[00:00] <SpamapS> patapouf: sure
[00:00] <SpamapS> patapouf: http://uec-images.ubuntu.com/releases/10.04/release/  that may help :)
[00:00] <SpamapS> patapouf: https://help.ubuntu.com/community/EC2StartersGuide  and that
[00:02] <nxt_user> hello
[00:03] <patapouf> SpamapS, Thanks for the link...
[00:04] <patapouf> SpamapS, I then select the Ubuntu 8.10 LTS 32 bits AMI .. I'm done creating the instance (I think)
[00:04] <patapouf> What is the step to access it ?
[00:18] <Undeon> Hey guys... Anyone here knows an howto to make an Active Directory user login automatically into squid?
[00:19] <Undeon> I've already read a few howtos here, but i'm struck...=\
[01:05] <patapouf> Hi, Anybody is well aware of the pricing chart of Amazone EC2 service ??
[01:06] <patapouf> I need a server running 24/7 which will not be heavyly used (very low traffic). Any suggestion ? What do I need to buy ?
[01:06] <wido> a vps somewhere?
[01:07] <patapouf> wido, are you asking me ?
[01:07] <wido> jep
[01:07] <wido> a virtual private server
[01:08] <patapouf> I don't have a virtual private server
[01:08] <wido> no, but you could rent one somewhere
[01:08] <wido> a few dollars a month
[01:09] <patapouf> wido, I don't understand the need if it's provided by Amazon ...
[01:09] <wido> or if if you wanted to be on decent hardware on a decent network some more
[01:10] <wido> you just said that you need a server with low traffic running 24/7
[01:10] <wido> so i suggest a vps
[01:10] <wido> but take amazon if you please
[01:10] <patapouf> I understand, but I don't know any
[01:10] <RoyK> patapouf: a vps is what is provided by amazon et al
[01:11] <patapouf> And amazon seas to satisfy me need so far. If you have any other suggestion/name. I will sure consider it before making my choice
[01:11]  * RoyK sticks to his dedicated server
[01:12] <wido> patapouf: what do you want to run on it? who are your users? where are your users? what is your problem with one minute downtime? and with one hour? and with one day?
[01:12] <wido>  do you need support? in english?
[01:12] <wido> etc, etc
[01:14] <patapouf> wido, I want to host various application and service (apache, tomcat, ssh, rsync, etc.), I really need root access. There will be various user. Very low traffic. I have issues if the service is down for a day or more.
[01:14] <patapouf> I don't need any support
[01:15] <wido> and your user base is located in?
[01:16] <patapouf> In fact, I need this kind of service because I want to get ride of the server I have at my office. The Internet is not reliable nor is the electricity.
[01:16] <patapouf> The user are base in North-America
[01:16] <wido> so, pick a hoster in a real datacentre
[01:17] <patapouf> do you have names ?
[01:17] <wido> with good connections near your users
[01:17] <wido> with a sla that pleases you
[01:17] <wido> and one that responds on e-mail
[01:18] <wido> no, i'm not from the us/canada
[01:18] <wido> so i can't give any advice
[01:18] <wido> but make sure you understand your needs
[01:19] <wido> sla, uptime, support, location, network, redundancy, etc.
[01:25] <qman__> patapouf, amazon ec2 doesn't seem the right fit for your needs, just a rough calculation on the pricing, you're looking at over $60 a month
[01:25] <qman__> you can get your own VPS at linode for $20 a month
[01:26] <patapouf> looking at the princing chart.. it's only 300/year$ for Small instance and 57$/year for micro instance
[01:26] <qman__> ec2 is CPU-usage based pricing, which is a lot for a 24/7 service
[01:26] <qman__> the pricing is better suited to things that aren't 24/7
[01:27] <qman__> servers on demand, that sort of thing
[01:28] <patapouf> I see .. so the princing is 300$/year + 0.03$/hours .... 300$ + 0.03$*24*365 = 562$
[01:28] <qman__> no, it's $300 a year OR .03/hour
[01:29] <qman__> but I don't think you can run what you want to run on a micro instance
[01:29] <qman__> you'd have to verify though
[01:29] <qman__> err, nevermind, you're right
[01:29] <qman__> reading the wrong numbers
[01:30] <qman__> anyway, it'd work but it'd be quite expensive for what you're trying to do
[01:30] <qman__> a regular VPS is more suited
[01:30] <wido> indeed
[01:30] <qman__> linode is one of the slightly more expensive ones, regarded as high quality from what I've heard
[01:30] <patapouf> qman__ do you know any in North-America
[01:31] <qman__> http://www.linode.com/
[01:31] <qman__> probably overkill for you, but still cheaper than ec2
[01:36] <wido> i don't see any info on their network though
[01:36] <giovani> definitely plenty of cheap VPS vendors out there
[01:37] <giovani> wido: what kind of network do you want/need?
