[01:39] <flaccid> smoser: any thoughts on this https://dpaste.de/Jk89B/ ?
[05:32] <centaur5> Is there anywhere that I could see an error log or hint as to why my Ubuntu install infinite reboots and doesn't even show a grub menu or error? I did run the boot repair disc and it claimed to be successful.
[05:36] <histo> centaur5: did you test the disk?
[05:37] <histo> centaur5: enable console mode and watch the messages fly by
[05:42] <centaur5> The drives test good. I did have to fsck all the partitions because it didn't successfully shut down.
[05:42] <centaur5> This claims it all repaired fine http://paste.ubuntu.com/5715473/
[05:44] <centaur5> I'm just wondering if somewhere grub keeps any kind of logs so I can troubleshoot.
[05:48] <centaur5> histo, can I enable console mode in a live disc then see errors?
[05:49] <histo> centaur5: what is looping the live cd or the hard drive?
[05:49] <centaur5> hard drive
[05:49] <centaur5> live CD boots fine
[05:51] <histo> centaur5: what version of ubuntu?
[05:52] <centaur5> 10.04
[05:52] <histo> centaur5: You said something about you had to fsck because the drives were unmounted or something what do you mean?  Is this a new install?
[05:53] <centaur5> No, capacitors on the mb died so I fixed them but then on boot it complained about needing to manually run fsck so I booted into live and did fsck on all the partitions.
[05:54] <centaur5> Ever since running fsck on the partitions it loops before booting grub.
[05:54] <centaur5> Before grub was at least booting the OS but failed to mount the partitions. Silly that an improper shut down was so unforgiving.
[05:55] <histo> centaur5: I would be more suspicous of your MB and the "fix" you did
[05:56] <centaur5> ha, I'd agree if it wasn't making it into a live CD just fine.
[05:57] <histo> centaur5: That doesn't mean you are using the same controller the hard disks are on
[05:57] <centaur5> Dedicated Adaptec raid card. 4 drives in a raid 5.
[06:00] <histo> centaur5: is the cdrom plugged into it?
[06:01] <centaur5> cd drive is onboard. I figured the capacitors north of the cpu wouldn't have anything to do with the southbridge though.
[06:03] <histo> centaur5: Also why are you installing 10.04 ?
[06:04] <centaur5> This server has been in production for 2.5 years with 10.04.
[06:04] <centaur5> I literally received my new hardware today for my new servers that will have 12.04 but I need this 1 running for at least a few weeks so I can learn how to do an HA cluster with the new ones.
[06:18] <histo> centaur5: No idea I would boot a live cd and check your menu entries and look for a force fsck file in / or whereveer
[06:20] <centaur5> histo, Alright, thanks for the response now I know there's really not much more troubleshooting I can do.
[06:28] <centaur5> histo, Just discovered my problem. I forgot to change my ROM boot setting for my raid card in the BIOS after it reset because of the capacitor change. However, it's still complaining about fsck on the partitions even though I ran it from the live disc.
[06:57] <czi_> Hey
[06:57] <czi_> Where can I find info on how to remove root password requirement for booting in single user mode?
[07:03] <histo> what?
[07:04] <histo> czi_: single user mode there should be no root password
[07:06] <Senor> what is the env varible name for library path?
[07:07] <Senor> in ubuntu
[07:26] <czi_> histo: Single user mode / recovery mode
[07:26] <czi_> It sure is password protected in the later ubuntu versions
[07:27] <histo> czi_: so what is it you uwant to do?
[07:27] <histo> czi_: It's protected with the user you installed with's password then. Root has no password
[07:28] <histo> czi_: if you are trying to change your pass boot a live cd chroot and passwd bob
[07:51] <czi_> histo: I want to remove the password protection
[07:52] <czi_> histo: I know how to boot up the server and change password, I do have the root password. But I want to remove the protection for future use ...
[08:19] <czi_> From the documentation: "If you decide to set a root password, you will be prompted for it in rescue mode."   But I dont wanna be prompted for it!!! :)
[08:25] <histo> czi_: then you'd have to remove it to not be prompted
[08:26] <histo> czi_: edit the recovery mode kernel line to rw and add init=/bin/bash
[08:27] <histo> czi_: actually you may not even need to change ro to rw just add init=/bin/bash
[08:30] <czi_> I don't want to remove my root password
[08:37] <histo> czi_: then add init=/bin/bash to end of kernel line
[11:09] <czi_> histo: Thanks
[11:29] <histo> czi_: np
[11:43] <zastaph> how would you log new incoming and/or outgoing connections? netstat and various tools only monitors in realtime the current state. I'd like a log of each unique new connection and their PID.
