[00:45] <lamont> kirkland: around?
[00:46] <lamont> libvirt vs windoze xp pro, drive C becomes readonly as of recently, was working a few days back
[00:46] <lamont> (lucid) - wtf?
[00:46] <lamont> wandering back and forth, so laggy
[01:29] <ryanakca> ... is there any reason why, under Lucid, I get a bunch of 000 executables in /sbin and /usr/sbin ? http://paste.ubuntu.com/405041/
[01:31] <Parabola> sup guys
[01:42] <ryanakca> ... I mean, I just reinstalled upstart and it is "12K ---------- 1 root root 9.3K 2010-03-17 19:38 /sbin/runlevel"
[01:42] <ryanakca> Filesystem problems?
[02:33] <zagabar_> Hello.
[02:34] <zagabar_> I have an ubuntu server and I recently got vnstat. Then I saw that my server is like constantly sending data. Approx 200 mb per hour is transmitted. Almost nothing is reiceved. I have no idea what this transmission is. How can I check it? It is a command line only serve.r
[02:37] <CBebop> I have an issue with Lighttpd not being able to use port 80
[02:37] <CBebop> saying it is in use
[02:37] <CBebop> but I have no processes using port 80 and I cant find why it would be in use
[02:37] <lamont> kirkland: interesting - 'twould appear that the issue was windoze vs self
[02:37] <CBebop> only happened after a reboot, used to work just fine, I am using Ubuntu Server 10.04
[02:38] <jeffesquivel> zagabar_, iptraf ?
[02:38] <CBebop> I dont have a lot of experience with this, but lighttpd is set up correctly as far as I know, and works fine if i specify a different port like 81
[02:40] <pmatulis> CBebop: so something else may be bound to port 80, please investigate
[02:41] <jeffesquivel> CBebop, can you try doing telnet localhost 80  on a terminal on your server?
[02:42] <CBebop> netstat -lntp showed nothing bound to port 80 is why I came here, I coulndt find much online but I most likely wasnt using the right words
[02:43] <jeffesquivel> CBebop, how about telnet localhost 80 ?
[02:44] <CBebop> connection refused
[02:45] <jeffesquivel> CBebop, hmm... do you have the error message provided by lighthttpd?
[02:45] <CBebop> one sec
[02:46] <CBebop> jeffesquivel: 2010-03-28 20:45:44: (network.c.345) can't bind to port: :: 80 Address already in use
[02:49] <jeffesquivel> CBebop, really strange
[02:49] <CBebop> I know =/
[02:49] <jeffesquivel> CBebop, what about lsof -i :80 ?
[02:50] <CBebop> Woa
[02:50] <CBebop> rtorrent 1369 cbebop  146u  IPv4  24225      0t0  TCP 192.168.1.141:56440->180.190.169.26:www (ESTABLISHED)
[02:50] <CBebop> That shouldnt be using that port at all
[02:50] <CBebop> I wrote the config file myself
[02:52] <CBebop> Killed rtorrent, re-ran the lsof command with no response and still port 80 lock from lighttpd startting
[02:52] <jeffesquivel> CBebop, yep, that seemed like an a connection to a remote 80 port
[02:52] <jeffesquivel> not the local one
[02:52] <CBebop> Ahh
[02:53] <FireCrotch_> CBebop: are you sure you don
[02:53] <CBebop> That would make sense, a tracker connection
[02:53] <FireCrotch_> 't have a rogue apache or something?
[02:53] <CBebop> Apache shouldnt be running, I disabled it on start (Ubuntu installed it when i upgraded from 9.04)
[02:53] <CBebop> its not in the process list at all
[02:54] <jeffesquivel> CBebop, could it be that you have the port used but on ipv6?
[02:54] <CBebop> Maybe, inadvertantly
[02:55] <CBebop> let me double check for ipv6 stuff in lighttpd
[02:55] <jeffesquivel> CBebop, I just read that lighthttpd first binds to port 80 on ipv4 and then on ipv6 addresses
[02:55] <jeffesquivel> CBebop, it may bind correctly on the first and then fail on the other one... and the error is the same
[02:57] <CBebop> I removed ipv6 from the config file and it started fine from the init.d script. I wonder why it just suddenly broke. Must have been an update to ligttpd that I missed.
[02:57] <jeffesquivel> CBebop, so it is working now/
[02:57] <jeffesquivel> ?
[02:57] <CBebop> I should pay closer attention. THank you very much for your help jeffesquivel and FireCrotch_
[02:57] <CBebop> Yeah
[02:57] <CBebop> Binds ipv4 just fine
[02:57] <jeffesquivel> CBebop, ok, great... I was running out of ideas here, he he :-)
[02:58] <CBebop> Must be a bug in the updated version from repositories
[02:58] <CBebop> that could be my only guess
[02:59] <jeffesquivel> CBebop, yep... if you get to the root of the problem, and it applies, please file a bug
[02:59] <jeffesquivel> well, gotta go... bbl
[03:37] <Roxyhart0> hi there somebody know any free software do do backup from windows/mac clients in a linux samba server?
[03:38] <aetaric> Roxyhart0: mac has native backup support
[03:39] <aetaric> and you can store them on samba servers by turning on a setting in the time machine prefs
[03:39] <Roxyhart0> mmm...i will chek it...thanks
[03:40] <aetaric> windows has a backup solution too. there is/was a backup program
[03:40] <aetaric> but you will have to setup scheduling
[03:40] <Roxyhart0> yes, i just was wondering if there are any free software client/server
[03:42] <aetaric> !info bacula
[03:42] <aetaric> maybe
[03:42] <Roxyhart0> thanks a lot! i will check it
[03:43] <aetaric> no problem
[04:14] <Sam-I-Am> hmmm
[04:14] <Sam-I-Am> anyone here notice ldap is STILL broken in lucid?
[04:15] <Sam-I-Am> glad to see no ones bothered to fix that in oh... 7 months since i reported the bug
[04:15] <twb> Isn't that because it's an important bug?
[04:15] <Sam-I-Am> yeah thats why it gets ignored :)
[04:15] <twb> Whereas some gnome icon being misaligned by 2px gets priority #1
[04:15] <Sam-I-Am> but i bet if there was a problem with EC it'd get some attention
[04:15] <twb> Ha
[04:16] <Sam-I-Am> no offense, but this being an LTS release you'd think people would want to fix these server-related bugs
[04:16] <Sam-I-Am> glad i'm not the only one who feels this way
[04:16] <Sam-I-Am> told my ubuntu lucid box to use ldap and now i'm locked out
[04:16] <Sam-I-Am> oh, and i cant even get into grub anymore because someone broke that too
[04:16] <twb> I'm only using Ubuntu because 1) apparently, the customers ask for it by name; and 2) it's still better than CentOS.
[04:16] <Sam-I-Am> so i had to boot off the damn CD
[04:16] <Sam-I-Am> well, hardy LTS was pretty damn solid
[04:17] <Sam-I-Am> had to backport a few things, but the general stuff worked
[04:18] <Sam-I-Am> oh, apparently i dont get to see what happens when my machine boots anymore... instead i get some splash screen which runs at what feels like 9600 baud on a framebuffer device :/
[04:19] <twb> There's apparently this "plymouth" crap that tries to start X during boot or something
[04:20] <Sam-I-Am> yep
[04:20] <Sam-I-Am> i tried removing it... and bad things happened
[04:20] <Sam-I-Am> none of the standard methods i know to turn my console back to 80x25 acually work
[04:20] <Sam-I-Am> either it gets ignored or it doesnt boot
[04:22] <twb> Oh, because it's loading vga16fb
[04:22] <twb> I had that problem, too.  In the end I removed the video card and used serial
[04:23] <twb> It's REALLY annoying, because on my hardware all it does is switch from an 80x25 I can read to an 80x30 I can't.
[04:25] <Sam-I-Am> twb: heh
[04:25] <Sam-I-Am> glad someone else here is feeling the love
[04:25] <Sam-I-Am> https://bugs.launchpad.net/ubuntu/+source/sudo/+bug/423252
[04:25] <Sam-I-Am> theres my bug... reported before karmic was released
[04:26] <twb> Sam-I-Am: use "files ldap" not "ldap compat".
[04:26] <twb> Sam-I-Am: does the problem magically go away?