[01:37] <wido> i didn't ask :
[01:37] <wido> :)
[01:37] <wido> it was patapouf
[01:37] <patapouf> I'm still browsing the website.. it's look interesting.
[01:38] <giovani> slicehost is another well-known one
[01:38] <giovani> both are pretty pricey
[01:38] <qman__> if you're willing to gamble on availability and customer service, you can find VPSes under $10 a month
[01:38] <giovani> if you searc the WebHostingTalk forums VPS section
[01:38] <giovani> you'll see tons of deals
[01:38] <giovani> search*
[01:38] <giovani> I have about 10 $5 and under/mo VPSes
[01:38] <giovani> most are great
[01:39] <patapouf> Well looking at the princing, I need something with more than 256Mb ram and 10GB of storage
[01:40] <giovani> patapouf: well, what do you need? there are tons of companies out there
[01:40] <giovani> I use PhotonVPS for my "big" box -- 20GB of storage and 512MB of RAM I think for $17
[01:40] <shauno> I'd be wary of overestimating how much ram you actually need.  256-512 is plenty of a lot of things
[01:41] <giovani> err, $11/mo rather
[01:41] <qman__> I run quite a bit on real hardware with 256MB, low traffic
[01:41] <qman__> got apache, ssh, an IRC logger, and a few others
[01:41] <patapouf> Hard to say, Something with 512Mb Ram, 20GB of storage is ok but more is better; 10GB of BW should be enough
[01:42] <giovani> patapouf: what are you actually needing to run?
[01:42] <giovani> because I run web/mail/dns servers with 64MB of ram
[01:42] <giovani> easily
[01:42] <qman__> yeah
[01:43] <qman__> to need 512MB, you've got to have a very heavy program load
[01:43] <patapouf> WEB service of many kind : apache + mod_mono + mod_php, tomcat
[01:43] <giovani> ok, tons of bloat
[01:43] <patapouf> kind of
[01:43]  * giovani sticks with nginx and lighty
[01:43] <wido> ah, tomcat :)
[01:44] <qman__> but, even so
[01:44] <patapouf> I currently have a VM with : 512Mb RAM, 150GB of Storage, 2.8 Ghz Core2Duo
[01:44] <qman__> you can still get a VPS with that fairly cheap
[01:44] <patapouf> It's running just fine, but as I mention earlier, the internet and the electricity is not very stable here
[01:45] <giovani> sounds like you want a cheap dedicated server then
[01:45] <giovani> that will be cheaper than a VPS of those specs
[01:45] <giovani> but theoretically less reliable as well
[01:45] <qman__> I don't know about cheaper
[01:45] <qman__> I couldn't find any from reputable providers for less than $100 a month last I looked
[01:46] <giovani> your definition of reputable is probably different than mine
[01:46] <giovani> but, you'd be hard-pressed to find a VPS from a reputable vendor with 150GB of HD and a dedicated core of a 2.8GHz C2D for less than $100/mo as well
[01:46] <qman__> that's true, but I don't see any reason to need that much disk space on a web server
[01:47] <giovani> yeah, I also don't see any need to run mod_mono and tomcat :)
[01:47] <qman__> indeed
[01:47] <patapouf> really ?? is that really hard .. cause Amazon whats providing this for ~562$/year
[01:47] <qman__> not with those specs
[01:47] <patapouf> 46$/month
[01:47] <qman__> that's easily a "large" instance, if not bigger
[01:48] <giovani> look
[01:48] <giovani> PhotonVPS will do 50GB disk, and 2GB of RAM for $35/mo
[01:48] <patapouf> Small Instance (Default) 1.7 GB of memory, 1 EC2 Compute Unit (1 virtual core with 1 EC2 Compute Unit), 160 GB of local instance storage, 32-bit platform
[01:48] <giovani> they're by far the cheapest, reasonably reputable, big VPS vendor I've found
[01:48] <patapouf> qman__ : I was looking for the small instance
[01:48] <qman__> I think what you need to do is find what your minimum specs really are
[01:49] <qman__> because I really doubt they're that high
[01:49] <qman__> and yeah, that's a great deal
[01:49] <patapouf> I need obviouly more storage than CPU
[01:49] <giovani> you do?