[11:44] <patdk-lap> connections have pid's?
[11:45] <patdk-lap> zastaph, hmm, you have heard about the firewall? iptables?
[11:54] <zastaph> sure, i've been looking for an application level firewall and tried most without any luck
[11:54] <zastaph> iptables has some LOG mode that I didn't quite investigate yet.. seems complicated :)
[11:55] <zastaph> watch -n 1 "netstat -tpanl | grep ESTABLISHED"
[11:55] <zastaph> that displays all new connections and their PID
[11:56] <zastaph> but in realtime.. i'd rather have a log
[11:59] <zastaph> to setup iptables, ufw, apparmor, first you need some intel about what needs your applications have
[12:27] <smoser> flaccid, i'm not sure about what went wrong for netdev_info and route_info (proably more information in /var/log/cloud-init.log . maybe)
[12:28] <smoser> but the subprocess that failed there was initctl. i suspect this is not ubuntu, in which case you'll need to set 'cc_ready_cmd'
[12:28] <smoser> in cloud-config
[12:45] <Daviey> zul / yolanda: Hey, can you both look spec out https://blueprints.launchpad.net/ubuntu/+spec/servercloud-s-server-app-banner-updates please?
[12:45] <yolanda> hi, ok
[12:45] <Daviey> zul: & https://blueprints.launchpad.net/ubuntu/+spec/servercloud-s-openstack-pkg
[12:46] <zul> eff
[12:46] <yolanda> i have to check with zul, i have few idea of it
[12:46] <Daviey> & servercloud-s-openstack-hypervisor
[12:46] <Daviey> zul: catch up :)
[12:46]  * zul will propose less blueprints next time
[12:49] <Daviey> zul: heh, take a look at all of them :)
[12:49] <Daviey> zul: if you want to share the burden a bit more, that is possible.
[12:50] <zul> Daviey:  can we fold the openstack-pkg blueprint into the openstack blueprint since it was more of a are we going to switch to git
[12:51] <Daviey> zul: yeah, that might make sense
[12:52] <Daviey> zul: mind you, servercloud-s-openstack-havana is chocked
[12:52] <zul> Daviey:  ill add the work items to the openstack
[12:52] <zul> Daviey:  its always is
[12:52] <Daviey> heh
[12:53] <yolanda> Daviey, should i add the openstack packages to the dep-8-tests blueprint?
[12:53] <Daviey> yolanda: i think so, what do you think zul ?
[12:53] <zul> Daviey: sure
[12:53] <RoyK> Trying to sort out bug 1171945 - can someone help me understand where MD raids are assembled?
[12:54] <zul> although checking to see if the help message runs is not much of a test but hey what are you going to do
[12:56] <Daviey> zul: it checks more than it appears
[12:57] <Daviey> xnox: Hey, are you able to look at the issue RoyK raised ^ ?
[13:02] <xnox> Daviey: yes. I have.
[13:37] <zul> Daviey:  what did you want me to do with hypervisors?
[13:38] <Daviey> zul: take a look at the Wi's and whiteboard to make sure the content is accurate and complete
[13:39] <zul> is is
[13:42] <pythonirc1011> We need to run an email server. How hard is it to configure one in ubuntu?
[13:42] <pythonirc1011> its very light emailing we need. Perhaps there is a better solution than running our own mailserver?
[13:43] <jcastro> https://help.ubuntu.com/12.04/serverguide/email-services.html
[13:43] <pythonirc1011> jcastro: https://help.ubuntu.com/community/Postfix  - was reading this
[13:43] <pythonirc1011> but it seems even setting up postfix is non-trivial correctly
[13:44] <pythonirc1011> I tried sending emails using gmail, but it starts complaining that we should use a web browser
[13:44] <jcastro> you probably want to search for something like "postfix smtp relay through gmail"
[13:46] <pythonirc1011> jcastro: well gmail screwed us up once. When we use smtp from gmail, after 5 emails, it starts complaining : SMTPAuthenticationError: (534, '5.7.9 Please log in with your web browser and then try again. Learn more at\n5.7.9 https://support.google.com/mail/bin/answer.py?answer=78754\n5.7.9 {WEBLOGINREQUIRED} ...
