[00:39] <zzz2009> I am setting up a server, the system has 4 500Gb drives, I am thinking of setting them up as a raid 0 which should give me about 2Tb of disk.
[00:39] <zzz2009> I was thinking of partioning: 100mB ext2 - /boot, 100Gb ext4 - /root, 100Gb ext4 - /var, 1Tb ext4 -/home
[00:39] <zzz2009> Is his a reasonable setup, for a mail/samba/apache server?
[00:39] <zzz2009> There are no local users.
[00:39] <zzz2009> possible global shared space,
[00:39] <debio264> I just installed slapd on Karmic and I have no idea what the admin password is
[00:39] <debio264> dpkg-reconfigure slapd doesn't set it anymore, it would seem
[00:40] <yann2> zzz2009, no, it is not :)
[00:40] <debio264> so I'm not sure how I should procede
[00:40] <yann2> with the raid0, the first disk failure would make you lose all your data
[00:40] <yann2> use raid5 at least, raid10 if you care about performance
[00:41] <zzz2009> OK, but it gives me the largest sapce + best speed
[00:41] <zzz2009> sapce = space
[00:42] <yann2> yes
[00:43] <yann2> just be aware that you lose one disk, you lose everything, and disks tend to fail regularly ;)
[00:43] <zzz2009> raid5 gives me about 500-600Mb which leaves me wondering
[00:43] <yann2> ? no it will give you 1.5TB
[00:43] <zzz2009> how soon an upgrade will be needed?
[00:43] <yann2> why would you want 100GB for /root ?
[00:44] <zzz2009> I am missing something, when I tried to setup a raid 5 I wound up with about 550Gb,
[00:45] <yann2> with 4 disks in raid 5, you would lose one disk :)
[00:46] <zzz2009> sorry /roo should be just /, I was typing what i was thinking not what i meant.
[00:46] <zzz2009> /roo = /root , I need to learn to type
[00:46] <yann2> the rest seems fine, just dont put it in raid 0 ;)
[00:47] <yann2> I wouldnt use raid5 neither if the mailserver is going to be used a lot though
[00:47] <yann2> depends on your usage I guess
[00:47] <yann2> gotta go, good night!
[00:47] <zzz2009> thanks yann2
[00:50] <zzz2009> f not raid 5 what then
[00:51] <yann2> raid10
[00:51] <yann2> but you need hardware that supports it (not sure you can do a soft raid 10?) and you would have only 1TB all together
[00:53] <qman__> you can do soft raid 10
[00:59] <qman__> but raid 5 performs pretty well, only really high traffic servers will need more performance than that
[00:59] <qman__> it depends entirely on the level of activity you expect
[00:59] <qman__> number of users, number of emails, files on the samba share, etc...
[01:07] <ninnypants> I had wemin 1.45 installed on my server using port 10000 and I unstalled it but when I try to install 1.49 it says that port 10000 is already in use why  is this?
[01:28] <twb> RAID5 has better read and worse write speed than RAID10, I heard.
[01:28] <twb> ninnypants: we don't support webmin here, sorry.
[01:28] <twb> ninnypants: it would be reasonable to assume that something is already using that port.  ss or netstat will tell you what it is.
[01:29] <JanC> if you want remote admin over HTTP for your site, install AjaxTerm behind SSL with server + client keys  ;-)
[01:36] <twb> javascript isn't http, grumble
[01:37] <twb> Do those web-based terminal emulators just do a normal socks5 tunnel over 443?
[01:38] <twb> (Or whatever; I've yet to find a box that I couldn't do ssh/putty -oPort=443 twb@foo.edu from.
[01:39] <twb> I'm forced to deploy webmin at work because of our retarded business model, where the customer pretends they can administer their own server.  AFAICT all they really need is a way to add and remove users, which could surely be done better by configuring libpam-ldap correctly or something.
[01:40] <twb> They also make me install ubuntu-desktop on servers, because "they're used to Microsoft SBS, so we need a GUI on the server in case they connect a monitor to it"
[01:40] <twb> So these boxes are sitting here starting mDNS and shit like that when they boot, sigh...
[01:41] <ninnypants> twb: the feedback that net stat gives is udp        0      0 *:10000                 *:*                                 13040/perl
[01:41] <JanC> twb: if you break things with something like AjaxTerm, at least you can be sure it's your own fault, webmin will do that without your help  ;)
[01:41] <twb> ninnypants: which says that a perl process (PID 13040) is using it.
[01:41] <twb> JanC: I *really* do not need to be convinced that webmin is utterly, utterly the wrong thing.
[01:41] <twb> If you want a giggle, run lintian on the deb sometime
[01:41] <twb> I think they manually build it with ar and tar
[01:42] <JanC> twb: in reality, checkinstall probably...?
[01:42] <twb> Shrug
[01:44] <JanC> anyway, webmin breaks Debian/Ubuntu, so don't use it
[01:46] <twb> webmin breaks other systems, too
[01:46] <JanC> I have no experience with that  ;)
[01:46] <twb> Are you a chef/puppet fan?  At one point, we considered writing a web UI for it, i.e. basically it looks like webmin, but it writes manifests.
[01:46] <twb> But we didn't have the budget for it
[01:52] <ninnypants> what is a good one to use then?
[01:53] <twb> There are no good web administration tools
[01:53] <twb> Officially Ubuntu recommends e-box, but I'm unimpressed with it
[01:57] <ninnypants> is there a decent one that handles mysql?
[01:58] <blak111> phpmyadmin?
[01:59] <StrangeCharm> using the karmic server RC, after attempting an install with encrypted lvm (and an unencrypted /boot in a normal partition), grub tells me 'error: no such disk'. what have i done wrong?
[02:00] <twb> "decent" and "mysql" don't belong in the same sentence
[02:00] <twb> StrangeCharm: boot a live CD and reinstall grub
[02:00] <twb> StrangeCharm: you can also ask #grub, but tell them you're using grub legacy (I assume 9.10 still doesn't use grub2).
[02:01] <StrangeCharm> twb, is there a live environment on the server disk?
[02:01] <StrangeCharm> actually, i seem to recall that 9.10 did use grub2
[02:01] <twb> I think there is; issue "rescue" at the boot: prompt.
[02:02] <twb> Personally I tend not to use d-i for rescue purposes.
[02:02] <StrangeCharm> twb, looks like grub2 is default: http://www.ubuntu.com/getubuntu/releasenotes/910overview#GRUB%202%20by%20default
[02:02] <twb> I have had the most luck with the CentOS 4 and 5 live CDs, since they have LVM and md RAID support pre-installed, as well as openssh-server (though it doesn't start by default).
[02:02] <twb> OK, good to know.
[02:03] <StrangeCharm> is there a command i can give grub to ask it to boot a cd, or should i just reboot?