[01:49] <giovani> what are you storing?
[01:50] <patapouf> currently, I have 80Gig used (mostly SVN)
[01:50] <qman__> do you need root access on the same server as you need your data? because you can get shared web hosting with tons of space and bandwidth for free
[01:51] <giovani> haha
[01:51] <qman__> and servers with tomcat and all that jazz for about $5 a month
[01:51] <giovani> now we're easily getting into the not-reputable catagory
[01:51] <patapouf> I know .. but I need root access
[01:51] <patapouf> There is many stuff this server's doing which required me root access
[01:54] <qman__> all I'm really saying is, you need to find out exactly what your needs are, because you can save a lot of money if you do
[01:54] <qman__> if it really all has to be on the same server, with all that space, a dedicated server is probably your best bet
[01:55] <qman__> when I was looking, you could get C2D-class systems with 2GB RAM and 500GB hard drive space, unlimited bandwidth, for about $100 a month
[01:55]  * giovani pets his $7/mo dedicated server in Toronto
[01:56] <patapouf> giovani : pets ??
[01:56] <giovani> patapouf: ?
[01:56] <patapouf> pets his $7/mo dedicated server in Toronto
[01:56] <giovani> yes
[01:56] <giovani> it's my pride and joy
[01:56] <patapouf> pets is the name of the provider ? do you have al ink
[01:56] <giovani> you don't want it though
[01:56] <giovani> it's an old box
[01:56] <giovani> 80GB HD
[01:57] <giovani> and no link -- they no longer offer it
[01:57] <giovani> hostmds is the provider though
[01:58] <qman__> they do have some good plans though
[01:59] <giovani> I wouldn't recommend paying anything over $30/mo for their servers :)
[01:59] <giovani> their network is pretty crappy
[01:59] <patapouf> and their price is in Canadian Dollard  (which is good for me since I'm from quebec)
[01:59] <qman__> ah
[02:00] <giovani> I kind of collect cheap servers/vpses
[02:00] <giovani> it's a strange hobby of mine
[02:00] <patapouf> well then it's nice to have you here -- Thanks
[02:00] <giovani> corenetworks is another cheap provider
[02:00] <giovani> I have two of their $25/mo celerons
[02:00] <giovani> phenomenal service
[02:01] <giovani> they'll hook up an ipkvm whenever you need it for free
[02:01] <giovani> with virtual media support
[02:05] <wido> giovani: you collect cheap vpses?! Funny :)
[02:05] <giovani> wido: yes, I mostly use them as honeypots
[02:05] <giovani> for some of my security research
[02:05] <wido> ah, nice
[02:06] <wido> on different networks?
[02:06] <giovani> oh yes, that's the point
[02:06] <wido> cool!
[02:06] <giovani> to find as many disparate, cheap providers worldwide
[02:06] <giovani> international (outside of north america and europe) are of course the hardest to get for cheap
[02:06] <giovani> but I keep my eye out for deals
[02:06] <wido> disparate?
[02:06] <giovani> yes ... spread out
[02:06] <giovani> geographically
[02:06] <wido> ah
[02:07] <patapouf> giovani, where are you from ?
[02:07] <giovani> patapouf: US
[02:08] <wido> the research is scientific, as in we can enjoy your outcomes somtime?
[02:08] <giovani> wido: I mostly share amongst other security researchers -- particularly durring outbreaks, etc -- but every once in a while there's a unique attack that I write up
[02:09] <giovani> I use them to write 0day IDS rules for my employer mostly
[02:09] <giovani> but it's a hobby, and not official
[02:09] <wido> aha
[02:09] <wido> so it's not spam or viruses?
[02:10] <giovani> it's for anything
[02:10] <giovani> I do spam traps, low-interaction honeypots for collecting new viruses
[02:10] <giovani> and high-interaction systems to trap advanced attackers
[02:10] <wido> aha
[02:12] <giovani> wido: you interested in security?