[13:48] <pythonirc1011> jcastro: Perhaps this used to work in 2008: http://ubuntu-tutorials.com/2008/11/11/relaying-postfix-smtp-via-smtpgmailcom/ - am not sure Gmail will put a captcha at some point on this solution?
[13:59] <patdk-wk> pythonirc1011, how can you put a *captcha* in that solution?
[14:10] <koolhead17> Daviey: lol
[14:12] <Daviey> koolhead17: ?
[14:20] <pythonirc1011> jcastro: is postfix the way to go for emails? or is there a better alternative by now?
[14:21] <Daviey> pythonirc1011: still the best.
[14:21] <jcastro> it's the default for a reason. :)
[14:21] <pythonirc1011> jcastro: If I setup postfix relay + gmail = captcha trouble at some point, isn't it?
[14:22] <jcastro> I think at this point you should probably just do postfix standalone
[14:22] <jcastro> or use an SMTP relay service
[14:22] <Daviey> Debian is also in discussions if they should switch from exim to postfix.
[14:22] <jcastro> http://aws.amazon.com/ses/faqs/ for example
[14:22] <maxb> Postfix is a sensible option for simpler setups. Exim is a great MTA if you need to do weird and wonderful things with your mail processing pipeline.
[14:22] <Daviey> pythonirc1011: i don't think they put captcha on imap, pop3 or smtp
[14:22] <jcastro> depends on how comfortable you are with postfix
[14:23] <pythonirc1011> http://www.authsmtp.com/auth-smtp/pricing.php - I wish there base was free
[14:23] <jcastro> you likely won't find any free service offering smtp relay, heh
[14:23] <Daviey> maxb: I'm not trolling, but have an example where exim excels where postfix lacks?
[14:23] <pythonirc1011> Daviey: http://ubuntu-tutorials.com/2008/11/11/relaying-postfix-smtp-via-smtpgmailcom/ - is this going to be reliable?
[14:24] <Daviey> pythonirc1011: worth a try :)
[14:24] <pythonirc1011> I don't care about speed/accuracy...only care about ease of use + reliability + spam filtering.
[14:24] <Daviey> pythonirc1011: i'd try it :)
[14:24] <pythonirc1011> Daviey: I tried using gmail email send/receive using python - blasted on my face. When we use smtp from gmail, after 5 emails, it starts complaining : SMTPAuthenticationError: (534, '5.7.9 Please log in with your web browser and then try again. Learn more at\n5.7.9 https://support.google.com/mail/bin/answer.py?answer=78754\n5.7.9 {WEBLOGINREQUIRED} ...
[14:25] <Daviey> interesting
[14:25] <Daviey> didn't know it did that
[14:25] <maxb> Daviey: Configuring a complex set of Exim routers, mainly. Postfix seems to be all about setting up particular options. Exim's approach of plugging bits of logic together in a given order feels more able to deal with complex stuff
[14:26] <lamont> Daviey: Tower of Hanoi is doable in exim, not so much in postfix.  your choice as to which this argues for
[14:27] <pythonirc1011> wish I could do : sudo -i; apt-get install emailserver; emailserver.configure --domain = "mydomain.com"  :)
[14:27] <pythonirc1011> which is easiest to use : sendmail/postfix/exim/...?
[14:27] <Daviey> maxb / lamont: So it's possibly a familiarity issue? :)
[14:27] <jcastro> don't even put sendmail on the table
[14:27] <lamont> pythonirc1011: I have a definite bias towards postfix, but then I've maintained it for 16 years
[14:27] <Daviey> wow, is it that long?
[14:27] <pythonirc1011> lamont: Maybe you can hand hold me for my install then :)
[14:28] <lamont> Daviey: since the alpha days
[14:28] <lamont> actually, I think we have a couple of months before we hit 16 years proper
[14:28] <maxb> Daviey: *Partially* maybe. But even as an experienced Exim admin, intent on seriously investigating Postfix, I couldn't figure out how to get it to approach that sort of thing, whilst staring at the docs
[14:29] <lamont> 0.0.1998121-0 was packaged Fri, 11 Dec 1998 22:31:37, but I kinda accidentally replaced sendmail with it in production about july of 1997, iirc
[14:29] <pythonirc1011> lamont: so for installing postfix for one domain - + want spam filtering + want to auth anyone who wants to send email thru the server - you've a list of commands for me?