[02:04] <twb> You need to rebot
[02:04] <twb> *reboot
[02:04] <twb> Personally I am completely fed up with grub
[02:04] <twb> Nowadays I deploy extlinux where I can
[02:05] <StrangeCharm> what's extlinux got over grub?
[02:06] <twb> StrangeCharm: it works in some md RAID failure modes that grub completely shits itself for
[02:06] <twb> Because extlinux doesn't write an internal table of disks -- it just bootstraps the disk it booted from.
[02:07] <twb> This is extremely useful when e.g. sda is dead and the BIOS "helpfully" renames sdb to sda, so that if the MBR on sdb tries to bootstrap (hd1), it will fail because there's only an (hd0) now.
[02:08] <ScottK> 9.10 uses grub2 for new installs, but not upgrades.
[02:09] <maxagaz> how to apt-delete the cache ?
[02:09] <StrangeCharm> this si a fresh install
[02:09] <StrangeCharm> *is
[02:12] <StrangeCharm> twb, what's a command for installing grub2?
[02:13] <smoser> kirkland, i put the euca2ools build up in my ppa now, but haven't tested it yet.
[02:14] <ScottK> StrangeCharm: If it's a fresh install of karmic, you have grub2.
[02:15] <StrangeCharm> ScottK, agreed, but it's not booting, so i want to make changes such that it will
[02:15] <ScottK> Right, but that's nothing to do with installing
[02:17] <StrangeCharm> ScottK, twb's instructions were to reinstall grub. do you recommend a different course of action?
[02:18] <ScottK> I see. I missed that.
[02:18] <ScottK> No, I don't have any specific recommendations as I haven't had to troublshoot it yet.
[02:23] <zul> smoser: go to bed :)
[02:31] <twb> StrangeCharm: you need to mount / and /boot (e.g. /mnt/target and /mnt/target/boot) and then say grub-install --root-directory /mnt/target
[02:31] <twb> If you're lucky, grub-install will work
[02:31] <twb> If not, you have to run grub by hand and piss about
[02:32] <StrangeCharm> though, that is made difficult by the fact that / requires lvm and dm-crypt, which you've said are not installed
[02:32] <paul_whipp> Is it sensible to use AWS/Ubuntu server to host websites?
[02:33] <twb> StrangeCharm: you'd need to anna-install them or whatever, then
[02:33] <twb> StrangeCharm: although d-i might be smart enough to do that on its own
[02:34] <StrangeCharm> anna-install? d-i?
[02:34] <twb> debian-installer (d-i) is the technology the Ubuntu server install CD is based on.
[02:34] <twb> anna-install is d-i's equivalent of apt-get install.
[02:55] <StrangeCharm> twb, i take it that you won't be able to give me info about fixing grub?
[02:56] <twb> StrangeCharm: what more information do you need?
[02:58] <StrangeCharm> well, i'm at a grub rescue prompt. i think i know why it isn't booting: it's looking at the wring disk. i want to tell it to load the grub config from the right disk, then try booting with that.
[02:59] <twb> Ask #grub
[02:59] <twb> I can't be arsed going through that now
[03:13] <erichammond> paul_whipp: Yes, Ubuntu is an excellent choice for running web sites on AWS/EC2.
[03:14] <erichammond> (or anywhere else for that matter)
[03:54] <smoser> good night all.
[04:34] <foo> hm, ubuntu desktop 8.10 is hanging during the load up screen. I'm trying to boot for disk. gah, *looks for newer disk*
[04:34] <foo> hm, can't seem to find one
[04:35] <foo> any tips?
[04:35] <ScottK> foo: 8.10 or 9.10?
[04:35] <foo> 8.10
[04:35] <foo> it's older
[04:35] <foo> I figured it'd still work, this is a p4 box
[04:35] <foo> (and I don't have something newer lying around, apparently)
[04:36] <JanC> you can download everything newer for free ;)
[04:38] <foo> I know, it's time :)
[04:38] <foo> but I might do that if I can't get this working
[04:38]  * foo tries an auditor cd
[04:40] <ScottK> foo: 8.10 is about your worst choice.  Use 8.04 (the LTS release) or 9.04.
[04:40] <foo> auditor seems to be working
[04:40]  * foo waits to run dd if=/dev/urandom of=/dev/hda
[04:40] <foo> that will completely wipe the drive, I hope
[04:47] <twb> Given a server with timezone data on it, how do I dump what *it* believes are the data for a particular timezone?
[04:47] <twb> Ah, zdump.  Thanks, google and wiki.debian.org!
[06:01] <edulacomadreja> hi everyone!, i need some help with heartbeat...
[06:01] <edulacomadreja> the linux-ha channel is dead or something
[08:08] <fahadsadah> Have an Ubuntu Server box, with remctl-client
[08:08] <fahadsadah> This box already has Kerberos configured.
[08:09] <fahadsadah> It does not have rDNS.
[08:09] <fahadsadah> To fix this, it is in a VPN.
[08:09] <fahadsadah> With the KDC.
[08:09] <fahadsadah> Which has rDNS.
[08:09] <fahadsadah> The KDCs VPN broke.
[08:10] <fahadsadah> So, I added to /etc/hosts on the client box, kdc.ip.address.here kerberos
[08:10] <fahadsadah> (so, basically, no rDNS for my box)
[08:10] <fahadsadah> remctl: GSS-API error initializing context: Unspecified GSS failure.  Minor code may provide more information, Server not found in Kerberos database
[08:14] <simplexio> fahadsadah: kerberos dosent know any server named "kerberos" , try edit hosts file and add full domain name
[08:14] <simplexio> not that i know its the problem, its my best quess based on information given
[08:14] <fahadsadah> The FQDN is in there too.
[08:15] <fahadsadah> Also, server not found in Kerberos database usually refers to the KDC not finding a machine account for my box.
[08:15] <fahadsadah> There is one, and I can successfully kinit/kadmin as it with /etc/krb5.keytab
[08:18] <simplexio> so something changes between configurations, which caues that kerberos server dosent regonize client anymore
[08:18] <fahadsadah> Thanks, I'll have to look into this further.
[08:29] <kaushal> hi
[08:29] <kaushal> I have questions about https://help.ubuntu.com/8.10/serverguide/C/samba-ldap.html
[08:29] <kaushal> can i seek help here ?
[08:30] <Lunar_Lamp> kaushal: it's probably best to just ask.
[08:30] <Lunar_Lamp> (though I'll say straight away that I am far from an LDAP expert, so probably won't be able to  help you)
[08:34] <kaushal> Lunar_Lamp: when i run smbldap-populate
[08:34] <kaushal> i get http://paste.ubuntu.com/302630/
[08:34] <kaushal> I have ldap server running on the same server as samba server
[08:35] <kaushal> ldapsearch -x -H ldaps://localhost returns fine for me
[08:40] <alvin> I can't find the link to the rss feeds that track changes in Hardy and Karmic. Is there anyone who has it?