[02:12] <bobobob> moin
[02:13] <bobobob> darn, I can't get the software RAID 1 to work... I guess my question is that once one of the disks fail, how do you replace it?  I had to make a / and swap on both disks....How do I tie in a replacements
[02:15] <wido> giovani: i work for an isp, but my profession is not security
[02:15] <wido> giovani: do you know http://iwatchedyourhack.org/ ?
[02:15] <wido> it is funny?
[02:16] <wido> ?=!
[02:17] <giovani>  nope, never heard of it
[02:17] <giovani> commercial or residential isp?
[02:17] <wido> commercial
[02:17] <wido> in the netherlands
[02:18] <giovani> cool
[02:18] <wido> the site might be interesting for you as a security expert
[02:18] <wido> it has nothing to do with my employer
[02:19] <wido> but it is funny
[02:20] <giovani> I'll definitely check it out
[02:20] <wido> this one was on a box of a colleague: http://iwatchedyourhack.org/node/8
[02:25] <shauno> heh, that is good.  I thought it was interesting when you see the direct switch from fumbling around with a very lose grasp for the environment, to just pasting in commands.  When it got to 'Shrek' .. hah
[02:25] <wido> :)
[02:25] <shauno> I'm curious, are they actually root under such a scenario?
[02:26] <wido> nope
[02:26] <wido> it's just a service
[02:26] <shauno> it really looks like all you've got to do is put # in PS1 and he wouldn't know the difference
[02:26] <wido> hence the odd shell
[02:29] <shauno> that is oddly interesting to watch tho.  he just tried to un-tgz a .pl.
[02:29] <wido> http://code.google.com/p/kippo/
[02:29] <wido> that's it
[02:30] <wido> kiddies
[02:30] <shauno> I may have to play with that sometime.  It could be more interesting than watching china fill my ssh log
[02:30] <wido> :)
[02:31] <shauno> (china rather than chinese, I've no idea if it's actually originating china or just coming via.  observation more than generalization)
[02:33] <wido> me either, but chinese IP's ..., thought china is big! :)
[03:01] <JasonMSP> how can I create multiuple access.logs for each site Im hosting and rotate those on a weekly basis like /var/log/auth.log  ?
[03:06] <wido> JasonMSP: first question, 'CustomLog /var/log/apache2/$access.log common' in your sites-available
[03:07] <wido> $access=unique per site
[03:08] <JasonMSP> wido:  here is an example CustomLog /srv/www/site.com/logs/access.log combined
[03:10] <wido> JasonMSP: here is an example CustomLog /var/log/apache2/mailman-access.log common
[03:11] <wido> and CustomLog /var/log/apache2/foto-access.log common
[03:11] <wido> and CustomLog /var/log/apache2/mail-access.log common
[03:11] <wido> and CustomLog /var/log/apache2/www-access.log common
[03:11] <wido> etc
[03:14] <JasonMSP> im not following
[03:15] <JasonMSP> each website has its own virtualhost file in the /etc/apache2/sites-enabled folder
[03:15] <wido> yep
[03:15] <wido> and they are generated by the sites-available
[03:16] <JasonMSP> id like to each week get rename the log access.log.old and start a new one.  and then delete the old .old file if there is one (on the 2nd week)
[03:16] <wido> in each sites-available file you can put an unique custom log
[03:17] <wido> JasonMSP: I'm not sure about the two weeks retention
[03:18] <wido> i answered your first question: unique logs
[03:18] <wido> the second one i'll have to check: more rotation
[03:20] <JasonMSP> each site already has a custom (unique) log.  I need to rotate them out so the file doesn't become enormus because it just keeps writing and writing to it
[04:02] <mnaines> How do I install Ubuntu Server on a machine with no graphics card?
[04:16] <qman__> mnaines, by installing a graphics card
[04:17] <qman__> or by creating your own live CD or liveUSB with a serial terminal or sshd running by default
[04:18] <qman__> installing with no video is one of those things that's possible but rarely necessary and therefore not covered in the normal installer
[04:19] <mnaines> Can Ubuntu Server run and be maintained without a graphics card after it is installed?