[14:29] <Daviey> maxb: I admit to being heavy on google juice whilst configuring a complex postfix.
[14:30] <lamont> Daviey: exim is certainly able to do more things than postfix.  Postfix picked a particular target and went for "do that well".  If you get too far in to the rough, exim is going to be better every time.  OTOH, these days, you have to go a ways to get into the rough that far
[14:31] <maxb> My main objection to it is that I've been unable to find a decent explanation of a mental model to understand its inner workings. Whereas, with Exim, just reading the routers section of the config file gets you pretty close
[14:31] <lamont> pythonirc1011: I'd have to send you to one of the howtos
[14:32] <maxb> I have entertaining requirements like "If the mail contains a text/calendar MIME part, deliver an extra copy somewhere else"
[14:32] <lamont> maxb: lots of little daemons that do one thing and only one thing, strung together to make a mail server.  That stringing makes it more complicated in some ways, and trivially nice in others.  It also means that reviewing it for security is much much simpler
[14:32] <lamont> maxb: reading exim config files makes my brain hurt. OTOH, that's probably a cyclical thing since I avoid making my brain hurt when possible
[14:32] <pythonirc1011> lamont: web link please? + ubuntu
[14:33] <lamont> Daviey: I think you know all the good ubuntu howto links for postfix better than I do...
[14:33] <Daviey> you'd think so...
[14:33] <maxb> Exim isn't simple, I agree. It shines when you *want* an MTA that isn't simple :-)
[14:33]  * lamont feels more like an upstream for postfix than a core-dev
[14:34] <lamont> maxb: fact
[14:34] <Daviey> ScottK and ivoks, spent a bunch of time on mail-stack-delivery package
[14:34] <pythonirc1011> can I ask postfix questions here? :)
[14:34] <Daviey> That should try and simplify the common use cases
[14:34] <ScottK> Daviey: Great.  It's needed some help for awhile.
[14:34] <ScottK> pythonirc1011: Yes.
[14:34] <lamont> pythonirc1011: sure.  every now and then, I even let myself get sucked into the conversations...
[14:35]  * lamont goes back to what he's supposed to be doing, since ScottK is here
[14:35] <ScottK> No!!!!  ScottK goes back to what he's supposed to be doing first.
[14:35] <pythonirc1011> I've a domain name "example.com". And want postfix to handle all emails to *@example.com.  What should "System mail name" be?
[14:36] <Daviey> pythonirc1011: does http://www.placedusoleil.net/node/6 help?
[14:36] <ScottK> Those are unrelated.
[14:36] <ScottK> System mail name is the name of the server.
[14:37] <pythonirc1011> Daviey: I don't have an active directory setup. How hard is that to setup?
[14:37] <pythonirc1011> ah! Windows  - nope
[14:37] <ScottK> The mydestination parameter in /etc/postfix/main.cf is what controls what domains are considered as local domains for delivery.
[14:38] <pythonirc1011> Is there a way to hardcode in a text file : "x@example.com" - forward it to "why@gmail.com", ... , "*@example.com" - forward it to "z@zz.zzz"?
[14:38] <pythonirc1011> ScottK: master.cf?
[14:39] <ScottK> No, main.cf
[14:39] <ScottK> Yes.
[14:39] <pythonirc1011> how?
[14:39] <pythonirc1011> Perhaps this can be modified for my purposes: http://www.placedusoleil.net/node/6 ?
[14:39] <ScottK> I would suggest you join the postfix users mailing list and ask there as they have more people with postfix expertise than we do here.
[14:40] <pythonirc1011> I just need postfix to mail forward the emails. And I want  a way to be able to use it to send emails by authenticated users.
[14:40] <ScottK> (I should really be doing work and not messing around on IRC)
[14:41] <ScottK> pythonirc1011: I'd look at http://www.postfix.org/documentation.html
[14:41] <ScottK> Most of what you're asking about is pretty standard and is covered in one of the documents linked from there.
[15:31] <hallyn_> ahs3: d'oh, https://buildd.debian.org/status/package.php?p=netcf
[15:56] <hallyn_> ahs3: if you're around, we need http://people.canonical.com/~serge/netcf-arch.debdiff to fix the non-lniux arches tryign to build
[15:56] <hallyn_> ahs3: for the armel and sparc failures, I *think* those were transient
[15:57] <hallyn_> bc the ifupdown version currently looks to be right on those arches
[15:58]  * ahs3 otp for a bit...bbiab
[16:00] <hallyn_> ahs3: np thanks!