[09:09] <maxagaz> how to display the hardware spec of my server ?
[09:09] <maxagaz> like the System Monitor tool in desktop ubuntu
[09:14] <alvin> maxagaz: $ sudo lshw
[09:59] <StrangeCharm> i think that grub is looking at the wrong disk to read its config (&c) from, from a live environment that can mount /boot, on the right disk, what changes should i make to get it to look at the right disk/partition?
[11:41] <StrangeCharm> after a fresh install of the karmic RC, with several, distinct encrypted disks, i'm encountring a problem decrypting them at boot: when prompted for the passphrase for the disk that contains /, all goes well, but when prompted for the phrase for md0_crypt, the message "* Reloading /etc/samba/smb.conf smdb only" appears on the same line, before I am able to completely enter the phrase. why is this happenning? how can I stop i
[11:41] <StrangeCharm> t?
[11:44] <simplexio> i think you can still insert pass
[11:47] <StrangeCharm> simplexio, typing the phrase afterwards seems to have no effect
[11:51] <StrangeCharm> that is, the system does not move on, it just keeps waiting
[11:53] <soren> ttx: bug 461829
[11:53] <soren> ttx: Do you know the effect of that?
[11:53] <soren> ttx: It's hard to set the importance when I don't understand the impact.
[11:54] <soren> (Doing triage for stuff from today as well, since we're cutting it close enough w.r.t. the release)
[11:55] <StrangeCharm> soren: are there a lot of bugs still being reported from the rc?
[11:55] <soren> StrangeCharm: Sure.
[11:56] <StrangeCharm> soren, would you say that the rc is gererally buggy, or bug-free?
[11:56] <soren> StrangeCharm: Buggy. Just like every other piece of software ever released.
[11:57] <soren> Ever.
[11:57] <simplexio> StrangeCharm: try hit ctrl+c, maybe it gives you question again
[11:57] <StrangeCharm> soren, i wasn't looking for an absolute statement. more: something relative, and useful
[11:57] <simplexio> those encrypted installations are little exotic installs
[11:58] <soren> StrangeCharm: I cannot give you anything relevant and useful in the form of a binary answer.
[11:58] <soren> StrangeCharm: Sorry.
[11:58] <soren> StrangeCharm: Some things are working quite well. Others not quite as well.
[11:59] <simplexio> StrangeCharm: you can get aroud it using key file on all other crypted partition exept root partition, just but files on root behind pass
[11:59] <soren> StrangeCharm: If I understood why you were asking, perhaps I could give a more useful response.
[11:59] <simplexio> off course then dont forget root partition passphrase
[12:00] <StrangeCharm> soren, what about the question stated thus: i am experiencing a problem. given my experience that ubuntu generally 'just works' should i expect that it's an actual bug, or that i'm doing something wrong?
[12:00] <simplexio> karmic for somereason couldnt mount all my partition on upgrade, so i think there is some work to do
[12:00] <soren> StrangeCharm: I don't know. What's the problem?
[12:00] <StrangeCharm> simplexio, putting keyfiles on the root partition sounds great, it'd be much more convenient only to have to enter one phrase. how do i do that?
[12:01] <simplexio> StrangeCharm: google "dmcrypt use keyfile" , i use passpharse only on home dir
[12:01] <zul> heylo
[12:01] <StrangeCharm> soren, at boot, for the second encrypted partition that's being mounted, after being prompted for the passphrase, another message comes up (related to smb) and i can't enter the passphrase
[12:02] <soren> StrangeCharm: "Can't enter" how?
[12:03] <StrangeCharm> soren, typing the phrase at that point hs no effect
[12:04] <soren> StrangeCharm: Are you sure? Perhaps it just doesn't echo anything. Try typing your password and hitting return, ignoring the fact that there is no feedback.
[12:04] <StrangeCharm> simplexio, are d-mcrypt keyfiles supported by the installer? i only recalled being able to choose 'passphrase' and 'random' for the key options.
[12:04] <elyezer> when trying to send an email from another computer to my email (postfix) server I get this message: THIS SERVER IS TO BE USED WITH AUTHENTICATION (#5.7.1)> #SMTP#
[12:04] <elyezer> in the return email with the error,what could be?
[12:04] <StrangeCharm> soren, how long after that point should i wait for something to happen?
[12:04] <elyezer> locally I can send and receive
[12:05] <soren> StrangeCharm: Not long.
[12:05] <soren> StrangeCharm: Wait a few seconds and see .
[12:06] <StrangeCharm> soren, nothing happens for a few minutes
[12:06] <soren> StrangeCharm: Then it's probably a bug.
[12:07] <StrangeCharm> soren, where do i report it?
[12:07] <simplexio> StrangeCharm: um.. you can do it by hand
[12:07] <soren> StrangeCharm: Launchpad. Probably against cryptsetup.
[12:07] <simplexio> StrangeCharm: lets see if i still have link to howto
[12:08] <soren> StrangeCharm: https://bugs.edge.launchpad.net/ubuntu/+source/cryptsetup/+filebug
[12:08] <soren> StrangeCharm: Oops, sorry.
[12:08] <kwork> is it safe to restart udev ?
[12:08] <soren> StrangeCharm: https://bugs.launchpad.net/ubuntu/+source/cryptsetup/+filebug
[12:08] <kwork> udevd for some reason takes 100% of cpu
[12:08] <kwork> i wonder is it safe to restart it from init ?
[12:09] <simplexio> StrangeCharm: what mountpoints you have crypt ?
[12:09] <simplexio> StrangeCharm: i have / and home on crypted partitions. so at start i just insert / pass, and then manually mount /home after it
[12:10] <simplexio> so my setup isnt automatic, but it works
[12:10] <StrangeCharm> simplexio, [ sawp, /, /tmp] on the disk i can enter the phrase for, [/var /var/log /var/mail and /home] on another disk, and [/data] on the disk on which i'm getting interrupted
[12:11] <StrangeCharm> simplexio, well, you could automate it with a small script
[12:12] <simplexio> StrangeCharm: i have 9.04, and it has "bug" that i dosent bring my raid5 setup up correctly so thats reason why i do it manually
[12:13] <simplexio> there is probably some fix, but no time or intres to fix it
[12:22] <simplexio>  
[12:22] <StrangeCharm> simplexio, is the raid encrypted?
[12:23] <simplexio> yep
[12:23] <StrangeCharm> simplexio, which bit isn't getting mounted right?
[12:29]  * soren lunches
[12:31] <StrangeCharm> are any of the encryption methods in the installer suitable for automatic decryption based on keyfiles?
