down properly
[03:00] <patdk-lap> rhb5d, push the power button :)
[03:01] <rhb5d> Is there a power button in VMware?
[03:03] <twb> Probably
[03:03] <twb> patdk-lap: unfortunately acpid isn't installd by default so that won't work >:-/
[03:04] <twb> rhb5d: unless you've installed acpid, your best bet will be to hit Ctrl+Alt+Del, which will trigger a clean *reboot*; then do a hard power off once you get to the bios or grub stage, where it is safe to do so.
[03:04] <twb> rhb5d: if acpid was installed, you could simply tell VMware to hit the (virtual) power button, which would trigger a clean shutdown
[03:05] <twb> Apparently ubuntu servers don't install acpid by default because "you could accidentally bump the power button when waddling through the machine room"
[03:06] <patdk-lap> twb, odd, works on mine :)
[03:06] <patdk-lap> but then acpid probably got installed for the monitoring software (cpu/temp/fan/....)
[03:07] <twb> patdk-lap: right
[03:07] <patdk-lap> works for me :) don't care about you :)
[03:07] <patdk-lap> but then he is using vmware, so no powerbutton to hit
[03:08] <rhb5d> ctrl + alt + del doesn't do anything.. i can restart it through VMware, but can't get it into grub
[03:08] <rhb5d> it just loops back to the login prompt
[03:08] <patdk-lap> you have to hold right shift for grub2 to load up
[03:09] <rhb5d> it has to be the right shift key?
[03:09] <patdk-lap> yes
[03:09] <rhb5d> maybe that's why it hasnt worked :)
[03:09] <patdk-lap> except for grub1 it's left control key, I think
[03:09] <twb> rhb5d: unfortunately in their infinite wisdom, the ubuntu developers made it nearly impossible to get into grub on fast machiens
[03:10] <patdk-lap> ya, my normal thing is to set grub to 2second timeout, and enable the menu
[03:10] <twb> rhb5d: unlike extelinux, where you can leave the scroll lock or num lock key down to get into it, with grub you have to hit shift or alt EXACTLY when grub expects it
[03:10] <rhb5d> yes! made it into grub!
[03:10] <rhb5d> you have no idea how happy that alone makes me :)
[03:10] <twb> rhb5d: OK, now pick single and break in
[03:10] <patdk-lap> twb, heh, openindiana does a default 30second grub timeout and menu on every boot
[03:10] <patdk-lap> that annoys me the other way :)
[03:10] <twb> Oh, if you have a root password you'll need to init=/bin/sh
[03:17] <qman__> grub1 it's escape
[03:17] <qman__> and it gives you a 3 second prompt to do so
[03:20] <rhb5d> thank you for your help!
[03:49] <scalability-junk> where could be the bottleneck, when I can get a maximum of 1MB/s transferrate through one connection?
[08:19] <SmokeyD> hey everyone, I am having trouble with an ubuntu server I have running in as virtualbox guest. I cloned this ubuntu server from another virtual machine that is running for a long time already without problems.
[08:20] <SmokeyD> But with this vm, I there seems to be a cronjob or something that makes the machine crash somehow.
[08:21] <SmokeyD> the machine runs fine during the day, but if I try to login the next morning through ssh, I immediately get asked for my password, but after entering it, nothing seems to happen, untill I hit ctrl-c
[08:21] <SmokeyD> then the server load and other login info is displayed and I get a bash prompt on the machine
[08:21] <SmokeyD> but then it doesn't react anymore to keystrokes or anything else.
[08:22] <SmokeyD> If I check the host OS. the virtualbox process doesn't have any cpu load or very high memory usage. Other vm's on the same host, and the host itself are also working normally
[08:23] <SmokeyD> so I am suspecting that the ubuntu server guest os somehow becomes unresponsive due to some cronjob that starts at night.
[08:23] <SmokeyD> does anyone have any clues as to what could cause this behaviour
[08:25] <SmokeyD> the same unresponsive behaviour exists when I log in directly to the machine (using vbox rdp to the host os), so it also isn't a problem with ssh
[08:25] <memoryleak> SmokeyD: Did you check syslog?
[08:26] <SmokeyD> memoryleak, I did on the host, and I am about to reset the guest os so I can log in again and check syslog on the machine. Right now I can't do anything on it
[08:26] <SmokeyD> on the host I couldn't see anything out of the ordinary
[08:26] <memoryleak> tail to syslog and try to login
[08:27] <SmokeyD> memoryleak, after the reset you mean? I am very certain that when I reset the guest os (tried that friday and tuesday) it will work as expected again, untill tomorrow. But I can check syslog on the guest os to see if some cronjob got stuck or something
[08:28] <SmokeyD> but I do need to reset the guest os before I can do anything on it again
[08:32] <SmokeyD> memoryleak, I just reset the guest os, and am checking syslog, but there are no messages between 05:29:02 this morning and 10:29:11 (which is right now when I reset the machine)
[08:32] <memoryleak> Well what about sshd log?
[08:32] <SmokeyD> while normally there is a dhcp request every five minutes in the syslog, there is not syslog message in the 6 hours this morning
[08:33] <memoryleak> SmokeyD: /var/log/auth.log ?
[08:34] <SmokeyD> memoryleak, there is a CRON login message in auth.log at 05:17 (30 minutes before the last dhcp request log in syslog), and then no more messages untill 10:29
[08:35] <SmokeyD> so something is causing the machine to freeze at around 05:49 this morning (almost 6 hours ago)
[08:35] <SmokeyD> correction: 5 hours ago
[08:35] <memoryleak> Whit what you virtualize ?
[08:35] <SmokeyD> virtualbox
[08:36] <SmokeyD> but there are other machines on the same host in virtualbox (including the one the problem os was cloned from) that run without any trouble
[08:36] <SmokeyD> and I checked syslog on the host os  (also ubuntu server) and there is nothing out of the ordinary there.
[08:39] <SmokeyD> the strange thing is that cron.daily on the guest os is not started untill 6:25 while the freeze seems to occur at 05:49, so it doesn't seem to be triggered by the daily cronjobs either
[09:14] <jamespage> Daviey: around?
[09:31] <Daviey> jamespage: hola
[09:31] <jamespage> Daviey, morning
[09:31] <jamespage> I'm just sweeping up any outstanding ISO tests
[09:32] <jamespage> having an issue with i386 RAID test
[09:32] <jamespage> install fine, boot fine, boot degraded from the first disk
[09:33] <jamespage> won't boot from the second disk - it looks like grub bombs out because it just resets itself
[09:33] <jamespage> and goes back to the grub menu - any ideas?
[09:33] <jamespage> i can't even run any commands from the grub command line - it just dies and resets
[09:34] <jamespage> I've complete torn down and re-setup my test VM - but its still not happy
[09:49] <koolhead17> hi all
[09:49] <lynxman> jamespage: Daviey: morning o/
[09:49] <lynxman> koolhead17: morning to you too :)
[09:49] <jamespage> moring all
[09:49] <koolhead17> hey lynxman :P
[09:49] <koolhead17> morning jamespage Daviey too :P
[09:50] <koolhead17> news of the day http://www.tehelka.com/story_main50.asp?filename=Ws101011MICROSOFT.asp :D
[09:50] <_ruben> interesting .. postfix 2.8.5 in lucid-backports .. guess i should dust off my apt-pinning skills :p
[09:56] <jamespage> Daviey: hmm I get the same on amd64 - wonder it its something about my vm
[10:01] <Daviey> jamespage: eeek
[10:01] <jamespage> Daviey: hmm
[10:02] <jamespage> I get the same issue with the amd64 install
[10:02] <jamespage> I've tried switching between virtio and default disk types
[10:02] <jamespage> and ensuring that the disk is always presented as the first ie. vda or hda
[10:02] <jamespage> but no success
[10:02] <cjwatson> jamespage: so if I did an install with kvm and two disks, would I probably be able to reproduce this?
[10:03] <cjwatson> pull one then the other
[10:03] <jamespage> cjwatson, well unless its something specific to where I'm running the test - yes
[10:03] <jamespage> thats exactly the configuration I use
[10:03] <cjwatson> jamespage: could you summarise your partitioning layout for me?
[10:04] <jamespage> I'm using 2 x 2GB qcow2 images to back virtio disks
[10:04] <jamespage> each disk is partitioned with three primary volumes - 1.5GB, 512M and the remainder - all set as type raid
[10:04] <jamespage> The 1.5GB's are RAID-1 into / as ext4
[10:05] <jamespage> the 512M as swap
[10:05] <cjwatson> all raid-1?
[10:05] <jamespage> and the remainder as /home as ext4
[10:05] <jamespage> yep
[10:05] <cjwatson> ok, let me have a go
[10:05] <jamespage> I can pull the second disk OK
[10:05] <jamespage> and restore it and resync
[10:05] <jamespage> I just can't pull the first and get it to boot
[10:05] <jamespage> amd64 or i386
[10:06] <jamespage> cjwatson, ta
[10:08] <cjwatson> jamespage: 1.5GB, 512M, and the remainder - but you're using 2GB disks, isn't the remainder nothing?
[10:08] <cjwatson> or is there a bit of rounding error at the end?
[10:11] <jamespage> cjwatson, hmm - now that is odd
[10:11] <jamespage> qemu-img info reports the disk a 2.0GB
[10:12] <jamespage> fdisk from within the kvm reports 2147 MB
[10:12] <cjwatson> ok, so different GB
[10:13] <cjwatson> i.e. GB vs. GiB
[10:13] <jamespage> ah
[10:13] <jamespage> so rounding error at the end I guess
[10:14] <jamespage> anyway - I end up with a 124MB /home
[10:15] <tjaalton> is there a "what next" document to read after installing ubuntu-orchestra-server? the packages only include the changelog & copyright in /usr/share/doc
[10:16]  * jamespage coffee - biab
[10:18] <lynxman> tjaalton: not yet, we're on it though :)
[10:18] <tjaalton> lynxman: ok, any pointers in the meantime? I noticed that the orchestra-provisioning "site" is not enabled by default
[10:19] <lynxman> tjaalton: hmm you mean the web ui?
[10:19] <tjaalton> lynxman: right, though any ui would do :)
[10:20] <lynxman> tjaalton: don't think we have one active yet :( it's all command line
[10:20] <tjaalton> ok that's fine
[10:21] <lynxman> tjaalton: let me see if I can find my notes for you, one sec
[10:21] <tjaalton> lynxman: thanks
[10:24] <bau-> I all, I have register a new domain with godaddy, now I can tell it to go to my server when I go to that site?
