2008/10/20/#ubuntu-server.txt

[02:39] <arrrghhh> will intrepid server use kernel 2.6.27-something?
[02:40] <hads> Yah
[02:40] <arrrghhh> schweet.  i hear the ntfs-nfs problem is no more in that kernel.
[02:41] <arrrghhh> i'm using the nfs-user-server package instead of the kernel-server package.  i'd much rather have nfs run in kernel-space instead of user-space.
[02:42] <arrrghhh> and i can just do an aptitude safe-upgrade on oct 30th and get intrepid?
[02:43] <hads> You should use do-release-upgrade
[02:43] <arrrghhh> i've never used that one
[02:46] <arrrghhh> can i do that right now and get the beta?
[02:47] <hads> Should be able to, might need a flag
[02:48] <arrrghhh> like -d
[02:48] <hads> Probably
[02:48] <arrrghhh> well on my server i'd rather wait.  i'd really like the new kernel, but i'd much prefer stability.
[02:53] <arrrghhh> the -m switch for mode - server or desktop.  is the mode assumed based on the current running system i hope?
[03:20] <tr-33> can anyone help me with getting iptables to accept incoming connections on port 80 and forward these connections to an internal ip
[03:55] <golb> with vm-builder, i'm supposed to have an Ubuntu intrepid installed first before running the script, aren't i?
[04:52] <rascov> tr-33: iptables -t nat -A PREROUTING -p tcp -d <publicIP> --dport 80 -j DNAT --to <privateIP>:<port>
[08:01] <scuser> hi all, how can I disable simple bind in openldap ?
[08:47] <kraut> moin
[08:58] <RockHound> hi everyone, question about mysql master/slave setup: the logrotate script on the slave fails as debian-sys-maint differs per host. How do I handle this? Just copy the debian.cnf to every host? or create a different maint user for each host?
[09:02] <mathiaz> RockHound: you could replicate each maint user on the master.
[09:03] <mathiaz> RockHound: so that every slave has all of the debian-sys-maint.
[09:03] <mathiaz> RockHound: hm - it may not actually work now that I think of it,
[09:03] <RockHound> ;-)
[09:03] <mathiaz> RockHound: Your only option is probably to use the same debian-sys-maint on every host.
[09:04] <RockHound> or create different users for each ost
[09:04] <RockHound> *host
[09:05] <mathiaz> RockHound: right - but then you'd have to modify the debian.cnf anyway
[09:05] <RockHound> true
[09:07] <RockHound> okay will just go ahead and run with the same debian.cnf ...
[09:07] <mathiaz> RockHound: seems like the easiest way to do it
[09:08] <scuser> hi all, how can I disable simple bind in openldap ?
[09:09] <RockHound> scuser: check the slapd admin manual on slapd.org ... big section about security .. don't know it from the top of my head
[09:12] <scuser> RockHound: I can't find slapd.org
[09:12] <RockHound> http://www.openldap.org/doc/admin24/
[09:12] <RockHound> sorry
[09:12] <RockHound> section 13.3
[09:42] <mathiaz> erichammond1: hi - have you been able to boot an ec2 guest built with vmbuilder?
=== mdz_ is now known as mdz
=== _ruben_ is now known as _ruben
=== xelpud is now known as fduplex
[10:30] <scuser> hi all, any help with that error "unwilling to perform simple authentication" ?
[10:40] <scuser> hi all, how to configure an ldap client not to do simple authentication ?
=== Deeps_ is now known as Deeps
[12:29] <wo0f> is there a ubuntu-server XEN iso out there?
[12:30] <wo0f> or is it a case of installing xenified drivers after iv virtualized it?
[12:31] <yann2> I think the latest python-vm-builder can create images usable by XEN
[12:32] <yann2> + it lets you configure a lot - you should have a look :)
[12:35] <wo0f> oo
[12:36] <wo0f> that looks exactly what im after :D (Y)
[12:36] <wo0f> cheers yann2
[12:38] <yann2> the clever one is in intrepid but you may find a backport :)
[12:47] <scuser> hi all, how to configure an ldap client not to do simple authentication ?
[13:22] <uvirtbot`> New bug: #286290 in postfix (main) "upgrade process hangs" [Undecided,Incomplete] https://launchpad.net/bugs/286290
[13:30] <scuser> hi all, can anyone tell me what are the available SASL_MECH options ?
[13:30] <scuser> hi all, can anyone tell me what are the available SASL_MECH options in the ldap.conf file ?
[13:52] <MuelleJ> A software package I am trying to install is requiring libstdc++-libc6.2-2.so.3, can I install the libstdc++-libc6.2-2 package from gutsy in hardy?