[04:29] <Sam-I-Am> twb: ldap needs to be first...
[04:29] <twb> Sam-I-Am: why?
[04:29] <Sam-I-Am> so root can be in ldap
[04:29] <twb> Oh, I missed that point.
[04:29] <Sam-I-Am> yeah, security policy says centrally managed root accounts... except for an 'emergency' password
[04:30] <twb> I'm amazed that *ever* worked
[04:30] <Sam-I-Am> wasnt a problem until karmic
[04:31] <Sam-I-Am> and even then i can log in as root fine, its just none of my users can use sudo or su
[04:31] <twb> Have you taken this upstream, to Debian?
[04:31] <Sam-I-Am> so if you dont enable your root account in ubuntu before configuring ldap, you lock yourself out
[04:31] <Sam-I-Am> debian doesnt seem broken
[04:32] <twb> OK, so what is different between Debian and Ubuntu wrt. nss or sudo?
[04:32] <Sam-I-Am> i... wish i knew
[04:32] <twb> You can find out from the PTS pages
[04:32] <Sam-I-Am> best i can tell this is a regression problem with whatever version of glibc ubuntu uses
[04:32] <twb> e.g. http://packages.qa.debian.org/s/sudo.html
[04:32] <Sam-I-Am> yeah, i dont think this is sudo specific
[04:32] <twb> At the bottom of the page is an "ubuntu" section with a link to the diffs
[04:33] <twb> Sam-I-Am: so check nss the same way
[04:33] <Sam-I-Am> since the error message comes out of the kernel
[04:33] <Sam-I-Am> or something deep within
[04:33] <twb> Where's the kernel message there?
[04:34] <Sam-I-Am> this thing...
[04:34] <Sam-I-Am> setreuid(ROOT_UID, user_uid): Operation not permitted
[04:34] <Sam-I-Am> is something very deep in the OS
[04:34] <Sam-I-Am> since su gets the same error
[04:34] <twb> Um, that's not a kernel message (as in printk/dmesg)
[04:34] <Sam-I-Am> so not really kernel, but definitely base libraries
[04:35] <Sam-I-Am> yeh
[04:35] <twb> setreuid(2) is a system call
[04:35] <Sam-I-Am> yeah, thats what i meant
[04:35] <Sam-I-Am> so anything that uses it gets bonked
[04:35] <Sam-I-Am> which is why i dont think its specifically sudo or su related
[04:36] <twb> Is the UID of the root account in LDAP 0?
[04:36] <Sam-I-Am> yes
[04:36] <Sam-I-Am> and right now i dont even have a root account in ldap
[04:36] <Sam-I-Am> for testing
[04:36] <twb> Then I'd say you need to investigate pam, nss and/or libc.
[04:37] <Sam-I-Am> i'm trying to get libnss-ldapd working here
[04:37] <Sam-I-Am> which might eliminate nss
[04:37] <twb> nss will still be needed, but if it works then you know the fault is in libnss-ldap
[04:37] <Sam-I-Am> of course, i kinda hoped someone with a bit more experience than me with the ubuntu internals would have looked at this bug
[04:37] <twb> You could also try #openldap for ideas
[04:37] <Sam-I-Am> since it effectively renders lucid DOA for anyone with enterprise/centralized auth
[04:38] <twb> Well, here we do not put root in ldap, because that would be stupid.
[04:38] <Sam-I-Am> why?
[04:38] <twb> It would mean that a network failure would turn into a boot failure
[04:38] <Sam-I-Am> it should use the local account if ldap fails
[04:38] <Sam-I-Am> thats how it works here
[04:39] <twb> So you have *two* accounts with UID=0?
[04:39] <Sam-I-Am> yep
[04:39] <twb> Suggest you look at sash, which does something similar, but strictly within the flat files
[04:40] <twb> See if it is broken, maybe
[04:40] <Sam-I-Am> well, right now i have no root account in ldap
[04:40] <Sam-I-Am> so theres only the local account
[04:40] <twb> But yeah, the way I would do it would not be to have a root account with a password in LDAP, but rather to use ssh keys everywhere, with puppet to add/remove keys from /root/.ssh/authorized_keys
[04:41] <twb> Sam-I-Am: is nscd installed?  If so, it might be caching the LDAP account
[04:41] <Psi-Jack-> puppet?
[04:41] <Sam-I-Am> not using nscd
[04:41] <Sam-I-Am> it causes other problems in general
[04:41] <twb> Sam-I-Am: nod.
[04:42] <twb> Psi-Jack-: cfengine NIHd by ruby wankers
[04:42] <Sam-I-Am> although people have reported its a 'workaround'
[04:42] <twb> Sam-I-Am: I think they're on acid
[04:42] <twb> Installing nscd should BREAK things, not fix them.
[04:42] <Sam-I-Am> yep
[04:42] <twb> I suspect more likely is that they installed nscd and then broke it, but nscd kept it working for a little while
[04:48] <Sam-I-Am> meh
[04:52] <twb> Sam-I-Am: let me know if you figure it out
[04:52] <twb> I'll probably be migrating to 10.04 in about three years :-/
[04:53] <Sam-I-Am> heh
[04:56] <Sam-I-Am> apparently there is a debian bug for this...
[04:57] <Sam-I-Am> getting the same level of attention
[04:57] <Sam-I-Am> what'll be funny is if this is a glibc problem... and it eventually works its way into redhat
[04:57] <twb> Sam-I-Am: what ticket number?
[04:57] <Sam-I-Am> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=545414
[04:57] <twb> Sam-I-Am: Debian and Ubuntu run eglibc, which isn't glibc.
[04:57] <Sam-I-Am> yeah my bad
[04:58]  * twb subscribes to that ticket via the command-line.
[04:58] <twb> Pity nobody has ITPd those launchpad CLI tools to Debian yet.
[04:59] <ScottK> twb: launchpadlib and ubuntu-dev-tools are in Debian.  Which did you mean?
[05:00] <twb> ScottK: hmm, maybe that happened since last I looked
[05:00] <twb> ScottK: basically I want an exact equivalent to bts(1) for malone
[05:00] <ScottK> twb: I'd love to have one of those too.  I don't thin it exists.
[05:00] <ScottK> thin/think
[05:01] <ScottK> I'd particularly love bts show <bugnumber>
[05:08] <twb> I wrote one for roundup, which was fun.
[05:18] <twb> Bug #123456
[05:22] <twb> ScottK: do you need bts show -m, or just bts show?
[05:22] <ScottK> bts show
[05:23] <twb> http://paste.ubuntu.com/405677/ <-- barebones version
[05:24] <twb> Probably even just a .bashrc of: mal () { ${BROWSER:-sensible-browser} https://launchpad.net/bugs/$1; }
[05:24] <ScottK> Right.  Seems like it'd be nice for ubuntu-dev-tools.
[05:24] <twb> Eventually it should look something like http://code.haskell.org/~twb/ru/ru
[05:25] <ScottK> There are a number of nice one liners in there.
[05:26] <twb> I haven't bothered because it's easier to reproduce problems on Debian and use reportbug and bts.
[05:34] <ScottK> Right, I want it more for when I see a bug number somewhere but don't want to fire up a browser and manually hunt through the LP U/I.
[05:44] <twb> ScottK: so what should it do, if not fire up your browser?
[05:44] <ScottK> twb: It does.  It's just an easy way to fire it up going directly to the right URL without having to deal with navigatin LP or remembering the URL pattern.
[05:46] <twb> ScottK: OK, no worries.
[05:48] <twb> Ugh, ubuntu-dev-tools doesn't have manpages for some/all of its binaries.
[05:49] <twb> Haha, and ubuntu-iso crashes when I try to call it with --help, because genisoimage isn't installed :-)
[05:49] <Sam-I-Am> the fun never ends :)
[05:53] <osmosis> why does the lucid installer, when creating a software raid, create a separate /boot that is not on the raid?
[05:55] <twb> osmosis: there's a guided install option for software raid now?
[05:55] <osmosis> twb, not part of guided partitioning, but it does have it in the installer without having to drop to the CLI. Its part of the "manual partitioning" option.
[05:56] <twb> Um, manual partitioning is manual.
[05:56] <twb> It shouldn't be doing anything you don't tell it to
[06:10] <osmosis> twb, yah i see. it was guided LVM. not raid.