[10:24] <lynxman> bau-: hmm wrong channel? :)
[10:24] <lynxman> tjaalton: Once you have orchestra installed and the isos imported (it does that automatically on the background the first time)
[10:25] <lynxman> tjaalton: it's time to tell cobbler to get to configure that machine by adding it to a profile
[10:25] <lynxman> tjaalton: your orchestra server needs to be the next-server for PXE boot
[10:26] <lynxman> tjaalton: then collect MAC address, give it a node name
[10:26] <lynxman> tjaalton: and run this on your orchestra server
[10:26] <lynxman> tjaalton: sudo cobbler system add --name="node01.internal.domain" --mac-address="AA:BB:CC:DD:EE:FF" --ip-address="AA.BB.CC.DD" --dns-name="node01.internal.domain" --hostname="node01.internal.domain --profile="profilename"
[10:27] <tjaalton> lynxman: thanks, I'll try that.. after lunch :)
[10:28] <lynxman> tjaalton: good luck :)
[10:47] <koolhead17> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/856417
[10:48] <koolhead17> i am still finding the same behavior of it on oneiric
[10:52] <jamespage> koolhead17, what else do you have running on that server? nova per chance?
[10:56] <jamespage> and if so which networking mode are you using?
[10:56] <koolhead17> jamespage: currently to do all my testing am running natty/daiblo setup.
[10:57] <jamespage> koolhead17, surely oneiric based on the kernel version?
[10:57] <koolhead17> jamespage: yes the system has nova running and  we are running managed vlan mode
[10:57] <koolhead17> jamespage: $ cat /var/log/installer/media-info Ubuntu-Server 11.10 "Oneiric Ocelot" - Beta amd64 (20111009)
[10:58] <koolhead17> 09/10/2011
[10:58] <koolhead17> i have dismentaled the machine to try newest built
[10:58] <jamespage> koolhead17: right - unfortunately I don't have a dual nice setup to reproduce on
[10:58] <jamespage> /nice/nic/
[10:59] <jamespage> but on my all-in-one I don't create the br100 interface - I create a br0 interface onto eth0 and use that for the private networking
[10:59] <koolhead17> jamespage: the same setup works well for the natty/daiblo setup. Am trying the newest image which just downloaded to see if this issue is resolved
[10:59]  * jamespage scratches his head
[11:00] <jamespage> koolhead17, I guess it may be a difference in the kernel handling of duplicate ops on br100
[11:01] <koolhead17> jamespage  what would you suggest to not come out of this then. :D
[11:02] <koolhead17> i mean am creating br100 on eth0 or eth1 should not make a difference
[11:02] <koolhead17> :P
[11:03] <jamespage> koolhead17, I don't think you need to create br100 at-all
[11:03] <jamespage> I only do that in my all-in-one to work around not having two interfaces
[11:04] <koolhead17> jamespage: so how will i assign IP to my internal network ?
[11:04] <koolhead17> and communicate with internal machines :(
[11:05] <koolhead17> jamespage: brb. trying out newest release
[11:05] <jamespage> koolhead17, firing in the dark a bit here - http://tinyurl.com/3c6fay8
[11:05] <jamespage> try configuring eth1 normally and then specifying --vlan_interface=eth1
[11:05] <jamespage> I think nova will then create the br100 interface using eth1
[11:05] <jamespage> *I think*
[11:06] <jamespage> actually I can test this
[11:06] <koolhead17> jamespage: cool
[11:09] <trapmax> how do i do grub-install on a kvm guest?
[11:09] <RoyK> damn - anyone here ever succeded in setting up AD-based authentication for Linux clients?
[11:11] <jMCg> RoyK: pam-ldap? Or just stuff like http auth? Kerberos integration? Or just.. blah blah blah.. There's a whole slew here. You can go from just using ldap to.. well.. complete integration.
[11:12] <RoyK> pam-ldap should do. I just wonder if I have to lock the config to a fixed context...
[11:12]  * RoyK thinks LDAP is largely confusing
[11:18] <RoyK> hrmf - failed to bind to LDAP server ldap://192.168.0.3/: Invalid credentials - but the same creds work with ldapsearch
[11:19] <RoyK> meaning kerberos must be setup?
[11:48] <jamey-uk> Suddenly my server has stopped booting, it just sits with a flashing cursor waiting to boot.
[11:49] <jamey-uk> Suddenly my server has stopped booting, it just sits with a flashing cursor waiting to boot. It's an EFI machine, and the disk seems to be GPT, but I don't know what's wrong: http://paste.ubuntu.com/706649/ Can anyone help?
[11:51] <jamespage> koolhead17: I think that works
[11:51] <koolhead17> jamespage: is it working for you the way you mentioned?
[11:52] <jamespage> OK - so lemme run this through
[11:52] <jamespage> eth0 - public interface
[11:52] <koolhead17> jamespage: yes
[11:52] <jamespage> eth1- private interface
[11:52] <jamespage>  --vlan_interface=eth1 in /etc/nova/nova.conf
[11:52] <tjaalton> RoyK: winbind + pam_krb5 should do
[11:52] <koolhead17> k
[11:53] <koolhead17> jamespage: and you assigned IP to eth1
[11:53] <jamespage> koolhead17, yes - well dhcp did - I know that is not that normal in server deployments
[11:54] <jamespage> I'm running this in a kvm so bear with me
[11:54] <jamespage> thank goodness for ssd
[11:54] <koolhead17> jamespage: what i did is assigned a static ip to eth1
[11:54] <koolhead17> so i have ip for eth1 and eyh1
[11:54] <koolhead17> *eth1
[11:54] <jamespage> I think thats spot-on
[11:55] <jamespage> I generated a set of networks -  sudo nova-manage network create private 10.0.0.0/8 3 16
[11:55] <jamespage> and then uploaded a image and ran one
[11:55] <jamespage> nova-network created:
[11:56] <jamespage> br100, vlan100@eth1 and veth0 with is on br100
[11:56] <jamespage> I can access the instance using its private address from the nova server
[11:57] <jamespage> I don't get any of those loop messages but I do see the br100 being setup
[11:58] <jamespage> koolhead17: make sense so far?
[11:59] <jamey-uk> Does anyone know how to get GRUB working with a GPT partition on a UEFI machine?
[12:00] <Jeeves_> qemu-img: Error while formatting '/dev/data_disk/fliep'
[12:01] <Jeeves_> That's when I try to qemu-img convert to an lvm-device
[12:01] <Jeeves_> Any known fixes for that?
[12:01] <koolhead17> jamespage: indeed :)
[12:01] <jamespage> koolhead17, it may be you get those loop messages because nova-compute is trying todo something to the bridge you already have setup
[12:02] <koolhead17> k
[12:02] <koolhead17> let me try as you mentioned
[12:03] <jamespage> koolhead17, I've not tried running nova + lxc in a kvm before - actually worked quite well!
[12:04] <jamespage> koolhead17, fwiw even if this does resolve your issue - I think that there might still be a bug in oneiric which you did not hit in natty
[12:04] <koolhead17> jamespage: so sumthing has changed in oneiric because we are still using the same old config creating br100 manually and it all works well
[12:11] <RoyK> 11.10 countdown...
[12:14] <jamespage> koolhead17, agreed - but I don't think that creating the br100 interface is required any more
[12:17] <koolhead17> jamespage: am still in same proble.
[12:17] <koolhead17> problem
[12:18] <koolhead17> is your eth1 and sudo nova-manage network create private 10.0.0.0/8 3 16  in same ip range
[12:18] <tjaalton> lynxman: actually, i got the cobber web-ui login screen up, but the password doesn't seem to work
[12:18] <jamespage> koolhead17, no - they don't need to be
[12:19] <jamespage> koolhead17, the address on eth1 will be on the native vlan
[12:19] <jamespage> whereas the private network ranges will sit on different vlans
[12:21] <jamespage> koolhead17: can you post /var/log/nova-*.log and your syslog and kern.log on the bug report
[12:21] <jamespage> we might be able to identify what is causing the issue from that
[12:25] <koolhead17> http://paste.ubuntu.com/706669/
[12:28] <koolhead17> jamespage: kernel log  http://paste.ubuntu.com/706670/
[12:29] <jamespage> koolhead17, can we get more of nova-network.log
[12:31] <jamespage> koolhead17, hmm - what parameters did you use with nova-manage network create?
[12:32] <jamespage> in my setup I see vconfig operating on eth1 not br100
[12:33] <koolhead17> yes
[12:33] <jamespage> cjwatson: any progress on reproducing my raid issue?
[12:37] <koolhead17> nova-manage network create --bridge_interface=br100 --fixed_range_v4=192.168.3.0/29 --label=proj nova-manage floating create --ip_range=10.10.10.224/27
[12:38] <koolhead17> this is what we used diablo plus natty
[12:39] <tjaalton> lynxman: I just get 'httpd does not appear to be running and proxying cobbler' when trying to add a system
[12:39] <tjaalton> same thing with 'cobbler check'
[12:40] <koolhead17> tjaalton: is apache running?
[12:41] <tjaalton> koolhead17: yep
[12:41] <zul> morning
[12:43] <lynxman> zul: morning o/
[12:43] <koolhead17> morning zul
[12:43] <koolhead17> tjaalton: try restarting the orchestra server once
[12:43] <koolhead17> jamespage: should i dump whole network log along with the bug report?
[12:45] <tjaalton> koolhead17: what's it called? can't find anything with that name
[12:46] <RoyK> hi all. giving btrfs a try again, now on the soon-to-be-released Ubuntu 11.10 - now, is there a way to do something similar on that as what I do on solaris/openindiana with 'beadm create'? that will create a clone of the root and allow me to boot into that, potentially after an upgrade (of that)
[12:47] <jamespage> koolhead17, this is creating the issue - --bridge_interface=br100
[12:48] <jamespage> but I don't understand why it worked OK with diablo on natty
[12:48]  * jamespage scratches his head
[12:49] <koolhead17> jamespage: i am cleaning everything and this time not give the br100 flag.