[04:19] <qman__> yes, if you install ssh
[04:19] <qman__> of course, you need to make sure your BIOS will boot the system without one
[04:20] <qman__> I have plenty of headless servers but none of them will boot without a graphics card installed
[04:20] <mnaines> I see.  Yeah...Mine throws a BIOS error 7F or 7E if I don't have a video card
[04:21] <mnaines> Basically BIOS error 7F on the board I use in the desktop means the video card is fubar
[06:22] <Psi-Jack> I'm running ubuntu 10.04-server under kvm in many instances, but kvm when it's sending the reboot or shutdown signal to the guests, they never even try to start shutting down as expected. I have CentOS servers as well, and they work without fail on this issue just not Ubuntu, so eventually libvirt has to force power-off the guests during a shutdown of the system.
[06:23] <Psi-Jack> Is there any resolution to this issue which is pretty big?
[06:25] <Psi-Jack> Aha... nevermind... Just resolved my own answer. needed to install acpid.
[08:30] <sherr> Psi-Jack: You might also want to install something like milly-guard on the host, so when the host is rebooted/shutdown, you can safely shut down the guests. Otherwise, their power is pulled - not good.
[08:30] <sherr> Who's "milly"? I meant : molly-guard!
[10:20] <matrix64> hey, how do I start upstart from recovery console?
[12:42] <JuJuBee> I am using NIS to handle log in for my students. However, yppasswdd is not running on the server and I cannot figure out why.
[13:15] <au> pths: please fix your connection
[13:50] <pmatulis> NIS is so old
[13:56] <willinja> pmatulis, what is nis ?
[14:00] <RoyK> lol
[14:00] <RoyK> !nis
[14:01] <RoyK> http://en.wikipedia.org/wiki/Network_Information_Service
[14:01] <RoyK> ubottu: nis is http://en.wikipedia.org/wiki/Network_Information_Service
[14:57] <Patrickdk> people still use NIS?
[15:02] <pmatulis> Patrickdk: yes, JuJuBee does
[15:06] <Patrickdk> heh, would think everyone would of probably dumped it for ldap
[15:11] <RoyK> Patrickdk: we also use nis - moving to AD these days, but with a hundred boxes from solaris 8 and up, it might take some time
[15:12] <willinja> where can i find ubuntu-server images that already configured php mysql torrentflux and samba ?
[16:57] <savid> Hi, I have a deployment directory on my server where website files are uploaded by different developers.   My problem is that I need the owner/group permissions to stay the same on all files within that directory, even files that are added.  Is there an easy way to manage this?
[16:59] <jo-erlend> chmod +s
[17:01] <jo-erlend> savid, mode s is for sticky. It does that. I think it's described in chmods manpage, but you'll find many references to it on the web as well.
[17:01] <savid> jo-erlend,  So if I do sticky on the web root directory,  any files added under that directory will inherit that permission?
[17:07] <jo-erlend> savid, sorry. I means setuid/setgid. Sticky is something else, but it is mode s and yes, it does what you want.
[17:07] <savid> Ah, ok
[17:07] <savid> Thanks!
[17:08] <jo-erlend> you'll have remember to set mode s on all the files and directory manually first though, since only new files and directories inherit the modes.
[17:09] <jo-erlend> it does completely different things if you apply it to an executable, so if you should have any executables in those directories, pay attention.
[17:12] <savid> jo-erlend, so (I'm assuming) for executables, it runs them as if they were run as the set uid?  that's why I need to be careful?
[17:13] <jo-erlend> right.
[17:16] <RoyK> jo-erlend: what does sticky do to an executable?
[17:17] <jo-erlend> I have no idea. I'm not sure it has any effect on executables at all.
[17:17] <jo-erlend> well... Any special effects, I meant.
[17:17] <p1l0t> I am trying to install server 9.04 (will upgrade once finished dont have anymore blank cds cant boot usb) on a compaq server with dual 18gb hdd but it is reading them as one 36gb drive and giving me an input output error no matter what I try :(
[17:18] <RoyK> jo-erlend: it did http://en.wikipedia.org/wiki/Sticky_bit - but never on linux, it seems
[17:18] <RoyK> p1l0t: probably raid0 setup on the raid controller
[17:19] <savid> Oh interesting, it seems that setuid on directories is ignored.
[17:19] <jo-erlend> savid, setgid works.