[06:12] <twb> osmosis: the bottom line is that bootloaders historically have trouble with LVM and md RAID (except RAID1).
[06:12] <twb> Even now, it is flaky to do it.
[06:23] <osmosis> twb, i might create a 4 disk raid1  for /boot  and then a 4 disk raid10 for root
[06:26] <twb> osmosis: what, are you hosting a database?
[06:27] <osmosis> twb, nah, just for fun. to push the limits. some qcow2 files
[06:27] <twb> Otherwise, s/raid10/raid5/ and that's what I do.
[06:27] <twb> 256MB RAID1 /boot, everything else is a RAID5 LVM, and then allocate 2 or 4GiB to /, $bignum for /home, and leave the rest of LVM unallocated.
[06:28] <osmosis> twb, actually testing the concept in a virtual machine with 4 virtio disks
[06:28] <osmosis> 4 qcow2 files to make a raid10
[06:57] <jeffesquivel> hi, which is the official ubuntu way to have HA?
[07:04] <twb> jeffesquivel: what does the ubuntu-serverguide say abuot it?
[07:06] <jeffesquivel> twb, sorry... asked to quickly, already on the Ubuntu's HA Team wiki, thanks anyway :-)
[07:07] <jeffesquivel> sometimes I just get excited about a topic and ask before doing my reading, he he
[07:47] <Ichat> could anyone plz help me with installing a headless torrent (client)  on   10.04 server   ebox.1.5
[07:48] <Ichat> im really kind of  Lost,
[07:49] <twb> headless, or GUI-less?
[07:50] <Ichat> webgui
[07:50] <twb> Last time I looked rtorrent was an acceptable bittorrent client, with an ncurses UI.
[07:52] <twb> Ichat: searching aptitude for tags "bittorrent" and "web" turns up torrentflux.
[07:52] <Ichat> its a home-server kind of system,  no monitor,   all goes via  ebox (or webmin),  the problem is,  that   for whatever reason,
[07:53] <Ichat> when i install   transmission daemon  (has a webui),    i cant get it to work outside of    localhost.
[07:54] <twb> Ichat: that's probably because transmission doesn't listen to anything but localhost by default, for security reasons.
[07:54] <twb> Ichat: find transmision's configuration file, and change that.
[07:55] <Ichat> in previous versions i had to change its config.json      in   /etc/transmission-daemon/settings.json     but now there is a settings file  in /etc/t ......mon/settings.json      and in   home/username/.config/settings.json     and in one other place
[07:56] <Ichat> the problem is,   what is the real file and what is not
[07:57] <Ichat> than i tried to change,  its setup params  in   /etc/default/transmission-daemon     to load a  config-dir  param  but it just plainly ignored that
[07:58] <Ichat> same prob with  8.04   9.10
[07:59] <Ichat> and i have no clue to what im doing wrong
[08:01] <twb> I don't use bittorrent, so I can't help you
[08:03] <Ichat> :) i wish i didn't either :$
[08:03] <Ichat> but its realy usefull for sharing lage files (sutch as  renderings (3d pictures)
[08:08] <twb> If you say so
[08:08] <twb> If it was strictly between two peers, I'd use socat or scp.
[08:36] <twb> I'll say this for Ubuntu: it's not SLES.
[08:50] <Omahn> Hurray for the Solaris 10 license change. Makes justifying Ubuntu *much* easier.
[08:57] <Jeeves_> :)
[09:16] <andol> Omahn: What license change?
[09:22] <joschi> andol: solaris 10 could be downloaded for free. oracle now changed the license, so you need to buy a support contract from them in order to use solaris 10. otherwise you need to delete it after 90 days
[09:23] <twb> Is sol10 the same as opensol10?
[09:23] <joschi> twb: no
[09:23] <twb> What's the difference?
[09:24] <joschi> twb: http://blogs.sun.com/jimlaurent/entry/faq_difference_between_opensolaris_solaris
[09:24] <twb> Thank you.
[09:26] <twb> Looks to be a similar breakdown to RHEL vs. Fedora.
[09:26] <joschi> twb: basically yes
[09:27] <twb> I'd go insane without dpkg/apt and a GNU userland.
[09:28] <twb> That and the QA from debian-policy/lintian/debbugs
[09:29] <lifeless> joschi: got a link to oracles announcement ?
[09:29] <lifeless> twb: there is nexenta, but it might be terrible, I don't know anyone that has used it in anger.
[09:30] <twb> I'm not touching nexenta until the dpkg and nexenta people settle their differences.
[09:30] <twb> If it was actually Debian GNU/kSol, I'd certainly try it.
[09:30] <joschi> lifeless: http://www.sun.com/software/solaris/popup.jsp?info=17
[09:30] <joschi> lifeless: "Please remember, your right to use Solaris acquired as a download is limited to a trial of 90 days, unless you acquire a service contract for the downloaded Software. "
[09:31] <joschi> lifeless: http://www.cuddletech.com/blog/pivot/entry.php?id=1120
[09:32] <andol> joschi: thanks
[10:08] <jiboumans> morning
[10:37] <maxagaz> hi
[10:41] <RoyK> hi all. this cloud thing, is that a replacement for general virtualization? if so, what does it use? kvm?
[10:53] <Cromulent> RoyK: no
[10:53] <RoyK> Cromulent, no what?
[10:54] <Cromulent> no it isn't a replacement for general virtualisation
[10:57] <twb> RoyK: cloud computing is a NIH reinvention of some of plan 9's functionality :-)
[10:57] <RoyK> Cromulent, then what is it, really? I don't get it...
[10:58] <twb> RoyK: it's a way to rent time on big iron, like you used to do back in the 60s.  The only difference is that now it's an HPC cluster instead of an actual big iron.
[10:59] <twb> RoyK: http://en.wikipedia.org/wiki/Cloud_computing
[10:59] <RoyK> well, it sounds like virtualisation to me
[10:59] <RoyK> I've read that
[11:00] <twb> RoyK: it leverages virtualization, but for a different goal.
[11:00] <twb> If you have one server and one VM, it would be meaningless to deploy cloud infrastructure.
[11:01] <twb> The point of cloud computing is when you have >>1 hardware node and >>1 VM.
[11:01] <RoyK> well, of course :)
[11:02] <RoyK> I was just thinking if cloud in-house computing could be a neat replacement for something like XenServer
[11:02] <RoyK> we have like five machines with 30 VMs
[11:03] <twb> That really depends on whether the benefit of load balancing VMs across the HNs is worth the hassle of configuring it.
[11:03] <RoyK> sorry, what's HNs?
[11:04] <Cromulent> hardware nodes
[11:04] <RoyK> ok
[11:04] <maxagaz> how to get the uuid of a device ?
[11:04] <twb> The things with the wires and electrickery.
[11:04] <RoyK> can ubuntu clouds failover a VM from one HN to another?
[11:04] <twb> maxagaz: udev has a thingy
[11:04] <twb> RoyK: that's the *point* of a cloud
[11:04] <RoyK> k
[11:05] <twb> RoyK: without that, it'd just be virtualization
[11:05] <RoyK> so what is it using for virtualisation? kvm?
[11:05] <twb> RoyK:
[11:05] <twb> RoyK: NFI.  Probably whatever you want.
[11:06] <Cromulent> the real benefit of clouds hasn't yet been realised on ubuntu yet imo which is having instances span multiple hardware nodes
[11:07] <maxagaz> twb, udevadm info --query=all doesn't work
[11:08] <RoyK> Cromulent, I guess you'll need a pretty cool infiniband network in the back for that to be efficient
[11:09] <maxagaz> for example, how to get the uuid of all my device ?
[11:09] <maxagaz> like, sda1 uuid, sda2 uuid...
[11:09] <Cromulent> RoyK: not really it just makes development of distributed applications somewhat simpler - if you present the resources to the application as one machine yet have the distribution handled in the background
[11:09] <maxagaz> twb, ok, I found this : blkid
[11:10] <Cromulent> most computationally heavy distributed apps work over the internet currently anyway
[11:11] <Cromulent> its just removing the setup from the application developer and putting it in the OS developers hands / cloud layer developers hands
[11:11] <Cromulent> that is why you have technologies such as OpenMP
[11:11] <Cromulent> or MPI rather
[11:12] <maxagaz> what does the last '2' means in fstab: "/dev/sdb1 /mnt/extra ext3 defaults 0 2" ?