[12:49] <koolhead17> jamespage: same is my situation/scarchy :)
[12:49] <jamespage> koolhead17, worth a try
[12:50] <koolhead17> i have working natty/diablo running with same commands :)
[12:51] <jamespage> koolhead17, I suggest you pastebin the log for the time being
[12:52] <koolhead17> tjaalton: pas-ef|grep orchestra gives notthing :(
[12:52] <koolhead17> jamespage: i dont think the network log has anything more relevant to this issues besides http://paste.ubuntu.com/706669/
[12:53] <koolhead17> * ps -ef | grep orchestra
[12:53] <jamespage> koolhead17, OK - so try with the --vlan-interace set in /etc/nova/nova.conf and don't specify the bridge_interface when creating the network with nova-manage
[12:53] <tjaalton> koolhead17: nope
[12:54] <koolhead17> jamespage: sure cleaning up all the the installs and try as yuo mentioned. :)
[12:55] <jamespage> koolhead17, ok - lemme know how it goes.
[12:55] <koolhead17> sure
[12:55] <koolhead17> lynxman: hellos
[12:56]  * koolhead17 pokes RoAkSoAx: 
[12:56] <lynxman> koolhead17: hey hey
[12:56] <koolhead17> jamespage: the only problem/confusion i have if its working on natty/diablo what have we changed in in oneiric :P
[12:57] <koolhead17> lynxman: what is default orchestra process, tjaalton needs some help :)
[12:57] <lynxman> koolhead17: cobbler
[12:57] <tjaalton> right, restarting cobbler didn't help
[12:59] <koolhead17> lynxman: am having nightmares in understanding RBAC of nova/
[12:59] <koolhead17> i created like 15 users and added like 5-6 projects
[12:59] <koolhead17> :P
[12:59] <cjwatson> jamespage: I've reproduced it; I'm working on debugging it (amid other things)
[13:01] <jamespage> cjwatson, great - let me know if I can help in anyway
[13:02] <cjwatson> will do
[13:18] <hggdh> jamespage: good morning, to which Jenkins test we map the tracker "Ubuntu Server EC2 HVM (US-East) amd64"?
[13:19] <jamespage> hggdh: we don't
[13:19] <hggdh> ah
[13:19] <jamespage> normally utlemming or smoser run that one by hand
[13:22] <koolhead17> tjaalton: check if there is allready a question with same else post  one with all details, am sure someone will look into it and answer back. :)
[13:29] <tjaalton> koolhead17: i suspect the network settings are somehow screwed, will fix them first
[13:31] <hggdh> jamespage: k, I will wait for utlemming or smoser
[13:48] <koolhead17> jamespage: success finally :P
[13:49] <koolhead17> but need to get 100% sure first :)
[13:49] <jamespage> koolhead17, \o/
[13:51] <koolhead17> jamespage: you made my day<rather 7.20 pm> for me
[13:51] <koolhead17> jamespage: so what should i do to the reported bug?
[13:52] <jamespage> koolhead17, so it looks like the bridge was getting in a right twist with the config you where using
[13:56]  * jamespage has another look at the bug
[13:57] <koolhead17> jamespage: yes. indeed
[13:58] <koolhead17> jamespage: shall i request for its clouser then?
[13:58] <jamespage> koolhead17: still thinking
[13:58] <koolhead17> :(
[13:58] <koolhead17> i mean only thing i don`t understand why it was working with natty and not with oneiric
[14:00] <jamespage> koolhead17: are you sure that the nova configuration was identical on natty?
[14:01] <koolhead17> jamespage: same configuration to be precise. :)
[14:04] <koolhead17> Daviey: ping
[14:04] <Daviey> koolhead17: hey
[14:05] <jamespage> koolhead17: Its an odd edgecase; lets document what we have found out today and mark it as a priority low....
[14:05] <jamespage> I'll update the report now
[14:05] <koolhead17> jamespage: cool
[14:06] <koolhead17> i have all logs/details added to it. hardware make was allready provided to it
[14:07] <koolhead17> Daviey: keystone/diablo will have no role based access control
[14:11] <koolhead17> TeTeT: hellos :)
[14:11] <Daviey> koolhead17: oh?
[14:12] <Daviey> utlemming: Hey, around?
[14:12] <utlemming> Daviey: sir, I am
[14:12] <Daviey> utlemming: Great!  Is the /current AMI's from your perspective final?
[14:12] <Daviey> anything concerning you?
[14:13] <utlemming> I'm doing some final digging around, but 20111011 is the one that smoser and I tagged as the candidate
[14:14] <utlemming> the tests look good (other than EC2 being EC2)
[14:15] <utlemming> Daviey: also, the desktop images are functional too
[14:15] <Daviey> utlemming: rocking!
[14:16] <Daviey> jamespage: you also smoked the ec2 images?
[14:16] <utlemming> https://jenkins.qa.ubuntu.com/job/oneiric-server-ec2/
[14:16] <jamespage> Daviey: that link ^^
[14:16] <jamespage> but looked good to me as well
[14:16] <jamespage> the only tests not covered are hvm
[14:17] <Daviey> good o'
[14:17] <Daviey> thanks!
[14:24] <tjaalton> koolhead17, lynxman: ok I got cobbler moving, had a wrong ip for the server :/
[14:25] <lynxman> tjaalton: cool!
[14:26] <tjaalton> now the old one is still at least in the profiles
[14:26] <koolhead17> tjaalton: awesome.
[14:27] <koolhead17> i wonder if we should comeup with new section in launchpad and assign section per project mentioning "common typos" :D
[14:27] <tjaalton> hmm maybe I'll just reinstall the stack, to get a fresh start
[14:27] <koolhead17> that way we can read few allready created typos beffore filling bugs :)
[14:27]  * koolhead17 just an idea :P
[14:28] <tjaalton> the problem here was that the machine used dhcp provided by the router, so when I switched to a static ip things were messed up
[14:32] <KM0201> does anyone know of a photo program, that has a PNP plugin, so i can browse a share? --- xbmc works for this, but I was hoping to use something designed just for pics
[14:33] <tjaalton> ha, purging cobbler doesn't remove /etc/apache2/conf.d/cobbler*.conf symlinks, so restarting apache2 fails
[14:34] <koolhead17> tjaalton: pass me the blog page once your done with your documented the typo :)
[14:37] <w00> hm, wasn't http://thisisthecountdown.com/ supposed to 'expire' today?
[14:37] <tjaalton> koolhead17: maybe i'll just file this one bug ;)
[14:37]  * koolhead17 kicks tjaalton :D
[14:38] <koolhead17> tjaalton: file it as a questions and another add as answer as comment. it will be cool :)
[14:38] <tjaalton> koolhead17: which one?
[14:38] <tjaalton> messing with the ip?
[14:39] <koolhead17> tjaalton: yes
[14:39] <smoser> jamespage, yellow is failure ?
[14:40] <jamespage> smoser: yellow indicates one or more tests failed
[14:40] <uksysadmin> hello all. who am I best bothering about Orchestra and OpenStack deployments?
[14:40] <jamespage> smoser: red is something really bad happened
[14:40] <koolhead17> uksysadmin: to all of us
[14:40] <koolhead17> :D
[14:40] <uksysadmin> :) koolhead17
[14:41] <jamey-uk> Every time I try to install Ubuntu Server, it successfully installs but then my UEFI system will just stay with a blinking cursor waiting to boot. I've formatted the disk in MBR format but this still happens. How can I get my system bootable?
[14:41] <koolhead17> uksysadmin: shoot your question :)
[14:43] <smoser> jamespage, so https://jenkins.qa.ubuntu.com/job/oneiric-server-ec2/10/ARCH=amd64,REGION=ap-northeast-1,STORAGE=instance-store,TEST=multi-instance,label=ubuntu-server-ec2-testing/testReport/i-dc3438dd_test_remote/RemoteEC2Test/testRootSSH/? that test failed, where would the output fo the ssh be ?
[14:43] <jamespage> smoser: its in the console output
[14:43] <uksysadmin> Where can I find information on using Orchestra to do OpenStack deployments? Recent events in Paris teased me with people talking about Juju and OpenStack installs... point me in the direction and you've got a willing volunteer for testing and development
[14:47] <RoAkSoAx> uksysadmin: howdy, we are in the process of finishing up the documentation
[14:47] <RoAkSoAx> uksysadmin: the orchestra part so far is documented here: https://help.ubuntu.com/community/Orchestra
[14:47] <RoAkSoAx> though its still in process
[14:48] <uksysadmin> Excellent RoAkSoAx - I appreciate things are developing at a fast pace and OpenStack is also a moving target - patience is a virtue I'm told - but where's the fun in that?
[14:51] <RoAkSoAx> ehhe
[14:51] <RoAkSoAx> indeed
[14:51] <smoser> jamespage, https://code.launchpad.net/~ubuntu-server-iso-testing-dev/+junk/ec2-automated-tests is the tests for ec2 ?
[14:52] <koolhead17> RoAkSoAx: hey!!
[14:54] <RoAkSoAx> koolhead17: o/
[14:57] <jamey-uk> I can't get Ubuntu Server installed on my UEFI-based system, it just won't boot after installing grub (supposedly succesfully) at the end of the installation. How can I get it booting? :(
[14:59] <smoser> hggdh, do you know ?
[15:01] <marshall> hey ubuntu-server
[15:01] <marshall> whenever I try to run this script, it immediately just says "Killed"
[15:02] <koolhead17> "this script"
[15:03] <marshall> koolhead17: it's asf_mapready, the GIS software
[15:04] <marshall> koolhead17: I was able to run it before, but now even when I try to do `asf_mapready --help` it just says "Killed". I was wondering if it's a thing in linux or ubuntu that kills it before anything happens.
[15:07] <kirkland> uksysadmin: kim0|holiday has some blog posts on the matter
[15:08] <uksysadmin> thanks kirkland
[15:11] <uksysadmin> I've just googled his blog - I'll go poking around it. cheers.
[15:18] <jamespage> smoser: sorry OTP
[15:18] <jamespage> smoser: https://code.launchpad.net/~ubuntu-server-ec2-testing-dev/+junk/ec2-automated-tests
[15:20] <Doonz> ey guys im having trouble connecting to my ssh server from outside of my network.
[15:21] <Doonz> my firewall is showing that it is passing the connection through
[15:21] <Doonz> but im not sure what to do with tcpdump data ive collected (dont know howto interpret it) any one willing to spend some time helping me on this
[15:24] <robbiew> RoAkSoAx: Daviey: any idea what the current status of bug 850880 is?  If it's not fixed, we should retarget.