[17:19] <RoyK> savid: use ACLs if you want full control
[17:19] <p1l0t> RoyK: do I need to move a jumper or change something in the bios then?
[17:19] <jo-erlend> in your scenario, I would set s on the group, not the owner. I always do that.
[17:20] <RoyK> p1l0t: no, boot on a smartstart CD and configure the raid
[17:20] <RoyK> p1l0t: what sort of server is this?
[17:21] <p1l0t> Its a compaq stackable dual 1g processors
[17:22] <RoyK> does it have a raid controller?
[17:23] <RoyK> p1l0t: that is - cat /proc/partitions - what is the device names there?
[17:23] <p1l0t> Prolient DL360 it has some kinda of smart array thing
[17:23] <RoyK> ok
[17:23] <RoyK> reconfigure the raid
[17:23] <RoyK> from a smartstart cd
[17:23] <RoyK> setup a mirror or seperate drives if you like
[17:23] <p1l0t> /dev/ida/c0d0 - 36.4 GB Compaq Smart Array
[17:23] <RoyK> ok
[17:24] <RoyK> this is not about linux, it's the 'hardware' raid setup that's the problem
[17:24] <p1l0t> So I need to figure out how to 'unraid' it then first
[17:26] <RoyK> yes, boot on a smartstart cd
[17:26] <RoyK> you can download the images from HP
[17:27] <p1l0t> Ok thanks RoyK I will have to go buy some blanks then
[17:28] <p1l0t> That or unplug one of the harddrives and see what it does :)
[17:29] <b0gatyr_> Hello, I have set up a static ip on ubuntu server but after a few hours it goes back to dhcp although the interfaces file has the static address in it.. why is this?
[17:29] <dominicdinada> question.... why do things such as phpscripts require ftp access... for updates when they are using sockets? or what not. Better yet for instance wordpress to update it or do anything is asking for ftp access
[17:30] <dominicdinada> should i check apache for url_include or what not  ?
[17:30] <giovani> dominicdinada: what do you mean "they are using sockets"?
[17:30] <giovani> they want FTP access so they can upload new copies of the php files
[17:30] <giovani> which is an entirely dangerous thing to hand out on the web, but, people do it because it makes upgrading easy
[17:30] <dominicdinada> arnt they using fsockopen.... to transfer files. or say url include
[17:31] <giovani> I'm unclear on what you're talking about
[17:31] <giovani> a wordpress upgrade is simply a replacing of a few files
[17:31] <giovani> that's done over FTP, usually
[17:31] <dominicdinada> giovani correct in a production enviroment they do not require ftp access...
[17:31] <giovani> who is they?
[17:32] <dominicdinada> oh on a public site i have never been asked... for ftp access
[17:32] <giovani> I have no idea what you're talking about
[17:32] <dominicdinada> various php scripts
[17:32] <giovani> wordpress, the company
[17:32] <giovani> uses the FTP protocol, to put new copies of the wordpress php files onto your server automatically
[17:33] <giovani> this has nothing to do with php or scripts in general
[17:33] <dominicdinada> hmmm
[17:33] <giovani> just a particular method that wordpress, the company, uses to make upgrading easier and semi-automatic
[17:34] <dominicdinada> ugh and my ftp host is not allowing my user to access it lol wtf
[17:34] <giovani> how did you get wordpress onto the server in the first place?
[17:35] <dominicdinada> via download and placed the script and ran install
[17:35] <giovani> so do that again
[17:35] <giovani> you don't NEED to use their automatic update feautre
[17:35] <giovani> you can manually upgrade wordpress
[17:36] <giovani> through an almost identical process as installing it -- they have a howto on the website
[17:36] <dominicdinada> i was trying to avoid that
[17:36] <giovani> why?
[17:36] <dominicdinada> that is how i was doing it manually........ for the last xxx months but i guess i have no accounts setup for ftp access
[17:38] <ruben23> guys why VPS servers have very low memory allocation like, 256, 512Mb,1Gb- with this can this process apps like normal servers..?
[17:39] <giovani> ruben23: that's not "low" for the vast majority of uses
[17:39] <giovani> I run many servers with 64MB of ram
[17:39] <patdk-wk> heh, lucid uses 24megs on boot :)
[17:39] <ruben23>  giovani: whats apps are in there..?