[12:39] <StrangeCharm> is it possible to to move a mount point (like /var) to another disk, after intallation?
[12:46] <sommer> morning
[12:48] <bogeyd6> StrangeCharm yes it is
[12:49] <bogeyd6> StrangeCharm https://help.ubuntu.com/community/MoveMountpointHowto
[12:57] <StrangeCharm> how can i (indeed, can i) reconfigure an existing dm-crypt partition that is currently mounted with a passphrase, so that it is instead mounted with a keyfile?
[12:57] <brewmaste> I'm setting up a high availability cluster with 2 load balancers (1 being a failover) and 2 webservers. How many or what type of SSL certs do I need to buy to let people connect with https?
[12:58] <soren> kirkland: A while ago, you mentioned something in #ubuntu-devel about Richard Hughes had found a solution to "the encrypted-swap/hibernate" problem.. Do you have a bug reference for that? It sounds very interesting.
[12:58] <kirkland> soren: yeah, on sec...
[12:58] <soren> kirkland: Cool, thanks.
[12:59] <kirkland> soren: I don't want to over sell this ....
[12:59] <soren> too late :)
[12:59] <kirkland> soren: he simply patched gnome *not* to offer to hibernate, if encrypted swap was detected
[12:59] <kirkland> soren: http://bugs.freedesktop.org/show_bug.cgi?id=23196
[12:59] <soren> Ahah.
[13:00] <soren> did we merge that?
[13:00] <kirkland> soren: not for karmic, i don't think, pitti thought it was a little too invasive
[13:00] <kirkland> soren: we'll have it for lucid, of course
[13:00] <kirkland> soren: and as far as the *real* problem goes, I think kees or jdstrand or mdeslaur might take a look at getting hibernate working with encrypted swap for lucid, perhaps
[13:01] <kirkland> soren: i'm trying to lean on them about it, at least
[13:01] <soren> kirkland: I'm not sure how that would work?
[13:01] <kirkland> soren: there's a few propositions
[13:01] <kirkland> soren: currently, we're randomly generating a swap key every boot (at least in the default enc swap setup, if you enc home)
[13:01] <soren> kirkland: right.
[13:02] <kirkland> soren: we could opportunistically wrap that random key when you login, in PAM
[13:02] <kirkland> soren: and store the wrapped key in LUKS
[13:02] <kirkland> soren: we could do that for at least a couple of users
[13:02] <soren> I thought about that, but how would that work for resume?
[13:02] <kirkland> soren: allowing any one of the valid, logged in users this boot to resume from hibernate
[13:03] <kirkland> soren: we'd need to teach initramfs to obtain that password, and unwrap the LUKS contents
[13:03] <kirkland> soren: there will necessarily be a password prompt on resume
[13:03] <soren> kirkland: You don't know if there's anything to resume before you decrypt it, so you have to always ask for the passphrase..
[13:04] <kirkland> soren: hmm, on cold boot?
[13:04] <soren> Unless of course you store a bit on the root filesystem saying that there's stuff to be resumed.
[13:04] <soren> kirkland: Yes.
[13:04] <kirkland> soren: right, that's the presumption
[13:04] <soren> Ok.
[13:04] <foo> ubuntu server 7.04 is hanging on usbhid: v2.6:USB HID core driver, any ideas?
[13:04] <foo> (during an install)
[13:04] <soren> It doesn't have to be in the initramfs, though, but I see what you mean.
[13:04] <kirkland> foo: 7.04?  really?
[13:04] <foo> kirkland: yup
[13:04] <soren> foo: Noone will care.
[13:04] <kirkland> foo: not sure, but 7.04 has been unsupported for a few years now
[13:04] <soren> foo: 7.04 went out of support almost a year ago.
[13:05] <foo> I've been doing some googling and tried with parameter irqpoll but that didn't to help
[13:05] <foo> oh
[13:05] <soren> foo: Don't use 7.04.
[13:05] <foo> err, I mean't 8.04
[13:05] <foo> my bad
[13:05] <soren> Ah.
[13:05] <kirkland> soren: yeah, one of the key design goals is *not* to prompt for any password on boot
[13:05] <foo> err, I mean 9.04
[13:05] <kirkland> soren: push that auth all the way back to GDM (or PAM as the case is)
[13:05] <foo> heh, wow... I think I even wrote it wrong on the CD
[13:06] <foo> Any thoughts? Google-fu is failing me
[13:06] <soren> kirkland: In a perfect world, we'd collect login credentials early enough to be able to use it for this, and then cache them and pass them to gdm.
[13:07] <soren> kirkland: In that case, there'd be no problem prompting for the password.
[13:07] <kirkland> soren: hmm, interesting
[13:07] <soren> kirkland: We already ask users for their password. What we want to avoid is asking *twice*.
[13:07] <soren> :)
[13:08] <kirkland> soren: in the encrypted LVM setup you mean?
[13:08] <soren> Yes.
[13:08] <soren> It would require being able to fire up X with a login screen without a rw filesystem.
[13:11] <soren> ...but that's hardly on-topic for this channel :9
[13:11] <soren> ":)", even.
[13:11] <kirkland> soren: heh, yeah
[13:12] <kirkland> soren: well, solving the encrypted swap hibernation problem is something i hope gets fixed for lucid
[13:12] <kirkland> soren: though i can't imagine i'll have time to work on it
[13:12] <soren> kirkland: Nah.
[13:16] <foo> gah, I can't seem to find anything helpful on google. it looks like others have been having this problem, though
[13:18] <garymc> [TK]D-Fender : What format have the ring tones got to be in and how do i add them?
[13:18] <soren> garymc: I'm going to guess you're in the wrong channel.
[13:19] <garymc> yeah good guess . sorry
[13:25] <soren> garymc: No worries.
[13:31] <zul> soren: fyi i fixed the opennebuela ftbfs last night
[13:31] <soren> zul: I saw. Thanks.
[13:34] <foo> well, making progress. Now ubuntu server 9.04 is hanging during install on input: AT Translated Set 2 keyboard as /devices/platform/i8042/serio0/input/input1. this is shocking
[13:36] <ttx> soren: about the populate_arp thing, I've no clue
[13:36] <soren> ttx: I'll defer to nurmi for an analysis.
[13:36] <soren> I guess he'll turn up in a few hours.
[13:37] <soren> I'm not sure what to do with the importance right now, though.
[13:38] <ttx> soren: by the way, duting triage day, you're supposed to triage bugs submitted yesterday, not necessarily the ones from today :)
[13:39] <soren> 11:54:24 < soren> (Doing triage for stuff from today as well, since we're cutting it close enough w.r.t. the release)
[13:39] <ttx> soren: cool, less work for me tomorrow :P
[13:39] <soren> In return you get to answer questions from me today :)
[13:39] <ttx> beh
[13:42] <soren> kirkland: can you triage bug 455901, please?