[11:13]  * RoyK throws maxagaz a "man fstab"
[11:18] <RoyK> hm. ic. is it possible to have redundant cloud controllers as well?
[11:21] <Cromulent> probably - I don't run my own cloud though so you'll be best off waiting for someone else to confirm - it would seem like a large oversight to not allow them though if you ask me
[11:33] <nucc1> how can i set my server's hostname permanently?
[11:38] <jeffesquivel> nucc1, change it on /etc/hostname then on /etc/hosts then $ sudo hostname new_hostname
[11:39] <vishalbelsare> I have a question regarding EBS. If I have use a Ubuntu EBS Boot AMI, i.e. the root disk is EBS based, do I incur the IO charges for EBS just for using that instance too? or is this applicable only to EBS volumes mounted on a running instance?
[11:48] <nucc1> jeffesquivel, thanks
[11:50] <jeffesquivel> nucc1, no problem
[12:09] <bronto2> i think i somehow unistalled part of ssh server, cant connect to the remote machine, any solutions other than calling the main admin?
[12:10] <jeffesquivel> bronto2, do you have any other remote access service (telnet, vnc, etc.) ?
[12:11] <bronto2> jeffesquivel, is there something thats enabled by default on ubuntu-server? (only ftp is working)
[12:11] <jeffesquivel> bronto2, don't think so
[12:11] <nucc1> jeffesquivel, are you familiar with postfix?
[12:11] <jeffesquivel> bronto2, are you sure you messed up your ssh? or could it be that your current ip is being blocked by something like fail2ban or denyhosts?
[12:12] <bronto2> ok, now if i have to mail the main admin, what is the command to reinstall the entire ssh thingy?
[12:12] <jeffesquivel> nucc1, it has been a long time since I had to admin a postfix server
[12:12] <jeffesquivel> nucc1, why?
[12:12] <bronto2> jeffesquivel, i belive its ssh proble, this is intranet
[12:12] <nucc1> jeffesquivel, my problem may be in postfix, or it may be dns. let me explain.
[12:12] <bronto2> *problem
[12:13] <jeffesquivel> bronto2, ok then
[12:13] <nucc1> jeffesquivel, i have a domain hosted on a shared hosting platform, and a linode node to which there is an A-record pointing to a web app running there.
[12:13] <bronto2> sudo apt-get remove ssh && sudo apt-get install ssh should fox things?
[12:13] <bronto2> *fix
[12:14] <nucc1> jeffesquivel, the problem is this: when the linode instance tries to send a message to addresses on the shared-hosted domain, it gets rejected with "sender verification failed"
[12:14] <nucc1> jeffesquivel, would it help if i told you the domain and subdomain?
[12:15] <jeffesquivel> bronto2, you can try sudo apt-get --reinstall install ssh
[12:16] <bronto2> jeffesquivel, thanks
[12:16] <jeffesquivel> bronto2, no problem
[12:16] <jeffesquivel> nucc1, any more information on what type of verification?
[12:17] <nucc1> sender verification
[12:17] <nucc1> i can paste the log
[12:17] <jeffesquivel> nucc1, I mean, IIRC there are several methods to do sender verification
[12:17] <jeffesquivel> nucc1, ok
[12:18] <nucc1> jeffesquivel, lemme pm you the link
[12:22] <jeffesquivel> nucc1, hmm... let me see if I understand correctly
[12:22] <jeffesquivel> nucc1, you have a web app installed on the same server as your postfix
[12:22] <nucc1> yes.
[12:22] <jeffesquivel> nucc1, and you want your webapp to be able to send e-mail msgs
[12:23] <nucc1> it sends the messages alright.
[12:23] <nucc1> jeffesquivel, the setup is like this
[12:23] <nucc1> the main website is on domain.com
[12:23] <nucc1> the webapp is on a linode virtual server: node.domain.com
[12:23] <nucc1> when the web app tries to send email to an address@domain.com , the email gets rejected
[12:24] <nucc1> jeffesquivel, for the reason in the log.
[12:24] <nucc1> jeffesquivel, other email servers seem to work fine. i've tested Yahoo and Gmail.
[12:25] <jeffesquivel> nucc1, I see...
[12:26] <jeffesquivel> nucc1, well the error message you are receiving (550) usually means that the user you are trying to send mail to doesn't exist (at least not on that domain)
[12:26] <nucc1> jeffesquivel, i think it says that the user I am sending the mail as cannot be verified
[12:27] <nucc1> and for the life of me, i don't know how to make my postfix instance reply acknowledge the verification requests
[12:27] <jeffesquivel> nucc1, well... you see the part where it says: "550 Sender verify failed (in reply to RCPT TO command"
[12:28] <jeffesquivel> RCPT TO is the SMTP command which defines the recipient not the sender
[12:28] <jeffesquivel> nucc1, you're getting bounced when you specify when you want to send the message to not when you say who you want to send it as
[12:29] <merlijn-> quick question, for how long will the 9.10 server release be supported?
[12:29] <nucc1> merlijn-, 18 months, i guess
[12:29] <jeffesquivel> s/when you want/who you want/
[12:29] <merlijn-> 18 months? desktop even gets 2 years
[12:30] <nucc1> merlijn-, lts releases are the ones that get 2 years.
[12:30] <nucc1> merlijn-, that is, 8.04, 10.04
[12:30] <merlijn-> lts server get 5 years
[12:30] <merlijn-> I found that here: https://wiki.ubuntu.com/
[12:31] <nucc1> jeffesquivel, is it possible that the domain.com mail server rejects the email with a false message? cos the addresses exist.
[12:31] <nucc1> i just sent my gmail address an email with the same postfix config.
[12:33] <nucc1> jeffesquivel, also, i get confused where it says "status verification failed for root@..." cos that is the user I am sending the email as
[12:34] <nucc1> merlijn-, lts releases get 5 years, non-lts releases get 18 months.
[12:36] <jeffesquivel> nucc1, hmm... I just read a little bit more about that error message
[12:36] <jeffesquivel> nucc1, it is also used to deny relay access
[12:37] <nucc1> jeffesquivel, i don't know why it tries to relay...
[12:37] <jeffesquivel> nucc1, which means, the machine your sending the mail to thinks you want to use it as a relay server and is bouncing you
[12:37] <nucc1> jeffesquivel, is what you read on the net?
[12:37] <jeffesquivel> nucc1, yes...
[12:38] <nucc1> where's the link?
[12:38] <nucc1> jeffesquivel, does that mean that i'm effectively hopeless?
[12:39] <jeffesquivel> nucc1, http://www.eudora.com/techsupport/kb/1593hq.html
[12:39] <jeffesquivel> nucc1, don't think so... gimme a sec
[12:43] <jeffesquivel> nucc1,
[12:43] <jeffesquivel> I just tried to do the same thing your server is doing but manually (telnet to port 25)
[12:43] <nucc1> yea
[12:43] <jeffesquivel> this is the message I get
[12:43] <jeffesquivel> 550-Message rejected because (vertex.tellantservices.com) [201.191.183.183] is
[12:43] <jeffesquivel> 550 blacklisted.
[12:43] <nucc1> whew!
[12:43] <jeffesquivel> but now, that's my IP
[12:43] <jeffesquivel> you should try from your server to see if that's the case for your IP too
[12:44] <jeffesquivel> most probably, it is
[12:44] <nucc1> jeffesquivel, vertex.tellantservices.com is not 201.191.183.183 !
[12:45] <jeffesquivel> nucc1, I know...
[12:45] <jeffesquivel> nucc1, that's why I told you that you should try from your server
[12:46] <jeffesquivel> nucc1, most probably your server's ip is blacklisted too
[12:46] <nucc1> hey, how do i specify recipienet address?
[12:47] <jeffesquivel> RCTP TO:
[12:47] <jeffesquivel> nucc1, http://www.earthinfo.org/example-smtp-conversation/
[12:47] <nucc1> lol. too many unrecognized commands. connection closed... lemme google for a quick tutorial :d
[12:50] <nucc1> jeffesquivel, here's what it told me: rcpt to: fanen.ahua@tellantservices.com 550-verification failed for <fanen@vertex.tellantservices.com>
[12:51] <nucc1> no such user here. sender verify failed.