[15:26] <tjaalton> hmm is it just me, or aren't the cobbler repo urls all broken? (http://archive.ubuntu.com/ubuntudists//oneiric)
[15:27]  * koolhead17 is happy!!! :)
[15:29]  * RoAkSoAx looks
[15:30] <robbiew> RoAkSoAx: looking at the bug or tjaalton's issue...or both? :)
[15:30] <Daviey> kirkland: around?
[15:30] <kirkland> Daviey: eeellloooooooo
[15:30] <RoAkSoAx> robbiew: both
[15:31] <RoAkSoAx> :)
[15:31] <Daviey> kirkland: have you seen bug 872553?
[15:31] <RoAkSoAx> tjaalton: yeah that's something we just noticed too and which I'll look at
[15:31] <RoAkSoAx> Daviey: he's aware
[15:31] <tjaalton> RoAkSoAx: ah, good :)
[15:31] <Daviey> RoAkSoAx / kirkland: Is there a plan?
[15:32] <RoAkSoAx> Daviey: im gonna nlook at it today, SpamapS seems to have found the cause
[15:33] <Daviey> RoAkSoAx: well right now, it's feeling safer to pull out the recommends of etckeeper and resolve it in P-Series properly
[15:33] <robbiew> +1
[15:33] <RoAkSoAx> Daviey: indeed
[15:35] <kirkland> Daviey: it should be trivial to solve via sru, either a) (preferably) lowering priority of debconf question, or (b) preseeding the value in orchestra itself
[15:35] <kirkland> Daviey: i asked RoAkSoAx to look at it yesterday, i think he's on it
[15:35] <Daviey> kirkland: Well right now juju+orchestra deployments are blocked.  Do you disagree that ripping out the recommends is safer at this stage?
[15:36] <Patriot> Hello to everybody
[15:36] <kirkland> Daviey: seems a little on the heavy handed side, but I can see why you're asking for it
[15:37] <kirkland> Daviey: it was a late addition (at my urging), and I take responsibility for this breakage
[15:37] <kirkland> Daviey: gimme 10 minutes to take a look
[15:37] <robbiew> can't we remove it, and then once fixed, SRU it back in?
[15:37] <Daviey> Oh sure, i agreed with the rational to add it.
[15:37] <RoAkSoAx> kirkland: there's no debconf question for installtion though
[15:38] <Daviey> robbiew: 0-day SRU IMO should either remove it, or fix it.  I don't think it's a good idea to have a later SRU reintroducing the recommends.
[15:38] <kirkland> Daviey: right -- I'm with you -- SRU'ing a recommends will almost certainly be nacked by the SRU team
[15:38] <robbiew> Daviey: fair enough
[15:39] <kirkland> interestingly, etckeeper has debconf questions, but no debian/config
[15:40] <kirkland> the only db_input I see in etckeeper's source is
[15:40] <kirkland> postrm:                         db_input critical etckeeper/purge || true
[15:40] <koolhead17> kirkland: etckeeper has no debconf options :)
[15:41] <RoAkSoAx> kirkland: yeah that's what I was saying, it only has debconf questions for the postrm
[15:41] <uksysadmin> kirkland, thanks for point me in the direction of kim0|holiday - gets me passed the first hurdle at least and seems the right place to keep an eye on for following developments towards OpenStack with Orchestra
[15:42] <kirkland> uksysadmin: kim0|holiday is awesome :-)
[15:43] <uksysadmin> he really should power off his computer though. save energy and all that ;-)
[15:44] <kirkland> aha!
[15:44] <kirkland> Daviey: okay, i'm on it
[15:45] <kirkland> Daviey: it's not the installation of etckeeper itself
[15:45] <kirkland> Daviey: its that juju installs a package that modifies /etc, and these changes need to be committed to etckeeper
[15:45] <kirkland> Daviey: we just need to configure etckeeper to do that automatically when etckeeper is installed by orchestra
[15:47] <Patriot> I am in desperate need of help, and if this is not the right chat can you please advise me which channel to go to
[15:48] <Daviey> kirkland: can you provide some more detail?
[15:48] <Patriot> Will do
[15:48] <Patriot> Ok this is the problem that i am pulling my hair out
[15:48] <Patriot> I have installed Ubuntu Server 10.04 LTS
[15:49] <Patriot> I am setting up the server for domain logons
[15:49] <Patriot> Everything is going 100%, until i want to add a Windows 7 machine
[15:49] <Patriot> The windows 7 machine logs on to the domain 100%
[15:50] <Patriot> Untill i restart then the windows 7 machine says can't find the domain
[15:50] <Patriot> I just re-installed the server again from scratch
[15:50] <kirkland> Daviey: hang on, let me get this fixed, and i'll explain it to you
[15:51] <Patriot> I also applied the windows 7 registry settings etc
[15:51] <Patriot> But no luck at all
[15:52] <Daviey> kirkland: ok, i have an upload in the queue backing out etckeeper.. if you can get a better fix today, i'm happy to run with yours.
[15:52] <smoser> jamespage, so i'm missing something.
[15:52] <kirkland> Daviey: i'm working as fast as i can
[15:52] <smoser> where is the stuff that runs those tests... ie launches an instance and collects console output.
[15:53] <Daviey> kirkland: rocking
[15:54] <boxybrown> quick question about fdisk
[15:54] <boxybrown> after creating a partition and writing it
[15:54] <boxybrown> is there anything special i need to do to format the drive?
[15:54] <boxybrown> before I can mount it?
[15:55] <iclebyte-work> you need to put a file system on the partition
[15:55] <jamespage> smoser: https://code.launchpad.net/~ubuntu-server-ec2-testing-dev/ubuntu-server-ec2-testing/ubuntu-server-ec2-testing
[15:55] <jamespage> sorry - misunderstood your question
[15:56] <boxybrown> iclebyte-work: okay yeah, looks like i need to run mkfs
[15:57] <iclebyte-work> yea.  mkfs.ext3 /dev/sdaX
[15:58] <boxybrown> iclebyte-work: ext3 or ext4, seems like the new ubuntus are using ext4...
[16:00] <Daviey> utlemming: around?
[16:00] <Doonz> any reason why ssh server works internally but not from any internet based ips
[16:01] <patdk-wk> firewall :)
[16:01] <patdk-wk> isp blocking it with their firewall? :)
[16:01] <Doonz> no and no
[16:02] <Doonz> i think its something with pfsense
[16:02] <Doonz> pfsense people say its the server
[16:02] <Doonz> i dont freaking know
[16:02] <patdk-wk> pfsense is a firewall
[16:02] <Doonz> wich the rules are there and its is allowing that connection and routing it to the proper port
[16:03] <patdk-wk> if it's not a firewall, then there is only one thing it could be, you modified sshd_config to not listen to the external ip
[16:03] <Doonz> i have to nics in the server
[16:03] <Doonz> eth0 is 192.168.1.32
[16:03] <Doonz> eth1 is 192.168.1.33
[16:03] <Doonz> inthe sshd_conf listen ip is 192.168.1.32
[16:04] <koolhead17> hi
[16:06] <Doonz> patdk-lap is that not correct?
[16:06] <patdk-wk> if you are connecting to ssh using that ip, yes
[16:06] <utlemming> Daviey, Jamespage: HVM AMI's passed too. So I'm happy with the 20111011 images.
[16:07] <Daviey> utlemming: did you see my Questions in -release?
[16:07] <utlemming> Daviey: just did...switching channels
[16:09] <Daviey> a
[16:09] <Daviey> ta
[16:11] <kirkland> RoAkSoAx: SpamapS: do you guys have an easy/quick way to reproduce this outside of orchestra (preferably in an EC2 instance)?  I have a fix i'd like to test
[16:16] <RoAkSoAx> kirkland: i wasn't able to reproduce it outside orchestra
[16:16] <RoAkSoAx> kirkland: maybe pass cloud-init config
[16:17] <panfist> i'm new to systems administration so sorry if this is a basic quesiton. i was wondering why when some packages are built/installed from source, they put themselves in the right place in the FHS,
[16:17] <panfist> and others just stay right where you extract them
[16:17] <RoyK> does xen support some sort of failover? I mean, afaik uec does, but with a rather rigid setup with hard to do a simple/clean setup without a single point of failure. software like vmware or hyper-v run in mesh mode, which is better
[16:18] <panfist> is there a reason for that, other than the package maintainer didn't have a part of the installation script that knows about your distro so it just defaults to not installing in a place that makes sense?
[16:18] <panfist> and if a package doesn't put itself in the right place, how do i know what that place should be?
[16:25] <patdk-wk> royk, xen doesn't support failover, but it has the ability for you to do it
[16:25] <ClaesBas> Does anybody know where installer finds out that it has been a "md-raid" (I have removed partions, dd 'ed 512 zeros to /dev/sdX, created new empty dos partion table and even created and formated a new ext2 partion on 20G), I want to start from scratch but installer only put a iscsi-thing in my face (it's with an "11.10 111010 server cd")?
[16:26] <patdk-wk> it has a way to move running domu's
[16:26] <ClaesBas> And of course have I removed superblocks with --zero-superblock....
[16:27] <patdk-wk> ClaesBas, wouldn't wiping the partition table be easier and faster?
[16:27] <patdk-wk> cause superblocks is a fs thing
[16:28] <ClaesBas> patdk-wk: I think I've done everything except dd'ing zero to whole disks (4x750G)
[16:30] <kirkland> RoAkSoAx: do you have an orchestra setup right now where you can reproduce it?  would you be willing to give a fix a quick test?
[16:30] <ClaesBas> Someone who coded this part of the "deb-installer" have been a bit to "clever"...
[16:30] <patdk-wk> claesbas, no, normally it uses blkid to find and locate drive configs
[16:30] <patdk-wk> and normally just writing 0's to the first meg of the drive is enough to remove that :)
[16:32] <ClaesBas> patdk-wk: I will try that, thanks...
[16:32] <patdk-wk> you didn't use gpt did you?
[16:32] <patdk-wk> if you did, you probably have to wipe out the first and last 1mb of the drive
[16:32] <patdk-wk> as gpt is redundant :)
[16:36] <koolhead17> jamespage: i was disconnected.
[16:38] <smoser> utlemming, it would appear that you have not started the promote-daily, is that correct ?