[13:43] <soren> kirkland: You're in a better position than I to ask for better feedback and such.
[13:43] <kirkland> soren: sure... can you move it over to qemu-kvm rather than kvm?
[13:45] <kirkland> soren: i marked it low, since the easy workaround is to use =1 cpu
[13:46] <kirkland> soren: and it's filed against jaunty/kvm-84; i suggest you ask them to test this against karmic/qemu-kvm-0.11
[13:47] <soren> kirkland: Good call.
[13:48] <kirkland> soren: i can do that, if you'd rather me do it
[13:49] <soren> kirkland: Nah, I'll do it.
[13:52]  * soren decides he needs gummy bears
[13:52] <soren> ...and coffee.
[14:00] <smoser> for anyone who didn't see amazon's announcement today, you can now get
[14:00] <smoser> instance type with 34GB and one with 68GB of memory for $1.20 and $2.40 / hour
[14:00] <kirkland> smoser: how many processors?
[14:00] <kirkland> smoser: url?
[14:01] <smoser> http://aws.typepad.com/aws/2009/10/two-new-ec2-instance-types-additional-memory.html
[14:01] <smoser> 4 and 8 cpu
[14:01] <smoser> but those are all fuzzy
[14:01] <smoser> they're measured in "EC2 Compute Units"
[14:01] <ttx> "Quadruple Extra Large", sounds like mcDonald advertising
[14:02] <smoser> yeah.
[14:02] <smoser> double secret probation
[14:02] <smoser> http://alestic.com/2009/10/ec2-4xlarge
[14:02] <smoser> thats erichammond's blog entry about it. he has /proc/cpuinfo for one of them.
[14:02] <elijahwright> those instance types are totally yummy.  the relational database service isn't bad either.   definitely good for a lot of biz uses.
[14:03] <smoser> yeah, the mysql is also cool.
[14:04] <smoser> and price also went down for the other types.
[14:04] <smoser> 8.5 cents per hour now for small
[14:04]  * kirkland needs to update his ec2-cost scripts
[14:04] <smoser> everything went down by 15%
[14:04] <smoser> for linux
[14:04] <ttx> smoser: is there an APi to query costs, so that kirkland doesn't need to update ec2-cost ?
[14:05] <smoser> i dont think so, but dont know for sure.
[14:07] <elijahwright> i'm pretty sure that there's not an API to calculate cost.
[14:07] <elijahwright> [someone writing a REST service to *do* that would be kinda cool, though...]
[14:18] <smoser> i think that rsync doesn't end up doing all that well with the compressed sparse tar files.
[14:18] <smoser> in regard to
[14:18] <smoser> rsync -aP uec-images.ubuntu.com::uec-images/karmic/20091027.1/karmic-uec-*.tar.gz .
[14:27] <soren> smoser: How do you compress them?
[14:27] <soren> smoser: "tar cvzSf" or some such?
[14:28] <soren> gzip wants --rsyncable in order for the compressed data to be... well, rsyncable.
[14:28] <soren> I don't know how to pass that from tar.
[14:30] <smoser> soren, http://bazaar.launchpad.net/%7Eubuntu-on-ec2/vmbuilder/automated-ec2-builds/annotate/head%3A/build-ec2-image calls tar. passes --use-compress-program=gzip-rsyncable
[14:30] <smoser> gzip-rsyncable is http://bazaar.launchpad.net/%7Eubuntu-on-ec2/vmbuilder/automated-ec2-builds/annotate/head%3A/gzip-rsyncable
[14:31] <smoser> i think its just that the images are so dense. there is some speed up, but rsync for the amd64 just averaged 290.49kB/s for me.
[14:32] <smoser> a lftp with pget would do as well or better than that on this link
[14:43] <zul> Nafallo: ping!
[14:43] <Nafallo> zul: pong
[14:44] <soren> smoser: Hm... Interesting experiment:
[14:44] <zul> Nafallo: which version of apache are you referring to for bug #460692
[14:44] <soren> smoser: Have you tried rsyncing the uncompressed image instead? Passing -z to rsync?
[14:44] <smoser> no.
[14:44] <Nafallo> zul: the server is running hardy, I checked that it still exist on karmic, so at least hardy to karmic I would assume :-/
[14:45] <zul> Nafallo: thanks
[14:46] <Nafallo> zul: does my "fix" seem sensible?
[14:47] <smoser> ttx, your bug 461301 comments really stink. we'll have to look at that in more depth shortly.
[14:47] <zul> Nafallo: yeah its probably easier to add the depends though ;)
[14:49] <Nafallo> zul: sure, and it would also make me less happy because it's a lot more ugly.
[14:49] <ttx> smoser: yes, there might be something fishy there
[14:50] <zul> Nafallo: yeah I know it would make me less happy as well ill look at it again in lucid
[14:50] <Nafallo> zul: oki
[14:53] <zul> Nafallo: obvouilsy its not going to get fixed for karmic ;)
[14:54] <Nafallo> zul: bah. coward ;-)
[14:54] <zul> Nafallo: good for an sru though
[14:54] <Nafallo> zul: want to keep track of that for me? :-)
[14:55] <zul> Nafallo: oh I will dont worry
[14:55] <Nafallo> awesome. thanks.
[15:33] <ttx> kirkland: did you have time to review the releasenote-potential euca bugs ?
[15:38] <acalvo> hi
[15:38] <acalvo> anyone could suggest a good ticketing system supported by ubuntu (and with LDAP, if possible)?
[15:38] <acalvo> I've been looking at osTicket, and is quite simple but there is no support for ubuntu
[15:39]  * soren calls it a day
[15:40] <acalvo> oh, soren, hi
[15:50] <ttx> zul: bug 449735
[15:50] <TeTeT> when starting the karmic image from UEC, it hangs waiting for /dev/sda2. see http://pastebin.ubuntu.com/302878/
[15:50] <ttx> zul: could you verify that against your proposed 3.4.2 and forward upstream if you can reproduce ?
[15:50] <kirkland> ttx: i did not have time to
[15:51] <zul> ttx: sure
[15:51] <TeTeT> any idea if that happens all the time or just in my cloud?
[15:51] <TeTeT> soren + smoser ^
[15:52] <ttx> kirkland: do you think you'll have more time today ?
[15:52] <ttx> or should I handle them ?
[15:54] <kirkland> ttx: probably better if you handle them.  i have a hard stop today
[15:54] <kirkland> ttx: and i'm testing the uec images
[15:55] <smoser> TeTeT, it will do that, yes.
[15:55] <kirkland> ttx: i spent yesterday verifying smoser's fixes, and tracking down a kvm issue that will need to be sru'd
[15:55] <smoser> you started with m1.small ?