[12:51] <jeffesquivel> nucc1, well.. good news, you're not blacklisted... it's just me :-)
[12:51] <nucc1> jeffesquivel, i'd have been puzzled, cos its a relatively fresh install :d
[12:52] <nucc1> with a new IP :d
[12:53] <nucc1> jeffesquivel, the message is quite confusing though. it says 550-verification failed for fanen@vertex.tellantservices.com
[12:53] <nucc1> which to me means that vertex.tellantservices.com did not acknowledge that it knew such a user
[12:54] <jeffesquivel> nucc1, did you used fanen@vertex... as your MAIL FROM:  ?
[12:54] <nucc1> jeffesquivel, yes.
[12:55] <jeffesquivel> nucc1, what happens if you use a different domain on your MAIL FROM:?
[12:55] <nucc1> jeffesquivel, let me use the main domain now.
[12:55] <jeffesquivel> I mean something which doesn't has tellantservices.com in it
[12:56] <nucc1> ha ha. ok. tellantservices.com just told me 451 temporary local problem... please try later
[12:58] <nucc1> jeffesquivel, it says 550 currently not permitted to relay through this server
[12:58] <jeffesquivel> nucc1, really weird... you changed just the MAIL FROM not the RCTP TO, right?
[12:59] <nucc1> jeffesquivel, yeap
[12:59] <nucc1> wonder why it doesn't say that for tellantservices.com domains.
[12:59] <nucc1> i mean vertex.tellantservices.com
[12:59] <nucc1> and since it's a shared host, our options seem quite limited
[13:00] <jeffesquivel> nucc1, the weird thing is that tellantservices.com shouldn't be trying to relay anything when receiving mail to @tellantservices.com addresses
[13:00] <nucc1> so it's a craptastic mail config on a shared host. explains why some people despise shared hosts...
[13:00] <nucc1> and leaves me unable to explain why stuff isn't working.
[13:01] <nucc1> :(
[13:01] <jeffesquivel> nucc1, yeah, I don't like not being able to understand stuff... he hehe
[13:02] <jeffesquivel> nucc1, hmm... let me try something from a different IP... gimme a sec
[13:02] <nucc1> ok. i'm here :)
[13:08] <jeffesquivel> nucc1, so I guess you have previously send an e-mail to fanen.ahua@tellantservices.com from some other service and it works ok, right?
[13:09] <jeffesquivel> actually, I should be able to try that
[13:10] <jeffesquivel> nucc1, yep... just as I thought
[13:10] <jeffesquivel> nucc1, tellantservices.com is misconfigured
[13:10] <nucc1> jeffesquivel, that is my actual email, which works.
[13:10] <jeffesquivel> nucc1, I cannot send e-mail to it even from my regular gmail account
[13:11] <jeffesquivel> nucc1, I just tried to send a message and it bounced me with the same message we are getting
[13:12] <nucc1> jeffesquivel, hmmm. maybe i should propose we close that account. use google apps for email, and host the website on linode altogether.
[13:12] <nucc1> oh boy.
[13:13] <jeffesquivel> nucc1, sounds like a good idea... he he
[13:15] <jeffesquivel> nucc1, did you tried to send an e-mail to that account from another account? did it worked?
[13:19] <jeffesquivel> nucc1, anyway, back to your original problem... I think it may be possible that tellantservices.com thinks it is also the mail server for vertex.tellantservices.com (don't know if that's even possible, I'm just guessing)
[13:19] <nucc1> jeffesquivel, we've been using it as office email,
[13:19] <nucc1> jeffesquivel, yes, i suspect that. but there's no way to fix that...
[13:19] <nucc1> that i can think of :(
[13:20] <nucc1> i'm actually looking at the pricing for google mail now
[13:20] <jeffesquivel> nucc1, yep... agree with you on that one
[13:21] <nucc1> jeffesquivel, googleapps wants to sell you the whole shebang for $50/user/year. which is just overkill.
[13:21] <jeffesquivel> nucc1, but it is free if you have less than 50 users
[13:22] <nucc1> jeffesquivel, it doesn't say so. i know it used to be that way.
[13:24] <jeffesquivel> nucc1, gimme a sec
[13:25] <nucc1> jeffesquivel, i really appreciate your time and assistance
[13:25] <jeffesquivel> nucc1, http://www.google.com/apps/intl/en/group/index.html
[13:26] <jeffesquivel> nucc1, I think that's the link to the standard edition which should still be free
[13:26] <nucc1> it was only showing me the business page :)
[13:29] <nucc1> yes, seems to.
[13:29] <nucc1> i'll recommend a move in the next meeting.
[13:31] <jeffesquivel> nucc1, ok
[13:44] <tyska> hello guys
[13:44] <tyska> someone here use UEC?
[13:44] <tyska> i wanna know how can i set Real IP Addresses to virtual machine. is this possible
[13:44] <tyska> ?
[13:47] <tyska> hello? someone is there?
[14:17] <zul> soren: vmbuilder in lucid seemed to be fubared
[14:30] <tyska> hello guys
[14:30] <tyska> someone who uses UEC is online?
[14:53] <ttx> tyska: yes
[14:55] <tyska> ttx: how can i assign a real ip address to a vm instance of my cloud? is this possible?
[15:01] <ttx> tyska: define "real IP address"
[15:01] <ttx> tyska: public IPs are assigned through VNET_PLUBLICIPS in eucalyptus.conf
[15:02] <ttx> I mean, VNET_PUBLICIPS
[15:04] <tyska> ttx: i need a IP that is reachable from any computer in the Internet, and not just the computers on the subnet of a LAN
[15:04] <tyska> ttx: did you understand?
[15:06] <ttx> tyska: the CC will assign public IP addresses from its public-facing network interface. If that is connected directly to the Internet, then it will serve reachable addresses...
[15:06] <ttx> Otherwise you'll have to do some NAT
[15:07] <ttx> so that the address from the VM can be reached through another one
[15:08] <tyska> and with nat i just put the NAT rules on the CC right?
[15:09] <RoyK> can I mix amd and intel machines in a cloud, possibly at different speeds, and the cloud will balance it all as best as it can?
[15:10] <ttx> RoyK: by cloud you mean... UEC ? EC2 ?
[15:10] <ttx> tyska: again, depends on where you put your CC. The machine doing NAT needs to be reachable
[15:10] <RoyK> uec - that is - I thought they were the same
[15:11] <tyska> yes, understand
[15:11] <ttx> RoyK: you can mix, however, eucalyptus counts in "cpus", it does not look into how fast they go
[15:12] <RoyK> ok, so tossing in an old P4 1,8 might be disasterous ;)
[15:12] <ttx> RoyK: so if you have a very slow quad-core and a very-fast 8-core, it will still see 12 "cpus"
[15:12] <ttx> RoyK: I think so, haven't tried :)
[15:12] <RoyK> ic
[15:12] <FF666> hi
[15:13] <FF666> is this the chanel for ubuntu eucaliptus?
[15:13] <edmoore> greetings. I hope this question does not stretch the scope of this channel too much: We've a bunch of old boxes lying around for our small research group - a webserver, a fileserver, email, some other bits and bobs. Someone advised me to consolidate them all into a bunch of virtual machines on a single box, and that seems to be The Done Thing (tm) nowadays. But I don't understand why, for our needs, just running each of those things natively di
[15:13] <edmoore> on a single ubuntu server wouldn't be a better idea?
[15:13] <ttx> FF666: yes, there is also #ubuntu-cloud, but here is alright
[15:13] <FF666> ok
[15:14] <ttx> edmoore: segmenting is good, so that an upgrade of one doesn't kill the other.
[15:14] <ttx> edmoore: but in the end it's a question of how much resource you can throw at it
[15:15] <ttx> since running them in multiple VMs will use more resources
[15:15] <RoyK> what sort of storage do you guys use for eucaliptus?
[15:15] <edmoore> we might be able to stretch to a new dell r710 with a couple of xeons and maybe 12GB of ram. It'll have a long upgrade path hopefully.
[15:15] <ttx> edmoore: VMs have another advantage, you can move them around
[15:16] <edmoore> so I guess the real advice is: unless you actually should virtualise (eg our actual number crunching machines), then you should really consider virtualising?