[16:38] <utlemming> smoser: I ran the promote daily yesterday. It's staged up and ready to go
[16:41] <RoyK> patdk-wk: moving domu's is nice, but do you know if there's a way to failover a domu to another dom0 in case of a dom0 failure?
[16:42] <RoyK> patdk-wk: by failover, I meen to administratively move it to another dom0 and boot it
[16:42] <RoyK> not true failover - afaik only vmware support that (and probably IBM zSeries, but that's another ballpark)
[16:43] <patdk-wk> oh, isn't that easy?
[16:43] <patdk-wk> just issue xm create domu on the other machine
[16:43] <RoyK> (and for vmware to support that, it needs infiniband or at least 10Gbps ethernet...)
[16:43] <RoyK> patdk-wk: automatically without risking collisions?
[16:43] <patdk-wk> heh? vmware supports that on anything,
[16:43] <patdk-wk> royk, that depends on you HA setup, I assume you used proper fencing and stuff from linux-ha
[16:44] <RoyK> patdk-wk: for a VM to be HA, vmware needs very high speed networking between the hosts to transfer the memory contents as it is changed
[16:44] <patdk-wk> royk, if the dom0 is down, there is no memory to move
[16:44] <RoyK> patdk-wk: vmware normally reboots a VM if the host it's on dies
[16:45] <patdk-wk> and I do move ram, gigs of it at a time, over gigabit with vmware
[16:45] <patdk-wk> but I think the way vmware moves ram is different from xen
[16:45] <RoyK> does that work ok? I mean - isn't gigE far too slow for that?
[16:45] <patdk-wk> as vmware will do like a double rsync to move it, and xen is all at once, locked
[16:45] <patdk-wk> royk, vm is only down for a few seconds for me, with 8gig+ ram vm's
[16:46] <RoyK> patdk-wk: that's moving a VM, yes
[16:46] <patdk-wk> with infiniband, it's down for a few seconds also
[16:46] <RoAkSoAx> kirkland: sure I have my own setup
[16:46] <RoyK> patdk-wk: but if you have the VM running on host A and you pull host A's plug
[16:46] <kirkland> RoAkSoAx: okay
[16:46] <patdk-wk> royk, then you will have to wait for the vcenter timeout
[16:46] <kirkland> RoAkSoAx: have you recreated the problem?
[16:46] <RoyK> patdk-wk: then the only way for that VM to survive, is to run on two hosts at the same time
[16:46] <patdk-wk> royk, and why would it move the ram then anyways? cause host A is gone
[16:47] <patdk-wk> royk, no, it won't
[16:47] <RoyK> patdk-wk: are we talking about the same thing?
[16:47] <RoAkSoAx> kirkland: on the process.. BW is not helping today
[16:47] <patdk-wk> what plug did you pull on host A?
[16:47] <patdk-wk> I assumed power :)
[16:47] <RoyK> the Big One
[16:47] <patdk-wk> then it's not running on Host A anymore is it? cause the power is off :)
[16:48] <patdk-wk> so it's only running on the new host
[16:48] <RoyK> patdk-wk: yes, that's why HA VMs can be set to run on two machines in parallel
[16:48] <RoyK> otherwise, you'll need to boot it up on host B
[16:48] <patdk-wk> oh, hmm, I know what your talking about now
[16:48] <patdk-wk> heh?
[16:49] <patdk-wk> wait
[16:49] <patdk-wk> your not using shared storage for the vm's?
[16:49] <RoyK> not for their memory
[16:49] <patdk-wk> dunno, my vmware HA clusters don't share memory across systems, probably cause I don't need that
[16:50] <RoyK> very few do
[16:50] <patdk-wk> your not talking HA
[16:50] <patdk-wk> your talking FT
[16:50] <RoyK> and then, that wasn't what I was asking about for Xen
[16:50] <patdk-wk> fault tolerance
[16:50] <RoyK> I sometimes get lost in those terms
[16:50] <RoyK> they are used for different uses across vendors :P
[16:50] <patdk-wk> no support for that in xen, as far as I know, but I dunno what xen 4.x does
[16:51] <patdk-wk> ya, HA just means the vm is running, on some host
[16:51] <patdk-wk> FT means, no reboots to do HA
[16:51] <RoyK> again, that wasn't the question, it was just a divert
[16:51] <RoyK> so can I do HA with Xen without scripting it myself?
[16:51]  * patdk-wk has no idea what the question is, due to lack of terms :)
[16:51] <patdk-wk> find someone else that scripted it? or buy the xen enterprise?
[16:52] <RoyK> patdk-wk: I know xenserver etc - I was merely asking......
[16:52] <patdk-wk> I don't think it's that hard, to use pacemaker to do that
[16:52] <RoyK> should work..
[16:53]  * RoyK can't beleive it's 2011 and Hyper-V still doesn't support memory overcommit
[16:53] <patdk-wk> but, they are working on it :)
[16:53] <patdk-wk> kvm doesn't
[16:53] <RoyK> yeah, wait till 2020 or so
[16:53] <Roasted> Question - can anybody make out what's going on here? This is my syslog from my DHCP/LTSP server. My client (connected via cat5) is hanging on boot. Here's the syslog: http://pastebin.com/Cg7qEMFv
[16:53] <RoyK> kvm does
[16:54] <patdk-wk> when did it get it?
[16:54] <patdk-wk> I know it had some mem dedup
[16:54] <RoyK> some time back
[16:54] <patdk-wk> but that is hardly overcommit
[16:54] <RoyK> hm... seems I might be wrong (again)
[16:55] <RoyK> I was _told_ it supported it, but it may seem there are some other methods used
[16:55] <RoyK> or ... http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Virtualization/sect-Virtualization-Tips_and_tricks-Overcommitting_with_KVM.html
[16:55] <patdk-wk> I will have to look some
[16:56] <patdk-wk> I'm reading the page from the kvm website on it
[16:56] <patdk-wk> I know kvm did kindof support memory ballooning
[16:56] <patdk-wk> but that doesn't work on high memory pressures, like xen/vmware do
[16:56] <RoyK> ok
[16:56] <patdk-wk> ballooning is like sharing the ram between vm's, depending on need
[16:57] <patdk-wk> but fails if they all need ram
[16:57] <RoyK> ic
[16:57] <kirkland> Daviey: RoAkSoAx: okay, i have a simple way to reproduce this outside of orchestra
[16:57] <RoyK> and that works better on xen?
[16:57] <kirkland> sudo apt-get -y etckeeper && sudo env -i PATH=$PATH TERM=$TERM apt-get install -y juju
[16:58] <kirkland> Daviey: RoAkSoAx: basically, just ensure that LANG is unset
[16:58] <patdk-wk> xen supports overcommit, it won't completely fail, it will get painfully slow though
[16:58] <patdk-wk> where vmware will maintain it's speed better
[16:59] <patdk-wk> I'm like a year out of date on kvm/xen though
[16:59] <RoyK> I guess it's time to test xen again, nevertheless
[17:01] <patdk-wk> ah, there it is
[17:01] <patdk-wk> looks like kvm does support overcommit now
[17:01] <ClaesBas> patdk-wk I had to wipe the disks in the end too as you proposed! (could be very tired for less)
[17:01] <patdk-wk> it's just so unbelievably painfully slow, it's not practical :)
[17:02] <patdk-wk> ClaesBas, ya, must of gotten setup with gpt partitions then
[17:02] <patdk-wk> normally only needed for >2tb drives, but I tent to do it on >1tb myself
[17:02] <kirkland> sudo apt-get -y etckeeper && sudo env -u LANG apt-get install -y juju
[17:02] <kirkland> slightly simpler ^
[17:02] <RoyK> ClaesBas: why?
[17:05] <ClaesBas> gpt? I've only used Ubunt server  CD's (advanced) install... (I don't really understand what you mean with gpt)
[17:05] <RoyK> global partition table
[17:05] <patdk-wk> I'm not sure, but I think ubuntu uses gpt by default these days
[17:05] <patdk-wk> gpt instead of mbr
[17:06] <RoyK> patdk-wk: not on small drives
[17:06] <ClaesBas> My disks is 750G
[17:06] <ClaesBas> 4x750G
[17:07] <RoyK> all data drives?
[17:07] <RoyK> for raid?
[17:07] <RoyK> if so, don't partition them, just give them to md as they are
[17:07] <RoyK> simplifies extending later
[17:08] <RoyK> as in when you want to swap them for 2TB drives to add more to the array without moving the data
[17:08] <Doonz> patdk-lapfigured out my problem... hdd dying/died
[17:08] <patdk-wk> fun
[17:08] <RoyK> if you partition them and try to replace them with larger ones, you'll have a small ordeal to go through to fix the partition tables
[17:09] <ClaesBas> One more stupid thing is that if you setup a "big" software raid the with the installation CD, it tells you to reboot and the grub has "quite", what happens with your not yet fully initiated raid then....
[17:11] <RoyK> ClaesBas: it'll initiate in the background
[17:11] <RoyK> ClaesBas: it'll even continue over reboots
[17:11] <RoyK> not start over - continue
[17:11] <RoyK> 'cept you get an unexpected reboot - that may break it
[17:11] <RoyK> unexpected as in someone pulled the power cord
[17:12] <patdk-wk> or best case, it just starts the resync over, unless bitmap is turned on
[17:12] <ClaesBas> You get a blank screen after the reboot....
[17:12] <RoyK> ClaesBas: are you using a separate drive for the root?
[17:12] <RoyK> ClaesBas: if not, I'd _strongly_ suggest you do
[17:12] <RoyK> get some old 10GB ATA drive or something
[17:13] <patdk-wk> or you will need to setup partitions
[17:13] <RoyK> then use the big drives for data
[17:13] <RoyK> patdk-wk: bingo
[17:13] <ClaesBas> No, but an own raid1 over 4 disks and no lvm for that partion
[17:13]  * patdk-wk is a fan of 80-160gig drives for boot
[17:13] <patdk-wk> raid1? or raid10?
[17:13] <RoyK> ClaesBas: please, beleive me on this :)
[17:13] <RoyK> use separate _drives_
[17:14] <ClaesBas> A stupid 4 disk raid1
[17:14] <RoyK> it makes life a lot easier when you want to extend that disk set
[17:14] <patdk-wk> heh, you will defently need to use partitions, and leave a good 1-2mb free at the start
[17:14] <RoyK> ClaesBas: even a couple of USB sticks may do the job
[17:14] <patdk-wk> I would just use a boot drive though, or boot usb stick
[17:14] <ClaesBas> RoyK But if your "server" is a machine with only 4 disks (I payed for it)!