[14:32] <scuser> hi all, does anyone know how to force the client not to do simple authentication, which configuration file should I modify ?
[14:33] <MuelleJ> scuser: can you elaborate, I don't understand
[14:35] <scuser> MuelleJ: I'm trying to login to a client which is configured to do simple authentication in order to login; the users are stored in an ldap database which is configured not to allow simple authentication so I want to disable simple authentication in order to have a successful login, got it ?
[14:37] <RockHound> scuser: with what are you trying to login?
[14:37] <scuser> RockHound: the client is configured to login via kerberos
[14:38] <RockHound> What is "the client"?
[14:39] <scuser> RockHound: a simple computer in the network, you can call it a terminal a PC or whatever
[14:41] <scuser> RockHound: any ideas
[14:42] <RockHound> it is using kerberos to access a kerberos real which has LDAP as its backend?
[14:42] <RockHound> or pam_ldap?
[14:42] <scuser> RockHound: exactly, no pam_ldap
[14:43] <RockHound> but there must be a pam module somewhere?!
[14:43] <RockHound> no specifically for ldap but for kerberos
[14:43] <scuser> RockHound: pam_krb5 is the one which I use
[14:45] <RockHound> okay ... so the problem lies in the communication between kerberos and openldap?
[14:46] <scuser> RockHound: yes
=== lamont` is now known as lamont
[14:50] <scuser> RockHound: so could you help me with that or just give me guidance ?
[14:54] <RockHound> scuser: I have not implemented such a setup myself but have you tried googling at all?
[14:54] <scuser> RockHound: for almost one week :D
[14:55] <RockHound> what is the exact error that you are getting?
[14:57] <scuser> RockHound: the server logs reports "unwilling to perform simple authentication" error
[15:00] <RockHound> http://www.semicomplete.com/articles/openldap-with-saslauthd/ maybe
[15:01] <RockHound> but to answer your questions about SASL_MECHs: http://linuxcommand.org/man_pages/saslauthd8.html
[15:02] <RockHound> and google really is your friend
[15:10] <nijaba> RockHound: do you know about http://manpages.ubuntu.com/ ? Quite useful as well --> http://manpages.ubuntu.com/manpages/hardy/en/man8/saslauthd.html
[15:12] <RockHound> nijaba: well thanks
[15:12] <nijaba> RockHound: thanks to kirkland whom put it together :)
[15:15] <nijaba> mathiaz, zul: I am proposing the merge of a couple additions to vmbuilder, could you please let me know what you think? -> https://code.launchpad.net/~nijaba/vmbuilder/libvirt-netopt/+merge/1355
[15:15] <zul> nijaba: suuuuuure :)
[16:01] <zul> looks fine to me
[16:02] <tonyyarusso> Say, what's the server team recommending these days for system backup?
[16:03] <tonyyarusso> Right now I'm thinking we'd probably want to do daily differential backups to a USB-attached external hard drive, maybe a weekly full backup the same way, and then a monthly backup of the most critical components to DVD+RW.
[16:03] <nijaba> tonyyarusso: backuppc or bacula
[16:03] <nijaba> backuppc is simpler, bacula VERY complete
[16:03] <tonyyarusso> Does complete = complicated?  :P
[16:04] <nijaba> tonyyarusso: as well, unfortunately
[16:04]  * tonyyarusso looks at package descriptions
[16:04] <tonyyarusso> We could probably get away with something simple for now (one machine), but it might be a good idea to learn the more scalable option while we're at it I suppose.
[16:05] <nijaba> tonyyarusso: I think it would be a sane investment, if you have a bit of time
[16:05] <tonyyarusso> In random other thoughts, samba should probably be a recommends instead of a depends for backuppc...
[16:06] <tonyyarusso> also odd that bacula is in main, yet lists MOTU as the maintainer.  :P
[16:06] <nijaba> tonnyarusso: mmh, that would not change much since with 8.10, recommends gets pulled by default as well, unless that's a policy comment, in which case I could not accurately comment
[16:08] <nijaba> zul: see ^^, should we fix that?
[16:08] <zul> tonyyarusso: yeah we should
[16:08] <zul> i dont think its a big deal though
[16:09] <tonyyarusso> Okay, next question:  Given recent changes to consumer data privacy law, I'd like to be able to keep the data in certain MySQL databases (but preferably not all of them) encrypted.  Is there a way to do that, or would you have to do whole partitions for it to work?  (I'm not familiar enough with how MySQL stores the data on disk.)