[15:16] <edmoore> sorry typo: unless you actually *shouldn't* virtualise, you should
[15:17] <ttx> edmoore: I'd say that, yes (but I'm biased)
[15:17] <RoyK> is virtualising number crunching machines really a good idea?
[15:17] <RoyK> I mean, is the overhead so low it makes sense?
[15:17] <edmoore> RoyK: that was a typo, i meant to say unless you actually *shouldn't* (eg our crunching machines), you should
[15:18] <RoyK> heh
[15:18]  * RoyK has no plans on virtualising the number crunching machines here
[15:18] <edmoore> RoyK: me neither!
[15:18] <edmoore> well, the crunchers are actuially run by sysadmins who know what they're doing
[15:19] <edmoore> ok that's sound advice, thank you very much.
[15:22] <FF666> is there any site where I can download an image of ubuntu server with eucaliptus installed?
[15:25] <ttx> FF666: not that I know of. Esepcially as it requires at least two machines at this point.
[15:25] <ttx> FF666: however installing it from 10.04 beta 1 ISO is not very difficult.
[15:26] <FF666> ttx:??, I'll use a virtual machine, so the number of computers is not a problem
[15:27] <FF666> I want to know if there is a site where I can find images of ubuntu, like vwmare images
[15:28] <ttx> FF666: well, running a virtualization node inside a VM will be... problematic. virt-in-virt isn't working so well.
[15:28] <ttx> FF666: you want an Ubuntu server image ? Or an Ubuntu-Server-preinstalled-with-Eucalyptus image ?
[15:29] <FF666> ttx: the second one
[15:29] <ttx> FF666: ok, we don't have that.
[15:30] <FF666> ttx: ok
[15:30] <FF666> ttx: but the performance is not a problem, I'm doing a proyect for college
[15:30] <FF666> ttx: I've to deploy an mysql on the cloud
[15:35] <RoyK> seems to me UEC might be a good replacement for this XenServer stuff we have now
[15:35] <RoyK> but again - what sort of storage would you recommend?
[15:37] <mdeslaur> smoser: so, does kvm support OVF?
[15:37] <smoser> its kind of not the right question to ask
[15:37] <smoser> ovf is really not something that KVM would support
[15:37] <mdeslaur> smoser: ok. uhh...do you like cake?
[15:37] <smoser> it would be supported more at libvirt or virt-convert level
[15:38] <mdeslaur> smoser: oh, is ovf just a description file? I thought it had the disk image also?
[15:38] <smoser> yeah, wouldn't that be easy :)
[15:38] <smoser> ovf is a description file.
[15:39] <smoser> and it does define the type of each disk image that is present in the OVA (archive file)
[15:39] <smoser> but makes no restrictions on what that type is
[15:39] <mdeslaur> oh, I see
[15:39] <smoser> yeah.
[15:40] <smoser> so the end goal of "heres the ovf, it can be run on any hypervisor"
[15:40] <smoser> is , well, not easily realized at the moment.
[15:40] <mdeslaur> my google search for OVF led me to some guy called "Scott Moser <smoser@us.ibm.com>", but he's apparently not answering his email any more :)
[15:41] <jcastro> I don't think that guy is trustworthy tbh
[15:41] <mdeslaur> jcastro: I've already complained to IBM management :)
[15:41] <smoser> mdeslaur, he's not worth anything
[15:42] <smoser> i'd not bother with him
[15:42] <smoser> :)
[15:42] <mdeslaur> hehe
[15:47] <ruben23> hi, how do memory testing and checking on my ubuntu-server..?
[15:50] <soren> ruben23: Reboot, choose the memtest86+ option in grub.
[15:51] <ruben23> soren: how do i check i i have memtest, i havent installed it yet.
[15:52] <soren> ruben23: It's installed by default.
[15:52] <ruben23>  soren: how do i select it on startup, i havent seen it..on startup
[15:52] <soren> ruben23: It should be in your grub menu.
[15:53] <ruben23> how do i access grub menu, something i need ot press..?
[15:58] <sherr> ruben23
[15:58] <sherr> Sorry - ESC for grub2
[15:59] <elb0w> Anyone use UEC?
[15:59] <elb0w> had some questions about it
[15:59] <zul> ttx: do you want me to take care of the phpsysinfo one?
[16:00] <ttx> zul: sure, though at this point beta2-targeted bugs and papercuts have precedence.
[16:00] <ttx> elb0w: shoot
[16:01] <zul> ttx: okie dokie
[16:01] <elb0w> im looking for redundancy options and failover
[16:01] <elb0w> it looks like it doesnt have any?
[16:02] <ttx> elb0w: there is some redundancy when using multiple clusters and multiple node controllers, but not something I'd really call redundancy (or failover).
[16:04] <elb0w> so
[16:05] <elb0w> I always pictured  a clouds (I know the term is overused) purpose was that if something goes down something else picks up the slack
[16:05] <elb0w> Is all the UEC is meant to do is spawn VM's?
[16:07] <ttx> elb0w: cloud computing is really "utility computing", power at your fingertips. That doesn't imply redundancy. That's what clusters do. However, redundancy so that the VMs you started actually are guaranteed to run is a welcome addition.
[16:07] <elb0w> so UEC is not HA then I take it
[16:07] <ttx> elb0w: foer example, Amazon EC2 reserves the right to terminate an instance of yours anytime (last time I read the fineprint)
[16:08] <ttx> elb0w: at this point, now.
[16:08] <ttx> no, even.
[16:08] <elb0w> ok
[16:10] <elb0w> thanks ttx
[16:20] <zul> ttx: the bacula "papercut" is not exactly a "papercut" its more of a feature
[16:24] <ttx> zul: right, wanted to see if Ante already had the fix in his PPA or not
[16:25] <zul> dont think so
[17:43] <RoyK^> hi all
[17:44] <RoyK^> what do you guys use for uec storage?
[17:47] <RoyK^> hum... "64-bit can run both i386, and amd64 instances; by default, Eucalyptus will only run 1 VM per CPU core on a Node" <-- only one VM per core by default seems _very_ low
[17:47] <RoyK^> I mean, overbooking by 4-10 is common
[17:52] <tyska> hello
[17:53] <tyska> someone can help me with UEC?
[17:53] <RoyK^> heh - trying to read up about it myself
[17:53] <tyska> read up where?
[17:53] <RoyK^> https://help.ubuntu.com/community/UEC
[17:54] <tyska> this how-to doesn't help with problems solution
[17:54] <tyska> for example, now im trying to find the file where is set the IP of the Cloud Host
[17:54] <RoyK^> well, I think it might be better to ask about the problem ...
[17:54] <tyska> but there is no documentation about this
[17:55] <RoyK^> the node controller?
[17:55] <RoyK^> or the nodes?
[17:59] <tyska> eucalyptus.conf
[18:00] <tyska> but when i try to run a instance of vm i get this error:
[18:00] <tyska> FinishedVerify: Not enough resources: vm instances.
[18:02]  * genii temporarily removes "coffee" from his highlight list
[18:03] <RoyK^> genii += 0xc0ffee
[18:03] <genii> RoyK^: :)
[18:05] <jeffesquivel> hi, I want to start learning about virtualization on Ubuntu Server, any ideas of what a reasonable hardware platform for testing would be?
[18:07] <RoyK^> jeffesquivel: the tutorial is quite clear about that https://help.ubuntu.com/community/UEC/PackageInstall
[18:08] <RoyK^> that's for cloud, though. if you plan to setup a vm or two with virtualbox or something, you don't need that amount of iron
[18:08] <RoyK^> brb
[18:09] <sherr> jeffesquivel: virtualisation or UEC cloud? You can test virt. on a laptop
[18:09] <jeffesquivel> RoyK^, I don't want to get into cloud just yet... I was thinking more like a ubuntu hypervisor (is that the correct term?) running on a server with probably two or three VMs running concurrently
[18:10] <jeffesquivel> sherr, virtualization
[18:11] <jeffesquivel> sherr, so is it ok for me to just buy the cheapest VT-enabled CPU and just use that?
[18:12] <jeffesquivel> I'm asking because I thought maybe the cheapest CPUs don't have some feature which would be nice to include in my testing...
[18:14] <RoyK^> you need VP
[18:15] <sherr> VT
[18:15] <RoyK^> ermm
[18:15] <RoyK^> yes
[18:15] <RoyK^> but I guess even Atom has that now
[18:15] <sherr> You don't need VT if you use Xen. But Ubuntu Xen support is poor.