[17:14] <RoyK> ClaesBas: then an usb stick
[17:15] <RoyK> no need for a terabyte for the root
[17:15] <patdk-wk> you only need boot/mbr on the usb sticks
[17:15] <RoyK> 8GB should do for most
[17:15] <patdk-wk> you don't even need root on it
[17:15] <Daviey> zul: how did you get on with bug 871278
[17:15] <Daviey> ?
[17:15] <RoyK> patdk-wk: it helps for setting up things
[17:15] <patdk-wk> royk, ya helps, but not needed
[17:15] <patdk-wk> I do my laptop that way, for security
[17:16] <patdk-wk> remove usb stick, no booting the encrypted laptop
[17:16] <RoAkSoAx> kirkland: http://me.roaksoax.com/etckeeper.png
[17:16] <RoyK> patdk-wk: plain raid5 md on the disks, simple root fs on the plug
[17:16] <RoyK> KISS
[17:16] <kirkland> RoAkSoAx: yep, cool
[17:17] <Daviey> jamespage: still around?
[17:17]  * RoyK has an old 60GB ATA drive for the root in his home server
[17:17] <patdk-wk> yuk, ata
[17:17] <RoyK> well, it works just as well as sata
[17:17] <zul> Daviey: sruable
[17:17] <patdk-wk> except I don't own a motherboard with ata on it :)
[17:17] <RoyK> a bit slower, but hell, I don't write data to the root nayway
[17:17] <Daviey> zul: yes, but how far through fixing it are you? :)
[17:18] <patdk-wk> I do have like 80 250gig wd ata drives, if you want me to ship them :)
[17:18] <zul> Daviey: working on it
[17:18] <RoAkSoAx> kirkland: so where's the fix
[17:18] <RoyK> patdk-wk: I have a dozen or so 500GB ATA drives - works well for making coasters
[17:19] <RoyK> my family and friends tend to like shiny costers :D
[17:20] <kirkland> RoAkSoAx: http://paste.ubuntu.com/706853/
[17:20] <kirkland> Daviey: RoAkSoAx: FYI, etckeeper fix works like a champ
[17:20] <RoyK> !etckeeper
[17:21] <Daviey> kirkland: enough to state your name against it in an SRU?
[17:22] <RoAkSoAx> kirkland: looks good to me gonna try to setup a local repo and get it tested
[17:23] <RoAkSoAx> SpamapS: what was the name of the tool you use to create repos?
[17:24] <kirkland> RoAkSoAx: reprepo?
[17:24] <RoAkSoAx> local repos
[17:24] <RoAkSoAx> kirkland: ahh yes couldn't remember it . thanks
[17:24] <zul> why default it to en_US?
[17:28] <kirkland> zul: the critical part is the UTF8
[17:28] <kirkland> zul: defaulting to en_US.UTF8 because that's what ubuntu defaults to
[17:28] <kirkland> zul: without the UTF8 bit, etckeeper (actually, bzr) chokes on filenames that contain UTF8 characters
[17:29] <kirkland> zul: our /etc has some of those (most obviously in the /etc/ssl certificates, but could be anywhere)
[17:29] <jamespage> Daviey: a bit
[17:29] <kirkland> zul: and if LANG is already set in your env, then that'll override
[17:30] <zul> k
[17:30] <SpamapS> kirkland: LANG=C didn't work ?
[17:31] <kirkland> SpamapS: let me test with that ...
[17:31] <SpamapS> that would be the least opinionated ..
[17:31] <kirkland> SpamapS: does C imply UTF8?
[17:31] <kirkland> SpamapS: i'm firing up a new instance now
[17:31] <SpamapS> C implies byte ordering and no encoding
[17:31] <SpamapS> it will never mangle anything
[17:32] <SpamapS> and really, bzr should be defaulting to C if there's none set
[17:34] <aibo> hi, is it possible to install imagemagick without X?
[17:34] <zul> Daviey: whoops
[17:34] <RoyK> aibo: shouldn't be a problem - it might install some x libs, but the command line tools should work
[17:35] <RoyK> aibo: and the x libs won't hurt unless you're _really_ low on disk space
[17:35] <aibo> it's VPS, i have 4.5 gigs free
[17:35] <RoyK> not a problem
[17:36] <RoyK> doesn't seem to be a -nox package for it, so just try, and apt-get will tell you how much it'll use
[17:36] <RoyK> _before_ installing it
[17:37] <aibo> okay
[17:38] <RoyK> that's the fun serving VPSes - you charge close to nothing for 4GB and then when the customer wants another 4GB (retail cost ~zero) you add another $10 a month :P
[17:39] <kirkland> SpamapS: no-go with LANG=C, same breakage
[17:39] <SpamapS> kirkland: very interesting!
[17:39] <kirkland> http://paste.ubuntu.com/706867/
[17:39] <kirkland> SpamapS: ^
[17:40] <RoyK> kirkland: what distro?
[17:40] <aibo> RoyK, yep, that's exactly what happens
[17:40] <RoyK> erm
[17:40] <RoyK> versjon
[17:40] <kirkland> RoyK: tamulinux
[17:40] <RoyK> heh
[17:40] <kirkland> wtf?  :-)  Ubuntu
[17:40] <kirkland> RoyK: 11.10 :-)
[17:41] <SpamapS> kirkland: well that does sound like a bug in bzrlib, but the "default to en_US" seems your only recourse
[17:41] <RoyK> kirkland: I just setup 10.11 with etckeeper without issues....
[17:41] <SpamapS> RoyK: it happens only when installing on a non-interactive shell
[17:41] <RoyK> oh
[17:41] <SpamapS> such as via cloud-init on first boot
[17:42] <RoyK> ic
[17:42] <SpamapS> kirkland: another option is to try and pick up the machine's locale.
[17:43] <kirkland> SpamapS: sounds overly complex;  this "fix" should handle 99.99% of the use cases
[17:43] <RoyK> SpamapS: shouldn't C work for them all?
[17:43] <kirkland> SpamapS: we're talking about etckeeper here
[17:43] <kirkland> SpamapS: and we're talking about the file encoding of /etc
[17:43] <kirkland> SpamapS: nothing else
[17:44] <kirkland> SpamapS: it affects nothing else in the distro, AFAICT
[17:44] <SpamapS> RoyK: thats what I was thinking, but bzrlib seems to be doing something wrong with C
[17:44] <SpamapS> kirkland: yeah I'm not worried about it breaking stuff.. just wondering if there's a simple way to pick up the right LANG
[17:45] <SpamapS> kirkland: that said.. if bzr requires UTF-8 specifically.. then yeah.. give it what it wans.
[17:45] <SpamapS> wants even
[17:45]  * kirkland reboots
[17:47] <huslage> howdy. I'm trying to set up this T-Mobile Rocket 4G USB modem. I've gotten usb_modeswitch to do the right thing and present me with a proper modem. Now I'm stumped about how to get the thing online. It has a usbnet interface as well as the standard AT-commands. I'm not sure how they relate (if at all).
[17:48] <huslage> the modem is a ZTE MF591
[17:48] <kirkland> Daviey: SpamapS: RoAkSoAx: I'm uploading the etckeeper fix
[17:48] <kirkland> Daviey: do you want it in -proposed?
[17:48] <kirkland> Daviey: or oneiric?
[17:49] <Daviey> kirkland: release pocket is still ok.
[17:50] <kirkland> Daviey: uploaded, fixes 2 different-but-similar etckeeper bugs, including our orchestra one
[17:51] <RoAkSoAx> kirkland: ok cool
[17:51] <kirkland> Daviey: RoAkSoAx: SpamapS: I'll leave it to the Ubuntu Server Team to decide if they want etckeeper recommended or not on Orchestra Clients
[17:51] <RoAkSoAx> ill wait for that one to be released them
[17:51] <RoAkSoAx> kirkland: Daviey I think we should keep it
[17:52] <kirkland> I'd like to see it -- but I understand if you guys take a conservative stance at this point
[17:52] <RoAkSoAx> will be a good way to find more errors for 12.04 and besides it is still in universe
[17:52] <kirkland> personally, I'd be happy to SRU fixes, as necessary
[17:52] <RoAkSoAx> +1
[17:52] <Daviey> RoAkSoAx: you think the benefit makes it worth while having a potential busted juju+orchestra deployment.
[17:52] <kirkland> but that's up to you blokes (/me puts in language that daviey understands :-)
[17:52] <Daviey> where busted = useless?
[17:52] <RoAkSoAx> Daviey: right now it is busted. The fix uploaded by kirkland makes it useful again
[17:53]  * kirkland finds lunch before his meeting in 7 minutes
[17:53] <Daviey> RoAkSoAx: how is it busted?
[17:53] <Daviey> The fact this pckage is in universe is not relevant, it's essentially supported.
[17:54] <RoAkSoAx> Daviey: busted in a way that it fails to create the necessary stuff, but from the testing I did today, it doesn't prevent juju from deploying a machine
[17:54] <RoAkSoAx> Daviey: in fact, the machine is up and running normally
[17:55] <RoAkSoAx> Daviey: the only issue is that etckeeper was unable to commit for the bug above
[17:55] <RoAkSoAx> Daviey: whchi should be fixed now and we should not have any problems
[17:55] <Daviey> RoAkSoAx: i've already backed etckeeper out.
[17:55] <Daviey> I'm not sure we should reintroduce it at this stage.
[17:56] <RoAkSoAx> Daviey: oh ok, so then nope, if it has been removed as recommends, then we should probbaly not reintroduced it
[17:56] <RoAkSoAx> Daviey: i thought it still was
[17:56] <RoAkSoAx> cause the installation i just did pulled etckeeper
[17:57] <zul> Daviey: queued
[17:58] <Daviey> RoAkSoAx: yeah, i really didn't want juju busted for the minimal feature improvement TBH.
[17:58] <Daviey> RoAkSoAx: i only uploaded it an hour or so ago
[17:58] <RoAkSoAx> Daviey: ok
[17:58] <RoAkSoAx> Daviey: that's fine then
[18:15] <wip> i've been hacked... i am using ubuntu hardy LTS that still runs PHP 5.2.4 which have a security issue. how can i upgrade PHP on ubuntu hardy!?