[16:37] <atrius> hello all.. is there a PPA for _recent_ IRC servers aside from ngircd?
[17:19] <ahs3> hallyn_:  hrm.  i'll add in the patch and re-submit.  hopefully you're right that the other failures are transient
[17:23] <SpamapS> hallyn_: hey I think I'm hitting bug 1031063 with a nested openstack  nova.. you around to help debug perhaps?
[17:23] <SpamapS>    domain = self._conn.defineXML(xml)\n', '  File "/usr/lib/python2.7/dist-packages/eventlet/tpool.py", line 187, in doit\n    result = proxy_call(self._autowrap, f, *args, **kwargs)\n', '  File "/usr/lib/python2.7/dist-packages/eventlet/tpool.py", line 147, in proxy_call\n    rv = execute(f,*args,**kwargs)\n', '  File "/usr/lib/python2.7/dist-packages/eventlet/tpool.py", line 76, in tworker\n    rv = meth(*args,**kwargs)\n', '  File "/usr/lib/
[17:25] <ruben231> hi guys anyone used here tungsten as mysql replicator.?
[17:26] <hallyn_> SpamapS: lunching, will look shortly
[17:26] <hallyn_> ahs3: thanks!
[17:26] <SpamapS> ruben231: I have not, but #mysql has a bunch of people who know about these things.
[17:28] <RoyK> anyone here that knows where ubuntu 12.04 and later assembles the raid? it's utterly broken for nested raids
[17:31] <SpamapS> RoyK: initrd
[17:31] <RoyK> SpamapS: well, where in there?
[17:31] <SpamapS> RoyK: look at the scripts mdadm puts in /usr/share/initramfs-tools
[17:32] <SpamapS> RoyK: It should just be pulling the stuff from the disk metadata.
[17:32] <RoyK> I have, but I can't find anything actually starting the raid
[17:32] <RoyK> some incremental tuff, but it seems to be broken
[17:33] <RoyK> normal raids starts well, but nested raids like raid-5+0 won't
[17:33] <RoyK> SpamapS: bug 1171945
[17:34] <RoyK> so, tested on lucid, precise, quantal and wheezy - works on lucid and wheezy but not the others
[17:35] <SpamapS> RoyK: mdadm is *really* broken on ubuntu. I tried for a little while to attack it, but its just a mess.
[17:36] <SpamapS> RoyK: it doesn't eat your data anymore, but it will often fail to start unfortunately
[17:36] <SpamapS> RoyK: I think the problem is that Canonical customers aren't mdadm users. :p
[17:39] <RoyK> SpamapS: well, do you know anything about *what* is broken in the mdadm startup?
[17:40] <RoyK> AFAIK it's only startup stuff that's broken - other things work well
[17:40] <RoyK> there were some issues at 12.04.nothing with udev getting a bit stressed, not picking up all the md members, but that was fixed at 12.04.1
[17:41] <SpamapS> RoyK: for nested, I think it is just assembling wrong. For some of the other issues, it is handling missing disks and stuff that should be ignored as critical errors.
[17:41] <SpamapS> RoyK: it used to handle those things really wrongly and occasionally wipe out your RAID .. so it's better than it was.
[17:42] <SpamapS> RoyK: basically what I did was make it more careful. The follow up to that would then be to find the places it is being too careful and handle them properly.
[17:42]  * RoyK thinks the initial 14.04 should be named 14.04.0 to state that "this is early"
[17:42] <SpamapS> RoyK: but there is so much rage in the bug tracker, it is hard to weed out the actual data from the vitriol.
[17:43] <RoyK> well, I don't think there's much in this bug
[17:45] <SpamapS> RoyK: anyway, xnox has done some work to make it better.. he may have more insight than I do.
[17:45] <RoyK> SpamapS: still, do you know what actual scripts that does the assembly for the raid? this is from 12.04 and forward, as explained
[17:46] <SpamapS> RoyK: dpkg -L mdadm , look at the scripts in /usr/share/initramfs-tools
[17:46] <RoyK> SpamapS: I've talked to him, but he couldn't give me a clear answer - I tried the udev part from wheezy, where it works, but no luck - see the bug
[17:47] <RoyK> s/bug/bug report/
[17:47] <SpamapS> RoyK: sorry, I have no interest in mdadm anymore.. no time to see the bug. ;)
[17:48] <RoyK> please?