[16:09] <tonyyarusso> zul: No, not big.  I'll drop a bug in though.
[16:09] <zul> k
[16:13] <tonyyarusso> Huh - according to http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=389793 samba actually is needed.  Probably a bug in itself, but that's beyond the scope of what I know.
[16:13] <uvirtbot`> Debian bug 389793 in backuppc "backuppc: Should not depdend on smbclient, samba-common | samba-tng-common" [Normal,Open]
[16:21] <tonyyarusso> I believe http://blog.i64.pl/PiosBlog/200609/01-secure-mysql-database/ answers the latter question.  Nifty.
[16:45] <Koon> nxvl: congrats on your motu-sru nomination. I could use some movement on bug 179447 :)
[16:45] <uvirtbot`> Launchpad bug 179447 in tomcat5.5 "Installation of tomcat5.5 fails if sun-java-jdk is not installed" [Medium,Fix released] https://launchpad.net/bugs/179447
[16:52] <mathiaz> nijaba: I'd add some checks for the net-virtio option as some distro don't support virtio at all (< hardy)
[16:52] <mathiaz> nijaba: I don't know however how easy that could be implemented.
[16:52] <nxvl> Koon: just send me an e-mail to check it after work hours
[16:52] <nxvl> :D
[16:52] <nijaba> mathiaz: hmmm, no, network virtio is supported on hardy, not the bd part.
[16:53] <mathiaz> nijaba: dapper doesn't support network virtio
[16:53] <nijaba> mathiaz: but I'll do some checking for older os, you are right
[16:53] <nijaba> s/os/suite
[16:55] <mdz> dendrobates: I just installed a recent server daily into KVM using LVM and it doesn't boot.  GRUB says "Error 15: file not found"
[16:56] <dendrobates> mdz: hmm,  can you create a bug?   kirkland can you reproduce it?
[16:59] <mdz> dendrobates: I just phoned cjwatson and he is looking into it
[16:59] <mdz> dendrobates: it looks like it is using GRUB and uuid when it shouldn't be
[16:59] <kirkland> dendrobates: i'm downloading the server iso now
[17:00] <mdz> kirkland: I'm about to head onto a phone call but if you tell me which package, I'll get a bug filed when I get off
[17:01] <kirkland> mdz: file against grub, assign to me, and i'll reclassify as per necessary
[17:08] <nijaba> mathiaz: --net-virtio check added in the branch
[17:13] <mathiaz> nijaba: ok - looks good to me. I don't know if it's a good thing to merge it in trunk now.
[17:13] <nijaba> mathiaz: how so?
[17:13] <mathiaz> nijaba: this is a new a feature - and we'll have to ask for a FFexception
[17:13] <nijaba> mathiaz: gah, right :(
[17:13] <mathiaz> nijaba: there is a bunch of new code already - most of it is bug fix.
[17:14] <mathiaz> nijaba: the ec2 plugin is the only one that qualifies as a new feature.
[17:14] <tonyyarusso> nijaba: Do you happen to know if the version of dvd+rw-tools in Ubuntu has the patch needed for Bacula mentioned at the bottom of http://www.bacula.org/en/dev-manual/System_Requirements.html ?
[17:14] <mathiaz> nijaba: while I agree that this is useful,  I don't think *now* is the best moment to merge it.
[17:15] <nijaba> mathiaz: ok, np.  maybe I'll add the check on libvirt / kvm only then
[17:15] <mathiaz> nijaba: yeah - that seems like a good check to have.
[17:18] <nijaba> tonyyarusso: feisty and > have 7.0 or >, so I'd ask on some bacula chan/ml if they know if the patch has been integrated, is still needed
[17:18] <nijaba> tonyyarusso: http://packages.ubuntu.com/search?keywords=dvd%2Brw-tools&searchon=names&suite=all&section=all
[17:19] <tonyyarusso> okay, thanks
[17:25] <nijaba> mathiaz: ok, rev 252 has it
[17:43] <kirkland> mdz: dendrobates: i am able to reproduce the problem
[17:43] <kirkland> mdz: dendrobates: looks like a bug somewhere in the new dm code, as the grub entry is set to root=/dev/dm-0
[17:43] <kirkland> mdz: dendrobates: which should not be the case
[17:44] <mdz> kirkland: please liaise with Colin about it, he's looking into it as well
[17:44] <kirkland> mdz: -> #ubuntu-devel, okay
[17:52] <tonyyarusso> nijaba: Ah, here we go - "Note, some Linux distros such as Debian dvd+rw-tools-7.0-4 package already have the patch applied"
[17:53] <nijaba> tonyyarusso: \o/
=== iftel is now known as sme2k8
[18:17] <zul> nijaba mathiaz: lp:~zulcss/vmbuilder/vmbuilder-bugfixes
[18:18] <nijaba> zul: great. Will look at it after dinner. Does it include the separate package for ec2?