[18:15] <sherr> You had best get it ... for options.
[18:16] <RoyK^> huh? UEC uses xen, right?
[18:16] <sherr> KVM, VirtualBox need need it (in general anyway)
[18:16] <RoyK^> virtualbox doesn't need VT
[18:16] <sherr> Amazon use Xen. But I think UEC is agnostic - but generally KVM mainly
[18:16] <RoyK^> kvm needs it
[18:16] <sherr> virtualbox - are you sure?
[18:17] <RoyK^> positive
[18:17] <RoyK^> it's even got a small icon telling if the cpu has it or not
[18:18] <RoyK^> but I read somewhere that xen was what made uec possible, or something
[18:18] <RoyK^> default install of a node, is that kvm?
[18:19] <sherr> RoyK^: I have not looked at UEC much. Interested in virt. but less so in "cloud".
[18:20] <RoyK^> afaics UEC is just virtualisation gone large
[18:20] <sherr> My impression was that UEC is based off eucalyptus which is an Amazon compatible AWS layer. But a layer that hides the implementation.
[18:21] <RoyK^> sherr: what are you virtualising? servers or general desktop/testing?
[18:21] <sherr> And Ubuntu's primary ikplementation is KVM. Xen has lost a lot of raction in Ubuntu, Fedora etc.
[18:21] <sherr> I am doing servers primarily.
[18:22] <sherr> And looking to use Xen because most of the available h/w I have available (at work) has no h/w VT.
[18:22] <RoyK^> how many VMs?
[18:22] <sherr> I have not had such good experiences with KVM either, performance wise.
[18:23] <sherr> RoyK^: Not cloud level :-) two or three maybe. I haven't got things running at work yet.
[18:23] <RoyK^> sherr: try virtualbox
[18:23] <sherr> Currently, I have a vbox vm on a desktop at home.
[18:23] <RoyK^> I use that on my primary (private) opensolaris server with three VMs
[18:24] <sherr> Well - 3 actually. 2x lenny, 1x centos (work staging web server)
[18:24] <sherr> VBox is great, works well.
[18:24] <jeffesquivel> RoyK^, is it possible to remotely manage virtualbox?
[18:24] <RoyK^> jeffesquivel: sure
[18:24] <sherr> Yes. VBoxManage is CLI.
[18:24] <RoyK^> jeffesquivel: the OSS version lacks RDP, though
[18:24] <jeffesquivel> RoyK^, so... I should be able to install VB even withouth X running on a server?
[18:24] <RoyK^> jeffesquivel: but download it form sun and you get RDP console access to each VM, and the rest can be managed through the command line or VirtualBox
[18:25] <RoyK^> without X on the server
[18:25] <tyska> hello guys
[18:25] <RoyK^> just remote X whenever you start VirtualBox
[18:25] <tyska> someone can help me with UEC?
[18:25] <jeffesquivel> RoyK^, sherr, I always thought of VB as more a "desktop virtualisation solution"
[18:25] <RoyK^> tyska: ask the question, not whether people can help you or not
[18:25] <RoyK^> jeffesquivel: but you can create VMs without X too
[18:26] <RoyK^> jeffesquivel: I think that was the initial idea, but it works well for servers too
[18:26] <tyska> RoyK^: i need to explain my problem, if no one is 'listening' i will just waste my time
[18:26] <sherr> jeffesquivel: that's because it has a GUI (if desired) that works well. And makes it easy to use.
[18:27] <RoyK^> tyska: it usually takes one (long) line - try it
[18:27] <sherr> People look down on "easy" :-) But time is short sometimes.
[18:27]  * RoyK^ likes easy systems
[18:28] <RoyK^> although I think kvm or xen might scale better than vbox
[18:28] <jeffesquivel> sherr, RoyK^: ok, I'll look into it too... even so, I prefer KVM + libvirt if possible (because of it being 100% FLOSS)
[18:28] <RoyK^> but then, since this box is running opensolaris, nothing else really works
[18:28] <Sripa> I am unable to upload my kernel using euca-upload-bundle command
[18:29] <RoyK^> jeffesquivel: it's your choice entirely :)
[18:29] <Sripa> It checks for the bucket name, and never returns..
[18:29] <Sripa> What might be the problem ?
[18:29] <tyska> when i run euca-describe-availability-zones verbose nothing happens
[18:31] <jeffesquivel> RoyK^, sherr: I read sometime ago that there is a feature in newer CPUs which does the virtual to physical page translation in hardware and should improve performance noticeably... do you know anything about it?
[18:31] <tyska> i think this problem happened because i had two NIC connected to a HUB that is connected to a DHCP server, in the installation the two NIC'S get dynamic IP, but later i change the configuration of one NIC to static IP
[18:31] <tyska> to do a LAN with the nodes in another subnet
[18:33] <tyska> now i've fixed the IP's configuration, but now i have 0000/0000 no free/max on the euca-describe-availability-zones verbose command
[18:33] <tyska> someone can help me?
[18:34] <sherr> tyska: Sorry, no knowledge of UEC.
[18:34] <Sripa> sherr: got any idea abt mine ?
[18:34]  * sherr thinks UEC might need better support and docs soon. Maybe a seperate channel.
[18:34] <Sripa> I use eucalyptus installed on centos 5
[18:35] <sherr> Sripa: Sorry - I know little about UEC *or* eucalyptus.
[18:35] <tyska> Sripa: see ubuntu-cloud
[18:35] <tyska> Sripa: see #ubuntu-cloud
[18:35] <sherr> I am vaguely interested, but life is short and I have no business case.
[18:36] <Sripa> tyska: yeah but that is specific to ubuntu right ?
[18:36] <sherr> #ubuntu-cloud - all 5 people .. :-)
[18:36] <tyska> UEC really needs a better documentation
[18:36] <sherr> I would assume ubuntu-cloud and UEC and euca are all relevant in that channel
[18:37] <tyska> ubuntu enterprise cloud
[18:37] <sherr> But this channel seems to turn into a UEC/Cloud channel a lot anyway ...
[18:37] <RoyK^> jeffesquivel: I thought that was VT
[18:49] <jeffesquivel> RoyK^, I think it may be called IOMMU
[18:49] <jeffesquivel> RoyK^, but not sure if that was what I read
[18:49] <jeffesquivel> about
[18:49] <RoyK^> I guess I'll just have to setup an ubuntu test cloud to see what it can do
[18:50] <RoyK^> storage on opensolaris seems like a good idea
[18:50] <jeffesquivel> well, gotta go
[18:50] <RoyK^> don't want non-snapshotting filesystems anymore
[18:50] <jeffesquivel> bbl
[18:50] <jeffesquivel> RoyK^, sherr, thanks for all the help!
[18:59] <tyska> if i run the command to discover the nodes, but the IP address configuration of the server was wrong, what i need to run for fix the credentials in the node????
[19:29] <tyska> i get to create a instance
[19:29] <tyska> but now i cant connect
[19:29] <tyska> im getting this message: Permission denied (publickey).
[19:30] <tyska> someone now how to solve this?
[19:30] <tyska> someone know* how to solve this?
[19:31]  * tyska waiting response
[19:32] <smoser> ok, for the sake of discussion, lets pretend that someone wanted what was provided by http://mod-auth-shadow.sourceforge.net/
[19:32] <smoser> and they were using some old debs (probably from hardy)
[19:32] <smoser> what would be the recommended way to do "auth shadow" ?
[19:34] <smoser> i'm guessing its pwauth / libapache2-mod-authnz-external
[19:49] <sherr> smoser: no idea - but that would seem a reasonable solution. For a start, it is recommended from the m-a-shadow page.
[19:49] <smoser> sherr, thanks. yeah, i think that is right.
[19:50] <sherr> It is also maintained - security patches etc.
[20:07] <zul> smoser: thats what I would do
[20:10] <guptaxpn> hello
[20:10] <guptaxpn> anyone in here?
[20:10] <guptaxpn> what's the deal with the virtual machine install option?
[20:11] <guptaxpn> how is it different than a normal server installation?
[20:11] <RoyK^> it uses a kernel made for kvm and some support stuff for that
[20:11] <guptaxpn> kvm?
[20:11] <guptaxpn> so would it work with virtualbox?