[17:50] <SpamapS> RoyK: ok, I read it. Still nothing to add. :-P
[17:50] <sarnold> RoyK: heh, there's probably nothing really new in the bug anyhow -- he knows it won't do nested..
[17:51] <RoyK> shouldn't such a bug be prioritised somewhat higher? I mean - if someone want to build large storage...
[17:51] <SpamapS> I do recall that the udev rules we have in Ubuntu are wrong and need fixing.
[17:51] <SpamapS> RoyK: they buy a large storage controller? ;)
[17:51] <RoyK> SpamapS: same udev rules on debian - works there
[17:51] <RoyK> SpamapS: wrong answer
[17:52] <SpamapS> RoyK: my point isn't that nobody wants mdadm. Its that nobody is paying Canonical to prioritize it.
[17:53] <SpamapS> Paying customers have SAN, built in RAID cards, or gluster/ceph/etc.
[17:53] <RoyK> which seems rather odd, since canonical seems to be trying to get a larger market share
[17:53] <sarnold> RoyK: if wheezy's packages work alright on wheezy, why not try the wheezy packages on ubuntu?
[17:53] <SpamapS> RoyK: Canonical is trying to capture _untapped_ markets like OpenStack deployments and large hadoop clusters. Not home server / SMB server markets.
[17:53] <RoyK> well, afaik wheezy assembles the raid differently
[17:54] <SpamapS> RoyK: they won't stand in your way, but you're going to have to put some resources in to get that stuff prioritized.
[17:54] <yofun> my server wont unrar my rar file correctly
[17:54] <yofun> it says "failed" all it gives me
[17:54] <RoyK> yofun: tried 7z?
[17:55] <yofun> thats a ubuntu package?
[17:55] <sarnold> yofun: perhaps the file is corrupted, check the sha1sum of the file against a version on another machine that unpacks correctly.
[17:55] <RoyK> yofun: run 7z - it's a command
[17:55] <yofun> no other files didnt work either
[17:55] <RoyK> yofun: p7zip-full is the package
[17:55] <yofun> -bash: 7z: command not found
[17:56] <hallyn_> stgraber: ok, so is your host i386, or are you doing i386 host on amd64, and then nesting inside the i386 guest?
[17:56] <hallyn_> bc i don't think you can nest kvm inside i386-on-amd64 (but i'm not 100% sure on that)
[17:56] <yofun> atm im useing unrar-free
[17:56] <RoyK> try 7z
[17:56] <RoyK> better supported
[17:57] <RoyK> and takes all sorts of formats
[17:58] <hallyn_> lemme try on saucy first
[17:58] <yofun> Extracting  vmaps/Miningsifter01.m2.vmo     Unsupported Method
[17:58] <SpamapS> hallyn_: you mean SpampaS ? :)
[17:58] <hallyn_> oh, no, i see, i386 host
[17:58] <hallyn_> SpamapS: d'oh, yeah.
[17:58] <SpamapS> hallyn_: i386 host and guest
[17:58] <SpamapS> I think
[17:58] <hallyn_> SpamapS: yeah that's what it shows.  sorry, misread at first
[17:59] <RoyK> yofun: quite possibly a corrupt rar file
[17:59] <SpamapS> hallyn_: virsh capabilities does show x86_64 .. which is weird
[18:00] <hallyn_> SpamapS: in the bug virsh capabilities output showed i686
[18:00] <hallyn_> different machine?
[18:00] <SpamapS> hallyn_: Linux bootstack-vm 3.5.0-31-generic #52-Ubuntu SMP Thu May 16 16:30:01 UTC 2013 i686 i686 i686 GNU/Linux
[18:01]  * stgraber ignores the highlight
[18:01] <hallyn_> stgraber: thanks
[18:06] <SpamapS> hallyn_: so is this just "i386 can't nest kvm" ?
[18:06] <hallyn_> SpamapS: i386 should be able to nest.  oh, though /etc/default/qemu-kvm in quantal may still have KVM_NESTED variable defaulting to 0
[18:06] <SpamapS> hallyn_: or should I set my libvirt_type=qemu ?
[18:07] <hallyn_> no, defaults to on
[18:07] <SpamapS> KVM_NESTED=" nested=1"
[18:07] <hallyn_> SpamapS: this *could* just be the old bug about /dev/kvm not getting chowned right.  can you check ls -l /dev/kvm?