[15:55] <ttx> kirkland: ok -- we can add more tomorrow anyway
[15:55] <TeTeT> smoser: yes, the default. Shall I use a different container?
[15:55] <smoser> TeTeT, it shouldn't hang terribly long though, right ?
[15:56] <TeTeT> smoser: well, for me it doesn't get out of this state at all
[15:56] <TeTeT> smoser: last instance I terminated after 15 minutes. I can leave this running for some more time
[15:56] <azteech> acalvo: in looking over osTicket summary, doesn't look like it matters what serverOS you are running. the only prerequisites are php mysql, and a server running apache. are you referring to there being no download for a ubuntu based server?
[15:57] <smoser> TeTeT, hm... you see that in console-output i guess ?
[15:57] <TeTeT> smoser: yes
[15:57] <smoser> its related to https://bugs.launchpad.net/ubuntu/karmic/+source/ec2-init/+bug/458850
[15:57] <acalvo> azteech: when installing osTicket, you need to tweak some configurations to match ubuntu's base config
[15:57] <acalvo> it's not a major problem
[15:57] <smoser> TeTeT, can you euca-get-console-output $iid | pastebin ?
[15:57] <acalvo> but I really like to install packages that are supported by ubuntu
[15:57] <smoser> err.. pastebinit
[15:58] <acalvo> so, every time there a bug fix a minor update, I can run apt-get upgrade and get rid of problems
[15:58] <smoser> TeTeT, and is this 32 or 64 bit instance?
[15:58] <TeTeT> smoser: 32 bit
[15:58] <zul> ttx: i can reproduce it
[15:58] <ttx> on 3.4.2 ?
[15:58] <zul> yep
[15:58] <smoser> so 32 bit c1.small, right?
[15:59] <ttx> zul: i don't think that bug is upstream yet
[15:59] <TeTeT> smoser: yes, though how do I see if it's a 32/64 bit image?
[15:59] <ttx> zul: they have https://bugzilla.samba.org/show_bug.cgi?id=6829
[15:59] <zul> ttx: creating a share called thisisreallyreallylongshare doesnt get displayed
[15:59] <ttx> but that's about special chars in the share comment field
[15:59] <zul> ttx: it will be in a couple of momments :)
[15:59] <azteech> acalvo: understood. prefer that myself. but, from brief looking over the wiki and forum, there are others who are having to mod as well.
[15:59] <smoser> TeTeT,$ euca-describe-images emi-247011C0
[15:59] <smoser> IMAGE   emi-247011C0    i-20091027105900/karmic-uec-i386.img.manifest.xml      admin    available       public  i386    machine
[15:59] <smoser>  
[16:00] <ttx> zul: make sure it's clear from description that its a client issue
[16:00] <zul> ttx: I will
[16:00] <acalvo> azteech: I know, but since, I usually ask here for advice, and then make a choice
[16:00] <ttx> if they consider 6829 critical for 3.4.3, I'm pretty sure yours should raise that game
[16:01] <TeTeT> smoser: ok, so it's 64bit, sorry for the confusion. The get-console-output is still the same as in pastebin http://pastebin.ubuntu.com/302878/
[16:01] <zul> ttx: can you link bug 6829 to the samba bug in launchpad that has the russian character one is not being displayed properly
[16:01] <ttx> zul sure
[16:01] <zul> stupid uvirtbot trix are for kids
[16:02] <smoser> TeTeT, can i see 'euca-describe-instances i-3B23075F' output
[16:03] <smoser> and then 'euca-describe-availability-zones verbose' also
[16:03] <TeTeT> smoser: http://pastebin.ubuntu.com/302881
[16:03] <TeTeT> smoser: http://pastebin.ubuntu.com/302883
[16:04]  * ttx adds "do not upgrade samba" to his LTS good resolutions
[16:06] <smoser> TeTeT, what image is this ?
[16:06] <smoser> i cannot reproduce it on the 20091027.1 build
[16:07] <TeTeT> smoser: it comes from the UEC image store
[16:07] <zul> ttx: samba bug 6852
[16:08]  * zul kick launchpad
[16:08] <TeTeT> smoser: Ubuntu 9.10 Release Candidate for i386 (argh) 20091022
[16:08] <ttx> https://bugzilla.samba.org/show_bug.cgi?id=6852
[16:09] <zul> ttx: this would be a good situation to package a samba daily git tree in a ppa
[16:09] <smoser> TeTeT, thank you. well, either way, i'd really appreciate it if i could see this not on RC, but on this current build.  there were changes to mountall, and other changes also that could definitely have fixed this.
[16:10] <smoser> and i cannot reproduce with current "release candidate" :)
[16:10] <TeTeT> smoser: ok, I can d/l the current build and try again. what's the original d/l url? cdimage.ubuntu.com?
[16:10] <zul> smoser: i see the ssh keys in the console ;)
[16:11] <smoser> TeTeT, http://uec-images.ubuntu.com/karmic/20091027.1/
[16:11] <smoser> there are instructions on how to download and register with your cloud at
[16:11] <smoser> http://testcases.qa.ubuntu.com/System/UECCloudImages
[16:21] <smoser> zul, what tests are you running ?
[16:21] <smoser> i'll avoid duplicate effort
[16:22] <zul> smoser: i did the single test for ec2 us. im going to finish that off after lunch but we want duplicate effort
[16:23] <smoser> i will do eu-west-1 region
[16:23] <smoser> doing the 4 tests there (amd64,i386 / single,multiple)
[16:24] <smoser> ttx, kirkland i've run the 20091027.1 on UEC, but i dont see any other results there. you guys running too ?
[16:24] <ttx> smoser: I've competed both tsets
[16:24] <ttx> tests
[16:25] <ttx> http://iso.qa.ubuntu.com/qatracker/result/3357/342
[16:25] <ttx> http://iso.qa.ubuntu.com/qatracker/result/3358/343
[16:25] <smoser> ttx, apparently i cant read
[16:25] <smoser> or reload or something
[16:25] <smoser> thanks. i hadn't put my result on the amd64 and thought that i had, thus was expecting > 1
[16:26] <smoser> (there are now 2 on both)
[16:27] <zul> smoser: acked
[16:34] <zul> ttx: ours is a dupe of 6829
[16:35] <ttx> zul: interesting
[16:35] <zul> yep
[16:35] <ttx> zul: didn't look that way
[16:35] <zul> ttx: volker said it *shrug*
[16:37] <ttx> he knows better than we do.
[16:40] <quikone> Is this the right place to ask server/networking questions?
[16:40] <TeTeT> smoser: there's another problem in my UEC now, can't start any more instances.