[20:11] <RoyK^> kernel virtual machine
[20:11] <guptaxpn> is it smaller?
[20:12] <RoyK^> not really, just install linux-image-virtual or whatever it's called
[20:12] <guptaxpn> huh
[20:12] <RoyK^> yeah, linux-image-virtual
[20:12] <RoyK^> use that on the guest
[20:12] <guptaxpn> wil lthat take up fewer resources?
[20:12] <RoyK^> not really, but it'll tune up things a little
[20:13] <guptaxpn> like what?
[20:13] <RoyK^> the resources are spent in userspace anyway
[20:13] <RoyK^> i/o drivers, scheduling choices...
[20:13] <RoyK^> it generally works better on VMs
[20:14] <guptaxpn> okay :)
[20:14] <guptaxpn> i'll switch the kernels once i finish installing the normal base system
[20:46] <lamont> yay EX_TEMPFAIL
[20:48] <wack479> I have been doing some research on mdadm and my raid5 array, and am i correct in my thinking that, in order to have the array put together at boot, i need to have it configurd in my mdadm.conf
[20:54] <geneticx_wrk> hi everyone. can someone please shed some light, I have created a ipsec vpn connection with an ASA 5510 using ubuntu..but once im connected I can't reach remote LAN or go on the internet..any help is appreaciated.
[21:04] <RoAkSoAx> kirkland, i've just submitted my student proposal to the GSoC website
[21:12]  * soren glances at soren`
[21:13] <soren> soren`: shoo!
[21:15] <tyska> hello guys
[21:15] <tyska> i really need help with UEC
[21:15] <tyska> some one can help me?
[21:16] <tyska> someone*
[21:17] <bogeyd6> kind of
[21:17] <soren> tyska: As always: Ask your questions. That's the only way anyone will be able to answer them.
[21:18] <tyska> i have installed CLC and CC on the same machine
[21:18] <tyska> as in the how-to page
[21:19] <tyska> and i have two NIC's, one is the NIC accessible from the  others machines in my lan
[21:19] <tyska> and one NIC just for the NODE's
[21:20] <tyska> but i cannot log in with elastic fox
[21:22]  * guntbert wonders what tyska is talking about - really no idea :-)
[21:23] <tyska> UEC
[21:24] <guntbert> and that is? (too lazy to google -- I admit it)
[21:27] <wack479> I have been doing some research on mdadm and my raid5 array, and am i correct in my thinking that, in order to have the array put together at boot, i need to have it configurd in my mdadm.conf
[21:31] <sYskk> the tinyurl is expired
[21:31] <sYskk> how do I add a subdomain to an existing domain ?
[21:31] <sYskk> I'm working from the ssh
[21:32] <guntbert> sYskk: please explain - where is the name server?
[21:33] <sYskk> actually the name server is located at my registrar - what I'm really looking for I guess is to add a virtual host in Apache
[21:34] <RoyK^> sYskk: /etc/apache2/sites-{available|enabled}
[21:34] <sherr> wack479: That's what I do.
[21:35] <wack479> sherr: hello again! and u add config to the mdadm.conf?
[21:35] <sYskk> RoyK^: thx
[21:38] <sherr> wack479: backup/rename mdadm.conf, then try :
[21:38] <sherr> mdadm --examine --scan > /etc/mdadm/mdadm.conf
[21:38] <sherr> Should give : ARRAY /dev/md* ... lines
[21:39] <sherr> wack479: as usual, be careful/backup ..
[21:39] <wack479> mdadm.conf is empty
[21:39] <wack479> so i dont need to backup it
[21:40] <wack479> and the os isnt on the array :)
[21:40] <sherr> Odd - but maybe your issue ...
[21:40] <wack479> but when i do "sudo mdadm --examine --scan > /etc/mdadm/mdadm.conf" it says permission denied
[21:41] <sherr> Just try the command i.e. sudo /sbin/mdadm --examine --scan
[21:42] <RoyK^> mdadm.conf shouldn't really be necessay - it's stored in the metadata on the drives
[21:42] <sherr> RoyK^: I agree. But in my experience, it often seems to be for me.
[21:42] <wack479> yes, but if i reboot, it has to be reassmbled without it, correct
[21:43] <sherr> wack479: did the cmd work/give output?
[21:43] <RoyK^> yes, the drives are tagged
[21:43] <wack479> yes it did
[21:43] <wack479> ARRAY /dev/md0 level=raid5 num-devices=5 UUID=87a687c5:9017aa15:3034c7f4:7083a99d
[21:43] <RoyK^> and mdadm should be self-concionous with the drives without a config file
[21:43] <sherr> Well, try it and see if it improves. It can't hurt.
[21:43] <wack479> its the same array, so should there really be 2 uuid's?
[21:43] <RoyK^> wack479: what's the problem?
[21:45] <sYskk> what is the command to restart apache ?
[21:45] <wack479> syskk: /etc/init.d/apache2 restart
[21:45] <RoyK^> apache2ctl restart?
[21:45] <RoyK^> apache2ctl graceful?
[21:45] <RoyK^> wack479: that's the kill-em-all command, there are better ones
[21:46] <wack479> RoyKwell, im trying to make sure my raid 5 is working properly, and when i restart, my array is gone, and i have to reassemble it.
[21:47] <RoyK^> ic
[21:47] <wack479> Royk: so as i was reading on it, it seems that to autoassemble u have to have a conf file
[21:47] <wack479> or whatever its called
[21:47]  * RoyK^ doesn't use linux for storage much, since opensolaris and zfs is so fucking cool
[21:48] <wack479> sherr: should there really be 2 uuids?
[21:52] <wack479> nm, ran it again, and only one showed up
[21:52] <RoyK^> not to bitch linux or anything, but for storage management, zfs beats it all
[21:52] <wack479> ok
[21:53] <RoyK^> snapshotting, comression and so on and copy-on-write and ssd caching
[21:53] <RoyK^> btrfs is a runner-up for zfs, but it's miles behind
[21:54] <soren> mdeslaur: I'm curious about virtinst 0.500.1-2ubuntu4.. Who says vmmouse is preferred?
[21:56] <amanda1> http://www.mdhjakten.se/dela/?id=dti2d6s
[21:57] <mdeslaur> soren: it offers better tracking, and is what works good, no?
[21:58] <mdeslaur> soren: vmmouse should get used automatically when you install lucid
[21:58] <RoyK^> wtf is vmmouse?
[21:58] <soren> mdeslaur: I didn't think we even installed or supported the vmmouse driver anymore.
[21:59] <soren> mdeslaur: Not since the whole evdev hotplug magic came along.
[21:59] <mdeslaur> soren: yes, it's in main, and gets installed by default
[21:59] <soren> mdeslaur: Perhaps it works better now in Lucid.
[21:59] <mdeslaur> soren: there's a udev rule for it now
[22:01] <mdeslaur> soren: I use vmmouse from dapper all the way to lucid
[22:01] <soren> mdeslaur: For a few releases at least you had to put stuff in your xorg.conf to enable it.
[22:01] <mdeslaur> soren: yeah
[22:01] <sherr> wack479: sorry, there should be one UUID per array
[22:01] <mdeslaur> soren: but lucid should get auto-configured with the udev rule
[22:02] <soren> RoyK^: It's a mouse bus that was first used by vmware and then added to QEMu.
[22:02] <RoyK^> ok
[22:03] <soren> VMWare has this special port that it uses for communication with the guests.
[22:03] <soren> That's how they send commands to the vmware tools you install in the guest.
[22:04] <soren> One of the things they use it for is for sending absolute pointer coordinates to the guest which makes for much nicer pointer integration for desktop environments.
[22:04] <soren> They also use it to send shutdown events and probaby a bunch of other tidbits.
[22:06] <soren> It's quite a hack. Especially these days where the hardware help with virtualisation. It violates the usual iopl restrictions in a number of ways. I can't imagine the vmware developers are happy with it these days. They have to special case a bunch of things to account for it.
[22:25] <wack479> sherr: no worries, and thats what i thought, added the one, and it seems to be working properly
[22:52] <RoyK^> hi, the page at https://help.ubuntu.com/community/UEC/StorageController says something about using the storage controller, but little about what sort of storage that should be used
[23:04] <jeffesquivel> hi, is any of the members of the HA Team here?