[18:08] <hallyn_> SpamapS: or, see whether a reboot of the first level guest fixes it?
[18:09] <SpamapS> # ls -l /dev/kvm
[18:09] <SpamapS> ls: cannot access /dev/kvm: No such file or directory
[18:09] <SpamapS> *that* might explain it :)
[18:10] <hallyn_> SpamapS: d'oh, is it using virtual kernel?
[18:10] <hallyn_> i thought it started in raring, but maybe quantal also doesn't ship the kvm driver
[18:11] <SpamapS> hallyn_: generic kernel
[18:11] <hallyn_> hm
[18:13] <RoyK> SpamapS: isn't the virtual 32bit kernel with virtualisation now?
[18:13] <RoyK> erm
[18:13] <RoyK> SpamapS: isn't the generic! 32bit kernel with virtualisation now?
[18:14] <SpamapS> hallyn_: what creates /dev/kvm ?
[18:15] <RoyK> the kvm mod?
[18:15] <RoyK> modprobe kvm
[18:20] <hallyn_> really, kvm_intel
[18:20] <hallyn_> but that should be automaticlaly modprobed at boot
[18:21] <SpamapS> FATAL: Error inserting kvm_intel (/lib/modules/3.5.0-31-generic/kernel/arch/x86/kvm/kvm-intel.ko): Operation not supported
[18:21] <SpamapS> [70824.341783] kvm: no hardware support
[18:21] <genii-around> That doesn't sound good.
[18:22] <SpamapS> do I need to turn on "vme" and/or "vmx" ?
[18:22] <SpamapS> looks like my host VM is just not quite right
[18:25] <hallyn_> SpamapS: what does uname -a show again?
[18:26] <hallyn_> what does 'kvm-ok' show in the guest?
[18:26] <hallyn_> and /sys/module/kvm_intel/parameters/nested on the host?
[18:27] <SpamapS> hallyn_: kvm-ok says no-go
[18:27] <SpamapS> hallyn_: the vm was not configured with a CPU profile, so vmx is off
[18:28] <SpamapS> hallyn_: on host nested == Y
[18:29] <SpamapS> hallyn_: I think I just need to reboot w/ vmx
[18:29] <SpamapS> or re-configure w/ qemu only
[18:30] <hallyn_> SpamapS: hm, right, the qemu64 cpu type automatically enables vmx.  the others probably don''t
[18:32] <RoyK> SpamapS: using ancient hardware?
[18:33] <soy_el_pulpo> ancient rocks!!! lol
[18:33] <soy_el_pulpo> ;)
[18:33] <RoyK> get your old 486 to roll
[18:33] <RoyK> down the hill, perhaps
[18:33] <soy_el_pulpo> 386 sx with math-co!!
[18:33]  * RoyK is reading http://bofh.ntk.net/BOFH/
[18:34] <SpamapS> RoyK: no, nesting kvm
[18:34] <RoyK> soy_el_pulpo: heh - the math co was a 386DX, not only the FPU
[18:34] <SpamapS> RoyK: model name	: Intel(R) Core(TM) i7-3667U CPU @ 2.00GHz
[18:34]  * RoyK sticks to 64bit these days
[18:34] <RoyK> eats a bit more memory, but works better
[18:35] <soy_el_pulpo> jaja, the SX came with a socket where you can add it... if I'm not wrong...
[18:35] <SpamapS> RoyK: when trying to run a mini openstack across 6-10 VMs on your 16GB laptop.. saving 5% matters. :)
[18:35] <soy_el_pulpo> the sx came without it... but my board had an additional socket...
[18:36]  * RoyK has a bunch of VMs on his home server and it only has 8GB (until I plug in the new 8GB)
[18:36] <RoyK> perhaps time to do that
[18:36] <soy_el_pulpo> also I got simm slot raisers, I do not how they are called, but it allowed me to ad more than 8 chips...
[18:36] <soy_el_pulpo> it was crazy...
[18:38] <hallyn_> SpamapS: can you add "+vmx" to the cpu flags?
[18:38] <SpamapS> RoyK: right, but I need at the very least 1GB per VM just to start the python monstrosity
[18:38] <hallyn_> on the host.  (not sure how you're specifying them right now)
[18:38] <SpamapS> hallyn_: I've done that, but I am trying a few other things before I reboot
[18:38] <hallyn_> k
[18:38] <hallyn_> SpamapS: how did you do it?