[18:20] <zul> nijaba: of course
[18:20]  * nijaba hugs zul
[18:21] <zul> nijaba: tries to remember what he was doing ;)
[19:25] <LordDicranius> I was wondering if anybody could help me decode my "df" output :)
[19:25] <hylje> which part of it?
[19:25] <LordDicranius> the device names
[19:26] <LordDicranius> c0d0p*, c0d1p*, c0d2p*
[19:26] <LordDicranius> c0d0 conists of 2 physical drives
[19:26] <LordDicranius> c0d1 and c0d2 both consist of 3 physical drives
[19:27] <hylje> it's a hardware raid setup
[19:27] <LordDicranius> c0d1 and c0d2 both are one big partition
[19:27] <LordDicranius> ah, ok
[19:27] <hylje> as far as i can tell
[19:27] <LordDicranius> ok, that makes more sense
[19:28] <LordDicranius> at a new job and I'm tryin to decipher stuff since according to my IT manager he knows nothing about it, it was setup by the guy they bought it from
[19:28] <LordDicranius> heh
[19:31] <LordDicranius> is there a term for the c0d2p1 stretching across 3 physical hard drives?
[19:32] <hylje> probably striping
[19:32] <LordDicranius> in trying to research I've come across "physical partitions," "logical partitions", "physical volumes", "logical volumes" - and those are all just confusing me lol
[19:32] <LordDicranius> oooh
[19:32] <LordDicranius> rebooting the server and getting to the RAID controller software would probably tell me more, huh?
[19:32] <hylje> look up how RAID works first
[19:33] <hylje> you might not see much sense from the config you get without knowing RAID first
[19:33] <LordDicranius> ok
[19:34] <LordDicranius> I think that get me on the right path
[19:34] <LordDicranius> thx for you help hylje :)
[19:34] <hylje> you're welcome
[20:05] <nijaba> zul: bzr merge  lp:~zulcss/vmbuilder/vmbuilder-bugfixes
[20:05] <nijaba> Nothing to do.
[20:06] <nijaba> zul: did you forget to push your changes?
[20:06] <zul> yeah....
[20:06] <zul> lemme check
[20:07] <zul> *sigh*
[20:07] <zul> gimme a sec :)
[20:07] <nijaba> np
[20:09] <zul> how do you delete branches?
[20:10] <nijaba> zul: mark them obsolete
[20:10] <nijaba> zul: click on the yellow pen icon
[20:11] <nijaba> err s/obsolete/abandoned
[20:16] <zul> lp:~zulcss/vmbuilder/vmbuilder-ec2fix
[20:22] <nijaba> zul: re:         return '/boot/vmlinuz-2.6.27-7-server'
[20:22] <nijaba> 29
[20:22] <nijaba>  
[20:22] <nijaba> 29
[20:22] <nijaba>  
[20:22] <nijaba> 30
[20:22] <nijaba>     def xen_ramdisk_path(self):
[20:22] <nijaba> 30
[20:22] <nijaba>     def xen_ramdisk_path(self):
[20:22] <nijaba> 31
[20:22] <nijaba>         return '/boot/initrd.img-2.6.27-2-server'
[20:22] <nijaba> 31
[20:22] <nijaba>         return '/boot/initrd.img-2.6.27-7-server'
[20:23] <zul> f*ck
[20:23] <nijaba> zul: does this means we will need to update pvb at each kernel release?
[20:23] <zul> yep
[20:23] <nijaba> meh...
[20:23] <zul> unless if I come up with a way to automatically detect it
[20:23]  * nijaba tought he selected only one line before pasting...
[20:24] <nijaba> zul: mhh, let me think about that. /me not too please at the prospect of running after the latest kernel...
[20:25] <zul> nijaba: what I was thinking was to grab all of the config-2.6.27-*-server files in /boot and look for HVC_XEN and return the latest one by data
[20:25] <zul> date even
[20:27] <nijaba> zul: the way versioning works, latest should be last in alphabetical order as well
[20:27] <zul> true
[20:29] <nijaba> zul: have time to give a shot at this, or do yo need offloading?
[20:31] <zul> nijaba: can you look at it, I have to look at why Im doing something wrong here
[20:31] <nijaba> zul: np
[20:35] <nijaba> zul: regarding http://bazaar.launchpad.net/~zulcss/vmbuilder/vmbuilder-ec2fix/revision/254?start_revid=255, did you forget to add some python-vmbuilder-ec2.install file?