[18:16] <wip> here's what is installed: http://paste.ubuntu.com/706895/
[18:16] <wip> i think it's not a good idea to make a dist-upgrade?!
[18:19] <wip> or should i use a PPA thsat provide PHP 5.3?
[18:20] <patdk-wk> hmm, that php shouldn't have a security issue, if you update it
[18:20] <patdk-wk> what exact security issue?
[18:22] <wip> patdk-wk, here's my source.lists is everything good? http://paste.ubuntu.com/706899/
[18:22]  * wip finding information about the hack
[18:22] <patdk-wk> well, you have the most current php
[18:23] <patdk-wk> so either it was a issue in the php script, or a new issue that hasn't been fixed yet
[18:24] <kirkland> SpamapS: do you have time to help negronjl push the cloudfoundry charms to the "best" place in LP?
[18:24] <kirkland> SpamapS: I'm thinking somewhere like lp:~cloudfoundry/charm/oneiric/cloudfoundry-*
[18:24] <kirkland> SpamapS: negronjl said he was getting some permission denied errors
[18:24] <kirkland> SpamapS: I suspect this is something you can help him nail very quickly
[18:24] <wip> patdk-wk, here's what happened to me... http://www.jorgepablos.com/2011/03/watch-out-php-5-2-vulnerability-google-results-finditnow-osa-pl/
[18:25] <wip> patdk-wk, the article say to upgrade to PHP 5.3
[18:25] <patdk-wk> that isn't a hack
[18:25] <patdk-wk> that is what they did AFTER they hacked
[18:26] <wip> right, sorry english is not my first language
[18:26] <patdk-wk> the php 5.3 fix, isn't so much a fix as, it breaks code that used older insecure default settings for php
[18:27] <patdk-wk> so while 5.3 would fix it, it would break those applications your using that caused the issue
[18:27] <patdk-wk> it's really the applications fault though
[18:27] <SpamapS> kirkland: ack
[18:27] <wip> i cannot check all the sites installed (many artists that doesn't know a thing about wordpress, dupral etc...) so basically i can't do nothing?
[18:28] <patdk-wk> oh heh, ya, all it takes is any out of date wordpress install, wordpressed has all kinds of issues
[18:28] <patdk-wk> I would recommend 5.3 there yes
[18:28] <patdk-wk> but your customers probably won't be happy about that
[18:29] <patdk-wk> cause that WILL break somethings for them
[18:29] <patdk-wk> how many, all depends
[18:29] <wip> patdk-wk: is there a way to install PHP 5.3 and be able to downgrade easily?
[18:30] <patdk-wk> installing via a .deb would be easy, just remove those .deb and reinstall the old .deb
[18:30] <patdk-wk> that goes for ppa/... too
[18:30] <patdk-wk> but I have never attempted to install 5.3 on 8.04 though, so that, I can't answer where to get a nice clean good source of php 5.3 for that
[18:30] <wip> first aptitude remove libapache2-mod-php5, php5, php5-mysql etc...
[18:31] <wip> then find a good source for php 5.3 dpkg -i php5.3
[18:31] <patdk-wk> if you install from ppa, to downgrade, remove the php 5.3 ppa, update, then remove/reinstall those :)
[18:31] <patdk-wk> well, to upgrade, you shouldn't need to remove
[18:31] <wip> patdk-wk, in ubuntu i never found a complete way to remove a PPA?!
[18:32] <zul> Daviey: stilla round?
[18:32] <Daviey> zul: yuppers
[18:32] <patdk-wk> comment it out, or delete the file, from /etc/apt/sources.list.d
[18:32] <patdk-wk> then you have to remove those packages from that ppa manually (I think, atleats how I do it)
[18:32] <zul> Daviey: i have the nova branch queued up with the iscsitarget fixes and smosers database pool fix
[18:32] <Daviey> zul: have a diff?
[18:32] <wip> patdk-wk: ok looking for a PPA for hardy (php 5.3)
[18:32] <zul> Daviey: should i upload to -proposed?
[18:33] <zul> Daviey: uhhhh gimme a sec
[18:33]  * RoyK just started kismet to log traffic in case his network goes down :P
[18:35] <Daviey> zul: I think this needs to be -proposed.  Also not certain we need to 0-day it.
[18:35] <zul> Daviey: http://pastebin.ubuntu.com/706909/
[18:35] <wip> or maybe if i am lucky i won't be hack again...
[18:36] <patdk-wk> wip, unlikely
[18:36] <patdk-wk> once it's hacked once, it gets onto many websites it was, and then other hackers do it to get their name/reputation out there also
[18:36] <sdasd>  Hi guys, im running the linux-image-2.6.32-29-scst (Debian 10.04) and I have issues inserting the arcmsr module, it states "FATAL: Error inserting arcmsr (/lib/modules/2.6.32-29-server/extra/arcmsr.ko): Invalid module format". I've downloaded the linux-headers-scst (which basically is the 2.6.32.29-server generic + scst). The output of the building of the module can be read here :
[18:36] <sdasd> http://pastebin.com/JCkBE3A2. Any Pointers/tips :)?
[18:37] <wip> so either i brake things by installing PHP 5.3 or i get hack... wow
[18:37] <wip> am i the only admin having this issue?
[18:37] <Daviey> zul: can you propose your fix to upstream trunk asap?
[18:37] <zul> Daviey: the iscsitarget one?
[18:37] <Daviey> yah
[18:38] <zul> Daviey: taken from it
[18:38] <Daviey> +Description: Fix iscsi target path issues.
[18:38] <Daviey> +Author: Chuck Short <zulcss@ubuntu.com>
[18:38] <Daviey> +Upstream: no
[18:38] <patdk-wk> wip, for your case yes
[18:38] <patdk-wk> cause your using mod_php
[18:38] <zul> Daviey: gah..
[18:38] <patdk-wk> so all it takes is one of your customers/sites to get hacked to infect the rest
[18:39] <wip> patdk-wk: even if i set the user / group of apache different from the directory of my sites
[18:39] <Daviey> zul: where did you get the Upstream tag from? :)
[18:39] <zul> my arse
[18:40] <patdk-wk> wip, guess that depends on pache mode, and I know your apache in 8.04 is too old
[18:40] <patdk-wk> that user/group thing is only for use when running cgi's, mod_* is not a cgi
[18:41] <wip> patdk-wk: thank you for your time! when you say "cause your using mod_php" what should i use?
[18:41] <patdk-wk> to keep users seperate? only two options, cgi :) or fastcgi, where fastcgi needs a php instance running per user
[18:42] <patdk-wk> there are patchs and hacks somewhat to make it easier
[18:42] <patdk-wk> php-ftm? or whatever is suppost to help
[18:42] <patdk-wk> but that is only in php 5.3 I think
[18:42] <wip> patdk-wk, no i mean - what are you using if you don't use mod_php?
[18:43] <patdk-wk> fastcgi, it's much faster than mod_php
[18:43] <patdk-wk> and a lot less resource intensive
[18:43] <patdk-wk> isn't that what I just said though?
[18:44] <wip> patdk-wk, sorry yes! are you on ubuntu server hardy?
[18:44] <wip> patdk-wk, what version of PHP
[18:44] <patdk-wk> no, as hardy life is almost over
[18:45] <wip> that is quite funny, i installed hardy like 2 years ago because LTS and now people say the life of hardy is almost over
[18:45] <patdk-wk> lts is good for 5 years
[18:45] <wip> wow, a server is suppose to stay up and running for decades ;)
[18:45] <patdk-wk> it has 18 months left?
[18:45] <patdk-wk> hardy came out in 2008
[18:46] <patdk-wk> just cause you just installed it doesn't mean it lasts longer :)
[18:52] <wip> patdk-wk, what will happen after LTS, PHP will not get updated?
[18:52] <patdk-wk> nothing will get security patchs, or updates
[18:52] <wip> anyone ever done apt-get dist-upgrade from hardy?
[18:52] <patdk-wk> you would need to upgrade the server
[18:52] <wip> this is dangerous i guess?
[18:53] <patdk-wk> you should be doing dist-upgrades all the time
[18:53] <patdk-wk> or else your kernel will never get security updates
[18:53] <patdk-wk> you mean, do-release-upgrade
[18:53] <patdk-wk> to upgrade to 10.04 LTS
[18:53] <wip> oups sorry yes do-release-upgrade
[18:54]  * patdk-wk has been known to go directly from 7.04 -> 10.04 doing that, without issue
[18:54] <patdk-wk> defently not recommended or supported though
[18:54] <thesheff17> wip: I have
[18:54] <thesheff17> it worked fine
[18:55]  * wip is scared to death
[18:55] <thesheff17> wip, backup your data
[18:55] <wip> thesheff17, yes i have back-up, but this server have many stuff installed, it's a mail server, http server, sftp server etc...
[18:56] <patdk-wk> wip, how cleanly the upgrade goes, depends mainly on how you customized the configs, and how you modified the system
[18:56] <thesheff17> well there is a number of things you can do
[18:57] <patdk-wk> it should work pretty nicely
[18:57] <thesheff17> take an image of the whole thing
[18:57] <thesheff17> or drive
[18:57] <patdk-wk> but it all depends on what is going on in that system :)
[18:57] <thesheff17> or clone the drive to another one test upgrade
[18:57]  * patdk-wk clones systems to vm instances, upgrade that many times, then upgrades the real system
[18:58] <wip> by the way 10.04 LTS have what version of PHP?
[18:58] <patdk-wk> 5.3
[18:58] <jj995> any suggestions on a simple logger for a channel on ircd-hybrid?  just logging to a text file on the irc server is desired
[18:58] <patdk-wk> PHP 5.3.2-1ubuntu4.9 with Suhosin-Patch (cli)
[18:58] <wip> 5.3.6 is released awhile ago... oh well
[18:59] <patdk-wk> wip, you have no idea how security updates work, do you
[18:59] <patdk-wk> or any distribution
[18:59] <wip> oups 5.3.8 is release - ok yes the backport thing
[18:59] <patdk-wk> wip would make a good security auditor
[19:00] <wip> it's a bit misleading calling php 5.3.2 when it's patched (backport)
[19:00] <wip> that's all
[19:00] <patdk-wk> but it doesn't have the features 5.3.8 has, so that would be misleading also
[19:07] <sdasd> argh, stop being invaldid module format!