[18:39] <SpamapS> hallyn_: in virt-manager :)
[18:40] <hallyn_> cool
[18:45] <SpamapS> aaaaand I also need to up the RAM on this VM
[18:45] <SpamapS> 2013-05-30 18:45:24.109 1373 TRACE quantum.agent.dhcp_agent OSError: [Errno 12] Cannot allocate memory
[18:45] <SpamapS> doh
[18:51] <zul> SpamapS:  at least its not java
[18:51] <SpamapS> zul: nobodyw ould think to try and build a cloud on java...
[18:52] <zul> hah
[18:57] <SpamapS> hallyn_: ok, after rebooting w/ vmx we have the hvm capability.. but it seems to.. pwn the guest's CPU
[18:57] <SpamapS> like, to a level where nothing works anymore
[18:58] <SpamapS> kernel isn't even responding to arp at this point
[18:59] <hallyn_> SpamapS: hm, wonder if it would help to allocate hugepages to the first level guest
[18:59] <hallyn_> are you getting any dmesg info at the host?
[18:59] <SpamapS> [165124.945856] kvm [1917]: vcpu0 unhandled rdmsr: 0x345
[18:59] <SpamapS> [165124.948111] kvm_set_msr_common: 22 callbacks suppressed
[18:59] <SpamapS> [165124.948113] kvm [1917]: vcpu0 unhandled wrmsr: 0x680 data 0
[19:01] <GrueMaster> Does anyone know why an ubuntu-server (12.04.2) would release/renew it's ip address on reboot?  System is configured for dhcp and submits the hostname to the dhcp server each time.
[19:03] <SpamapS> hallyn_: so I did the "copy host configuration" button in virt-manager. Trying 'kvm64' now.
[19:03] <SpamapS> though I suppose kvm32 would be more appropriate
[19:04] <Chocobo> Hi all.  Is it possible configure a network interface to get an IP from DHCP without also setting nameservers and routes?
[19:04] <jcastro_> GrueMaster: dhcp server perhaps giving it out as a reservation instead of on-demand?
[19:05] <Chocobo> sorry, when I say routes I should say "default route"
[19:05] <GrueMaster> Sadly, we have no control of the dhcp server (Windows based and controlled by an uber paranoid IT group).
[19:07] <SpamapS> hallyn_: ok, that seems to have fixed it
[19:07] <hallyn_> SpamapS: sigh, i don't get it.  you copied your nested guest configuration to a new one and that fixed it, or befor you were trying amd64 guest and i386 guest works?
[19:09] <SpamapS> hallyn_: before I was trying "copy host configuration" in virt-manager, which used "SandyBridge" as my guest CPU config.
[19:10] <SpamapS> hallyn_: I changed that to 'kvm64' and that seems to ahve worked. I now have a good working nested kvm guest.
[19:11] <hallyn_> hm, so virt-manager wasn't reading cpu flags right or something?  interesting.
[19:11] <hallyn_> thanks
[20:04] <roborino> just got a dell r520 and installed 13.04 server with openssh, however every time I connect to the box shortly after it just disconnects with Write failed: Broken pipe
[20:05] <sarnold> roborino: how 'shortly'? seconds? near-instantly? minutes? hours?
[20:05] <roborino> sarnold: within a minute
[20:06] <roborino> sarnold:  also get ssh: connect to host calvin port 22: Connection refused, but when I look on the console sshd is running
[20:09] <roborino> sarnold: right now I hit return a couple times, no problem.  Third time I get Write failed:  Broken pipe
[20:10] <sarnold> roborino: broken pipe makes me think stateful firewall is missing rules to add associated packets...
[20:10] <sarnold> gotta run now, sorry, but look for ESTABLISHED and firewall rules...
[20:24] <roborino> enabled ufw allow 22
[20:25] <roborino> still get Write failed: Broken pipe after about a minute
[20:42] <thesheff17> no matter what I do I still get libsvn1 (= 1.6.17dfsg-3ubuntu3) but 1.7.9-0ubuntu0~ppa1 is to be installed
[20:42] <thesheff17> even though the PPA is gone
[20:42] <thesheff17> any idea what I can do?
[21:01] <Neozonz> can someone help me conver this .htaccess file into lighttp's url rewrite http://pastebin.com/wh93vjE8
[21:01] <Neozonz> *convert
[21:12] <hallyn_> smoser: mjt pointed me to http://mentors.debian.net/package/sgabios   i think you'd like it :)