[16:42] <TeTeT> quikone: just ask
[16:43] <smoser> TeTeT, well, based on your availability-zones verbose output (http://pastebin.ubuntu.com/302883) you can only run 2 instances at once
[16:43] <TeTeT> smoser: I terminated all instances, it's a different problem. I haven't been able to reproduce it reliably, but I get timeouts from walrus when it happens
[16:44] <quikone> I have removed network manager in favor of network administrator, since then I cannot access my websites locally or remotely.  Any ideas on how to resolve
[16:44] <smoser> TeTeT, kirkland or ttx are probably more help with that.
[16:45] <TeTeT> smoser: the instance is in pending all the time. Restarting the NC sometimes helps
[16:45] <smoser> but please do make sure you're running the most recent stuff. i know that there were arbitrary (low) timeouts in some cases previously.
[16:46] <TeTeT> smoser: I've updated everything, it's yesterdays image plus updates on the front end and NC
[16:47] <Reepicheep> quikone: did you change IP addresses?
[16:47] <quikone> Yes
[16:47] <Reepicheep> have you restarted the web server?
[16:48] <quikone> Not recently, although I have restarted networking
[16:48] <Reepicheep> not reboot .. but just restart the web server (apache, lighttpd, ..)
[16:48] <smoser> TeTeT, thank you.  ttx, kirkland, you have any thoughts on that problem?
[16:48] <quikone> No Ihave not
[16:49] <Reepicheep> it may be binded to the incorrect IP
[16:49] <Reepicheep> also double check firewall rules if you set rules to the old IP address
[16:49] <quikone> okay. I will restart.  just apace2?  Or more?
[16:50] <Reepicheep> just apache should be fine
[16:51] <quikone> This could be it;  (98)Address already in use: make_sock: could not bind to address 0.0.0.0:443
[16:51] <quikone> no listening sockets available, shutting down
[16:53] <Reepicheep> quikone: "netstat -tapn | grep LISTEN" will show you what is listening on what ports and by what processes
[16:56] <quikone> I don't see apache2 listening anywhere??
[16:57] <Reepicheep> quikone: make sure it's running.  Check the logs to make sure it isn't erroring on startup
[16:57] <Reepicheep> like maybe it is trying to bond to the old IP which doesn't exist anymore
[16:58] <quikone> That would be my guess, how do I get back to the config files
[16:58] <ttx> smoser: no time sorry
[16:59] <ttx> TeTeT: doesn't ring a bell
[16:59] <TeTeT> ttx: it's fine. now the cloud stopped working completely. I'm calling it a day and see what tomorrow brings
[17:02] <mathiaz> kirkland: hey - question about kvm - which files is used for the bios?
[17:03] <ninjah> Is ebox worth using?
[17:03] <mathiaz> kirkland: I'd like to replace the bios with one the gpxe version - so that I don't need to chainload the boot loader
[17:05] <alvin> ninjah: eBox looks nice, but in my personal opinion, it is not finished yet. It's also very Windows-minded. (no NFS for example)
[17:06] <ninjah> alvin: Hmm... I was a fan of webmin for some things. I was told to switch to ebox.
[17:07] <alvin> ninjah: I never used webmin. I will probably test out eBox again in the future, but now, it just can't do the things I want.
[17:08] <alvin> (like NFS, and ldap address book)
[17:08]  * alvin is going home
[17:08] <DrNick_> i'm not a fan of ebox
[17:10] <DrNick_> i vaguely remember trying it, and it installing a whole bunch of junk with it.  i wouldn't say it is any better than webmin (which i dont find that amazing either :p)
[17:14] <ninjah> DrNick_ I thought webmin was cool. Easy way to allow someone to configure a box if they don't know Linux
[17:14] <DrNick_> yeah its not bad
[17:17] <flagg0204> didnt care for ebox myself, nor webmin for that matter.  but it does have it place for allowing novices to work with some of the more complex elements of linux
[17:18] <DrNick_> agreed
[17:23] <jjohansen> smoser have you gotten a chance to do any testing on those EC2 test kernels yet?
[17:23] <mrchrisadams> i appreciate the irony of asking this in a ubuntu server channel, but... what's GUI equivalent to fdisk/partman in stock ubuntu?
[17:24] <ninjah> haha... Don't know if I've ever seen a GUI partition tool
[17:24] <smoser> jjohansen, :-(
[17:24] <smoser> sigh
[17:25] <jjohansen> well we are all busy, its not like I have manage to get it uploaded to eu region either
[17:25] <smoser> mrchrisadams, gparted ?
[17:29] <mrchrisadams> smoser: I'm using ubuntu jaunty
[17:29] <mrchrisadams> system > administration  > (should it be here?)
[17:29] <smoser> mrchrisadams, maybe i'm missing something: http://packages.ubuntu.com/jaunty/gparted
[17:30] <smoser> you might need to 'apt-get install gparted'
[17:30] <smoser> i dont know where it'd show up. i'd just run it from command line.
[17:34] <DrNick_> there is gparted yes, but that requires X11 installed.  you could try 'cfdisk' instead of strait 'fdisk' - it provides a menu driven interface - very easy in my experience
[17:36] <jsalisbury> I performed an apt-get dist-upgrade to get to the RC.  Is there a way to force another upgrade even if I'm already at the RC?
[17:36] <DrNick_> aside from that though, i normally use lvm for all my volume management, so most drives i stick in a server don't have a partition table as such
[17:36] <jsalisbury> I want to see if this caused the issue specified in Bug #455625.  However as expected, re-running apt-get dist-upgrade says 0 packages to upgrade.
[17:37] <DrNick_> even with one disk, lvm makes sense
[18:09] <smoser> mathiaz, zul better late than never, bzr pull lp:~smoser/+junk/ec2-test
[18:09] <smoser> https://code.launchpad.net/~smoser/+junk/ec2-test
[18:30] <kirkland> ScottK: ping
[18:37] <bdmurray> zul: wrt bug 458637 should samba be able to access windows 7 shares?
[18:38] <zul> bdmurray: it should
[18:39] <bdmurray> zul: I've seen a couple of other reports regarding that specific issue.  Should I retitle that bug and make them dupes or do something else?
[18:39] <zul> retitle the bug and make them dupes
[18:40] <zul> we'll probably have to do an SRU for it
[18:49] <acalvo> hi
[18:49] <acalvo> anyone know how can I reset/recover the admin password of a LDAP tree (not the config tree though)
[18:49] <acalvo> ?
[18:50] <ahasenack> acalvo: you mean the password of the rootdn dn?