[20:35] <zul> indeed, I guess I didnt do a bzr add
[20:36] <zul> it just contains the ec2 python directory
[20:36] <nijaba> ok, I'll add it. Thanks
[20:45] <sandstrom> I have installed php and mysql by themselves. Is there a difference between this and using the tasksel method?
[20:45] <sandstrom> would it be better to remove apache, php and mysql and install lamp using tasksel?
[20:49] <nijaba> zul: I would suggest you to install https://pastebin.canonical.com/10376/ as ~/.vim/python
[20:50] <nijaba> sandstrom: if it works, it means that you should have done it very close to what tasksel would do
[20:50] <ScottK> nijaba: I think it's a bit rude to use a private pastebin on a public channel like this.
[20:50] <sandstrom> nijaba: so there is no point in doing it all over with tasksel?
[20:51] <nijaba> ScottK: oh, sorry,  I thought I used a public one
[20:51] <sandstrom> tasksel wont glue it all together in any specific way?
[20:51] <ScottK> nijaba: No.  The canonical one is closed.
[20:52] <nijaba> ScottK: yes, but I thought I used the ubuntu one.  So here it is: http://pastebin.ubuntu.com/60231/
[20:53] <nijaba> sandstrom: no, it just makes sure you have a sensible set of packages whose install has been tuned to glue it all together.  So if you picked the same packages, you should be fine
[20:54] <sandstrom> I picked php5, mysql5 and thats it
[20:54] <sandstrom> (and apache2)
[20:55] <nijaba> sandstrom: may need a few more to glue is all, so would advise to just run tasksel to complete
[20:55] <sandstrom> but I didn't go into build-details on which php module to include or what apach2 modules to include
[20:55] <sandstrom> thanks
[21:36] <uvirtbot`> New bug: #286614 in openldap (main) "slapd apparmor profile needs update for multiple databases" [Undecided,New] https://launchpad.net/bugs/286614
[21:41] <uvirtbot`> New bug: #286616 in openldap (main) "slapd apparmor profile needs update for multiple databases" [Undecided,New] https://launchpad.net/bugs/286616
[22:07] <Linuturk> is there an rss feed of Ubuntu updates?
[22:07] <Linuturk> similar to the debian feeds?
[22:36] <uvirtbot`> New bug: #286643 in bacula (universe) "bacula client configuration is broken out of the box" [Undecided,New] https://launchpad.net/bugs/286643
[23:03] <mathiaz> nijaba: awesome - all of my test for python-vm-builder are working now!
[23:10] <nijaba> mathiaz: yes, and rev 253 now merges zul's fixes
[23:10] <nijaba> mathiaz: there are a couple new bugs for vmware, but I think we'll deal with that later. time for an upload
[23:22] <mathiaz> nijaba: right - I'll work on doing an upload tomorrow
[23:22] <mathiaz> nijaba: kvm+libvirt testing works for me now.
[23:23] <mathiaz> nijaba: I'll have to test the ec2 plugin tomorrow
[23:23] <mathiaz> nijaba: and then we get something uploaded.
[23:23] <mneptok> libvirt, you demon bitch goddess.
[23:23] <mneptok> oh, sorry. was that out loud?
[23:27] <tonyyarusso> Speaking of which, anyone want to explain the concept of KVM in oversimplified terms?  (I used VMware Fusion briefly, but I still don't really understand the different types of virtualization and whatnot, and what host/guest pairings work in each, etc.)
[23:27] <tonyyarusso> I have a friend who'd like to have a Windows guest on his Ubuntu server for the sole purpose of running a distributed rendering application, so I'm starting to look into how that could be done with KVM and how it works.
[23:29] <nijaba> zul: here is my experimental branch for xen https://code.launchpad.net/~nijaba/vmbuilder/xendiscoverkernel
[23:29] <zul> kvm+special hardware+heady goodness
[23:29] <zul> nijaba: coolio
[23:31] <gsanatci> hi
[23:33]  * nijaba -> bed
[23:34] <nijaba> goodnight
[23:34] <gsanatci> I installed ubuntu desktop edition 8.04, but i'm using like server (mysql). I'm not sure, is it correct ?
[23:34] <gsanatci> Maybe i'm loosing to much performance.
[23:36] <gsanatci> i think, all peoples sleeping :)
[23:36] <gsanatci> good night guys.
[23:48] <trashguy> [Solars],
[23:49] <trashguy> what was the torrent thing u showed me