[19:20] <RoAkSoAx> Daviey: im gonna upload fixes for lp #872926 nad lp #872892
[19:21] <RoAkSoAx> Daviey: and take care of adam's fix
[19:29] <RoAkSoAx> adam_g: ping
[19:31] <RoAkSoAx> err adam not her ethis week
[19:31] <Patriot7912> How do you make samba a member of a domain?
[19:33] <RoAkSoAx> smoser: ping
[19:36] <smoser> RoAkSoAx, here.
[19:38] <RoAkSoAx> smoser: did you re-review https://code.launchpad.net/~gandelman-a/ubuntu/oneiric/cobbler/lp850880-850866 ?
[19:38] <RoAkSoAx> smoser: did it satifsy all your concerns?
[19:42] <smoser> RoAkSoAx, reading
[19:42] <smoser> a couple things i would comment
[19:42] <smoser> * debug() should go to stderr, not stdout
[19:43] <smoser> i guess thats it
[19:44] <RoAkSoAx> smoser: cool I guess I can fix that myself and get it uploaded
[19:44] <RoAkSoAx> thanks
[19:47] <jfb_h20> what's the way to 'disable' the left slot in a pagetemplate or BrowserView?
[19:47] <jfb_h20> disregard, wrong irc
[19:51] <kirkland> Daviey: fyi, the etckeeper upload was rejected
[19:51] <kirkland> Daviey: please advise
[19:52]  * RoAkSoAx will be back later
[19:53] <skrite> anyone recommend some mysql benchmarking software? i need to find what settings and engines will work best in our applications
[19:58] <RoyK> skrite: dunno - there are several - but the real thing is to test it
[19:59] <RoyK> skrite: I'd recommend using postgres, though - less hassle with broken parts and better support for most of the stuff you want (or don't want)
[19:59] <skrite> my main idea now is just writing some scripts that will hammer it and measure the times, but was looking for a more professional (read knowledgeable way)
[20:00] <RoyK> any particular reason for using mysql?
[20:00] <RoyK> mysql is generally a hack, not a good DBMS
[20:00] <RoyK> but then, if you're using stuff like wordpress, you're stuck with it
[20:01] <Daviey> kirkland: infinity rejected it because it is a shipped package, which doesn't warrant a respin.  He felt that it should be a tradional SRU.
[20:01] <kirkland> Daviey: okay
[20:02] <Daviey> so reupload to -proposed
[20:02] <kirkland> Daviey: i'll reupload to -proposed
[20:02] <Daviey> cool
[20:02] <kirkland> Daviey: does it need the whole SRU blob in the bug reports, etc.?
[20:03] <kirkland> Daviey: uploaded
[20:07] <skrite> RoyK, well, i would be open to another database, but have so much code written and in place that i kinda cant
[20:08] <skrite> RoyK:  what db do you like?
[20:08] <zul> Daviey: ill do the nova upload tomorrow
[20:10] <Daviey> zul: cool, that doesn't need to be 0-day IMO.. we'll do the normal 7 day SRU process.
[20:10] <Daviey> kirkland: probably should.
[20:10] <kirkland> Daviey: arg
[20:10] <Daviey> kirkland: not really my call, you want to check that against the sru team :)
[20:10] <kirkland> RoAkSoAx: any chance you can do that?
[20:11] <kirkland> Daviey: i really have a million other things i need to work on today
[20:11] <RoyK> skrite: I use all sorts of DBMSes, from sqlite to oracle, but I really fancy postgres
[20:11] <Daviey> kirkland: well i don't think it matters if it is in the archive for release day?
[20:11]  * Daviey goes AWOL
[20:11] <kirkland> Daviey: that's kinda what i was thinking
[20:11] <skrite> RoyK: , yeah if i could pick today, that is what i would use
[20:12] <RoyK> postgesql
[20:12] <RoyK> postgresql, even
[20:12] <zul> Daviey: which segway into my next question should we have a schedule for os sru
[20:14] <allowoverride> yawns
[20:15] <RoAkSoAx> kirkland: sure
[20:15] <RoAkSoAx> kirkland: will do as soon as I get back
[20:15] <RoAkSoAx> kirkland: could you either attache a debdiff to the bug report or send me the pastbinit
[20:15] <RoAkSoAx> again
[20:16]  * RoAkSoAx will be back in 45 mins
[20:31] <adam_g> RoAkSoAx: im checking irc periodically. if you /msg me ill respond eventually
[20:36] <adam_g> RoAkSoAx: just pushed to lp:~gandelman-a/ubuntu/oneiric/cobbler/lp850880-850866/ to address scotts comments on debug()
[20:49]  * RoyK just pissed off some people from #zimbra, and it feels good
[20:50] <allowoverride> heheheh
[20:51] <allowoverride> whast that?
[20:52] <RoyK> a rather good email or groupware server
[20:52] <allowoverride> ohhh
[20:52] <allowoverride> ok
[20:52] <RoyK> built on postfix and a bunch of other software
[20:52] <RoyK> but they don't want to explain what they do
[20:58] <allowoverride> they dont wan to explain?
[20:59] <allowoverride> how did you come to that final theory
[20:59] <allowoverride> ?
[21:27] <RoAkSoAx> adam_g: cool thanks
[21:27] <RoAkSoAx> adam_g: i'll prepare SRU then
[21:57] <drahst> I've got ubuntu server up and running, now I'm starting to install virtuals, but I don't see an option for vmbuilder or virt-install to use an img file that I created (pre-allocated disk space file) to install the os on vs. default. is there any way to make virtbuilder use a specific img file? I created a pre-allocated file to help with i/o, I'd like to have the vm that is created to
[21:57] <drahst> use it
[22:02] <RoAkSoAx> Daviey: ping
[22:03] <gholms> soren: What would it take to make uvirtbot ignore eucabot in #eucalyptus so they can't accidentally trigger one another?
[22:08] <soren> gholms: I had no idea it was even in there.
[22:08] <gholms> It isn't yet.
[22:08] <gholms> But it will be.
[22:08] <soren> Says who?
[22:08] <gholms> Me.  I'm adding it.
[22:08] <soren> And how do you expect to do that?
[22:08] <gholms> Running a supybot instance on projects.eucalyptus.com
[22:09] <soren> Err..
[22:09] <soren> There can be only one uvirtbot.
[22:09] <gholms> I'm not looking to replace it.
[22:09]  * soren is confused
[22:10] <soren> uvirtbot isn't in #eucalyptus, but it will be, because you'll run another bot on p.e.c?
[22:10] <soren> uvirtbot: Thanks, that's really helpful. I appreciate it.
[22:10] <geekbri> so im trying to build this .deb file of php.  when i run dpkg-buildpackage -rfakeroot it actually runs my configure which i set with dh_override_auto_configure, and it seems to run a make as well, but it doesn't run make install.  Does this not happen automatically?  Do i need to specify in the debian/rules to run a make install?
[22:10] <gholms> No, eucabot will be in #eucalyptus, so uvirtbot, which is already there, should ignore it.
[22:10] <soren> gholms: Ah.
[22:11] <soren> gholms: What will eucabot do?
[22:11] <gholms> Bug listing with meetbot
[22:11] <gholms> s/with/and/
[22:11] <soren> gholms: If it does bug listing, why do you need uvirtbot?
[22:11] <gholms> ¯\(°_o)/¯
[22:12] <gholms> It's already there, which is why I'm asking.
[22:12]  * soren is confused again
[22:12] <gholms> Maybe when eucabot is completely set up it won't be necessary?
[22:13] <soren> I don't think I understand the problem :)
[22:13] <soren> uvirtbot's primary job is to announce bugs and look up bugs.
[22:14] <gholms> eucabot is a copy of an internal bot that, among other things, responds with bug links and summaries when someone runs ".lp 12345".
[22:14] <soren> You're concerned that a coming bot will say things that'll trigger this behaviour.
[22:14] <gholms> If uvirtbot sees that it will then respond.
[22:14] <gholms> Yeah
[22:14] <soren> ...but if this other bot can say such things, why do you still need uvirtbot?
[22:14] <gholms> I doubt we will.
[22:14] <soren> I have no particular desire to keep it in there. If it's in your way, I'll gladly remove it.
[22:14]  * gholms runs off to ask obino
[22:15] <soren> You guys work it out. You can just let me know when you want to have it removed.
[22:15] <soren> I need to sleep.
[22:15]  * soren heads bedwards
[22:16] <RoAkSoAx> Daviey: orchestra1.1 which has the removal of etckeeper has been rejected btw
[22:17] <gholms> soren: Ok; thanks
[22:26] <Daviey> RoAkSoAx: no, i uploaded to both -proposed and release pocket
[22:27] <Daviey> the -proposed was rejected at my request
[22:27] <Daviey> and the release pocket one is now published
[22:28] <RoAkSoAx> Daviey: ohh ok
[22:29] <RoAkSoAx> Daviey: alright gonna test fixes for lp #850880 and then I'll sru to -proposed
[22:30] <Daviey> RoAkSoAx: great, i don't care if that is there at release time fwiw.  I think that can be a normal SRU, but importiant for the future.
[22:31] <RoAkSoAx> Daviey: i'm closing couple other bugs with the SRU
[22:31] <Daviey> RoAkSoAx: if you wanted to add the love to the bug, but still keep it a sponsored package for adam_g - it might be helpful.
[22:31] <RoAkSoAx> bug #872926 and bug #872892
[22:32] <RoAkSoAx> Daviey: i couild 0-day sru those two, and then close adam's later
[22:32] <Daviey> ooo, more fixes \o/
[22:32] <RoAkSoAx> when I can test it throughtly
[22:33] <Daviey> does it warrant 0-day?
[22:33] <RoAkSoAx> Daviey: well to save uploads 1 woyuld be enough
[22:35] <Daviey> RoAkSoAx: I think it's a god compromise to have multiple fixes, following the otherwise normal SRU process.
[22:35] <Daviey> We aren't rushing in fices into -updates, but closing multiple bugs.
[22:35] <Daviey> That is the approach i would recommend
[22:35] <RoAkSoAx> Daviey: yeah that's what I'm saying that I'd rather do a single upload
[22:35] <Daviey> Yeah, i think that is a good compromise.. do that :)