[18:50] <acalvo> yes
[18:51] <ahasenack> acalvo: just change it in slapd.conf or, if using slapd.d, in a file inside that directory (I don't remember which one, but shouldn't be hard to find)
[18:51] <ahasenack> acalvo: it's the rootpw entry
[18:51] <ahasenack> acalvo: but, if there is no rootpw entry, then it means that rootdn has an entry in the directory itself
[18:51] <acalvo> mmm I'm gonna check it
[18:51] <ahasenack> acalvo: and the password should be reset in that entry
[18:51] <ahasenack> acalvo: I suppose you could add a rootpw temporarily in that case
[18:56] <bventura_> anyone know how to enable mysql warning logging on ubuntu?  I've tried setting 'log-warnings = 2' in my.cnf and it does not seem to work
[19:04] <Coosh> Is there anyway to see who is logged onto a pptp vpn in ubuntu? (aparantly logwtmp is broken ?)
[19:33] <quikone> Is there documentation somewhere to have multiple nics online and configured at the same time.  It seems that if one is up the other goes down.
[19:36] <billybigrigger> hey all
[19:36] <billybigrigger> anyone alive?
[20:05] <smoser> nekro_, ping regarding 461156
[20:10] <nekro_> smoser: I am unable to reproduce the problem. Eucalyptus should handle safe as well as unsafe b64 encoding in the latest upstream.
[20:11] <smoser> it absolutely reproduces with current karmic versions
[20:11] <nekro_> smoser: I don't know what to tell you.
[20:11] <smoser> you see the stack trace? it looks at least reasonable that it would fail there.
[20:12] <smoser> in the code i'm looking at
[20:12] <nekro_> smoser: yep. there was an issue at some point. Not sure which revno it was fixed in (a while ago so not sure why you see in with the latest karmic)
[20:14] <smoser> where do i get upstream revision control from ?
[20:14] <nekro_> smoser: lp:eucalyptus/1.6
[20:14] <nagappan> hi cr3
[20:15] <nagappan> cr3, I'm here :)
[20:15] <cr3> can someone recommend documentation about running two dhcp servers on the same network? is it even possible without wrecking havok?
[20:15] <cr3> nagappan: cool, now we wait for the gods to answer. I can be the virgin we offer as a sacrifice
[20:17] <nagappan> cr3, :D
[20:18] <_ruben> cr3: with which goal in mind? dhcpd supports clustering
[20:20] <cr3> _ruben: personally, I'm just curious. nagappan has another use case where he would like one dhcp server to offer static addresses and another dynamic addresses
[20:20] <smoser> kirkland, can you reproduce that bug ?  It is as easy as removing 2 lines in euca-run-instances, and then running: ec2-run-instances -k mykey --user-data " << FOO >" ${EMI}
[20:21] <smoser> nekro_, note, that you have to have patch/fixed euca-run-instances to not double encode data
[20:24] <_ruben> cr3 / nagappan : how would each dhcp server know whether or not to respond to a request?
[20:24] <calzifer> hi, i'm trying to run xvfb with jdownloader on my homeserver without x, but how can i connect on the xvfb ?
[20:24] <nagappan> _ruben, First DHCP server has to serve based on MAC address
[20:25] <_ruben> and i dont see why you cant use a single dhcp server which hands out both static and dynamic leases (since that's what i do at work)
[20:26] <guntbert> nagappan: don't use 2 dhcp servers if possible - they tend to get in each other's way
[20:27] <nekro_> smoser: I see. yeah I had an older version installed. I can reproduce it now, thanks.
[20:31] <nagappan> guntbert, okay
[20:32] <nagappan> guntbert, cr3 suggested me to write static ip in /etc/network/interfaces, let me try that out, this is our test system in VMware, after a batch of test is completed, we do reimaging (using Clonezilla)
[20:33] <nagappan> guntbert, I have set of scripts which run on each host, probably I can populate the data
[20:33] <cr3> nagappan: just to make sure, if you hard code the static IP address in /etc/network/interfaces, you won't need the second dhcp server on the network, right?
[20:34] <nagappan> cr3, we need it for reimaging by Clonezilla
[20:34] <guntbert> nagappan: sorry, I didn't follow - just jumped on the sentence with two dhcp servers - and of course it is possible to have one server to disk out fixed addresses and random ones
[20:35] <cr3> nagappan: so hard coding the static IP address might only solve one of your problems but, as suggested by guntbert, you probably shouldn't be running 2 dhcp servers. could you perhaps isolate both networks by vlan or something?
[20:36] <smoser> you could run 2 dhcp servers on one network, but you just want to make sure that for every MAC that the one is going to answer on , the other is going to ignore
[20:36] <qman__> yeah, things like clonezilla should generally be done on an isolated network
[20:36] <smoser> and vice versa. otherwise you're asking for confusion.
[20:36] <cr3> nagappan: if I may speculate, is the problem that you have one dhcp server managed by people outside your team and you need another dhcp server managed by your team for other purposes?
[20:38] <nekro_> smoser: I have confirmed that euca-run-instances --user-data works against amazon. I am adding this as an upstream task to fix
[20:38] <cr3> qman__: nagappan seems to have two nics per machine, so it should be feasible to isolate cloning to one physical network and the rest to the other more general network
[20:38] <nagappan> cr3, sure
[20:39] <smoser> nekro_, it is fairly high priority as a karmic-updates... user-data is generally broken without it
[20:39] <nagappan> cr3, second dhcp server maintained by our techops team
[20:39] <nagappan> cr3, fist dhcp server maintained by me, with Clonezilla setup
[20:40] <cr3> nagappan: aha! so there's probably a lot of overhead to implementing smoser's approach where you'd have to ask the techops team to specifically ignore each mac address you intend to server statically from your own dhcp server
[20:41] <smoser> i suspected that cr3 was correct.  it'd definitely be a pain to coordinate
[20:41] <cr3> nagappan: instead, you're probably better off isolating both networks, just using vlans if you happen to have that luxury or physically otherwise. then, both dhcp servers will not conflict with each other
[20:41] <smoser> i have done this before, but in a network where the "official" server only responded to known macs
[20:42] <smoser> so my dhcp server was free to respond to known macs that the first didn't know about
[20:42] <bventura_> anyone know how to enable mysql warning logging on ubuntu?  I've tried setting 'log-warnings = 2' in my.cnf and it does not seem to work
[20:43] <nagappan> cr3, sure
[21:13] <zul> mathiaz: do you have a suggestion for a fix for 462169?
[21:13] <mathiaz> bug 462169
[21:13] <mathiaz> zul: nope - not really
[21:27] <fbc-mx> What command would you use to tell ubuntu server to ping all the distros and use the closest one? Like the desktop does.
[21:29] <_ruben> ping doesnt relate to download speed at all
[21:31] <fbc-mx> _ruben, ok, then whatever wat the desktop determines it can I get the server version to?
[21:32] <pmatulis> huh?
[21:33] <fbc-mx> _ruben, ok then how to I get the server to use the closest distro is there some automated way of doing it like the desktop does?
[21:39] <pmatulis> archive.ubuntu.com
[21:48] <fbc-mx> exit