[17:40] <giovani> ruben23: lighttpd,nginx,djbdns,openssh
[17:41] <giovani> patdk-wk: depends on your configuration
[17:41] <ruben23> web server..? you host your own or is it hosted..?
[17:41] <giovani> ruben23: it's a VPS ...
[17:41] <jo-erlend> a webserver doesn't require much ram at all.
[17:41] <giovani> well, apache is full of bloat
[17:41] <giovani> but good web servers can be quite efficient with ram
[17:41] <jo-erlend> well... In normal instances. If you have lots of visitors, then it uses more.
[17:41] <giovani> nginx and lighttpd are good examples
[17:41] <patdk-wk> giovani, well, I mean just the bare needed stuff, for ubuntu-server
[17:42] <giovani> patdk-wk: but that's flexible
[17:42] <patdk-wk> add ontop of that whatever your server does
[17:42] <giovani> I do a lot of stripping down
[17:42] <ruben23> oh ok, now how VPS differs with cloud servers..?
[17:42] <jo-erlend> ruben23, it doesn't.
[17:42] <giovani> ruben23: there's no difference theoretically
[17:42] <giovani> in practice, most VPSes you buy are not fault-tolerant
[17:42] <giovani> while "cloud" servers like EC2 usually are
[17:43] <patdk-wk> heh? ec2 is not fault tolerant at all
[17:43] <jo-erlend> ruben23, the cloud just makes it less tangible, which is the whole point of the cloud. You don't worry about how it happens, as long as it does.
[17:43] <giovani> this is just industry practice and naming though
[17:43] <giovani> patdk-wk: they're run on live-migration hardware ... so, they are
[17:43] <giovani> most of the "VPS" industry runs your VPS on a single, non-portable server
[17:43] <patdk-wk> ec2 clearly states you can lose your instance, disk, ... at any point
[17:43] <giovani> patdk-wk: what they guarantee and what is implemented are different things
[17:55] <dominicdinada> !ftp
[17:55] <dominicdinada> !ftp server
[17:55] <dominicdinada> !FTP
[18:00] <rutri> hello, i need help getting my server up and running
[18:01] <giovani> rutri: that is an incredibly general request
[18:01] <rutri> ok
[18:02] <rutri> I need help getting access to KVM using virt-manager
[18:03] <rutri> I have followed all of the instructions on the Ubuntu server document site including adding my user to the libvirtd group however my connection is still getting refused when I use virt-manager form my desktop computer.
[18:06] <rutri> can anyone help me?
[18:13] <yann2> hi rutri
[18:13] <yann2> you do know that even once you have been added to the group
[18:13] <yann2> you need to logout and login again so that it is effective ? :)
[18:13] <yann2> maybe its just that
[18:13] <rutri> yann2: i rebooted after I set it all up is that enough?
[18:14] <yann2> yes :P
[18:14] <rutri> yea :( i have rebooted several times and added the sysadmin account as well as my desktop users account.
[18:15] <rutri> I have also disabled the firewall, if you can give me that link where I can post large bodies of text I will give you the virt-managers detailed report.
[18:16] <Znuff> Hi. Any ideas how can I install pdo_sqlite under ubuntu-server 10.04?
[18:16] <Znuff> all I get is configure: error: Cannot find php_pdo_driver.h.
[18:19] <rutri> any other suggestions?
[18:23] <rutri> yann2 you still there?
[18:25] <BobSapp> hey there, I want to roll out a very minimal ubuntu server cloud nodes for ai purposes.  As a result they only really need to have perl and some cpan modules installed.  no services other than ssh for communication.
[18:26] <BobSapp> I was thinking about using EC2 to host these operating systems.  Can I do this with ubuntu cloud or would I have to instanciate an ubuntu server every time and tell it to install the packages i needed each time?
[18:30] <rutri> Does anyone know why libvirtd would not be installed if I installed libvirt-bin?
[18:31] <yann2> rutri, I guess you might want to not run libvirtd, but just the tools to connect to an instance somewhere else, if I understand it right?
[18:32] <rutri> ok then what is used as the interface for qemu+ssh?
[18:34] <rutri> yann2: this link is what I used to set up the server http://www.ideyatech.com/2010/05/virtualization-with-ubuntu-1004-lucid-lynx/
[18:35] <rutri> I have KVM and libvirt-bin tools installed however when I try to connect with the manager its like the server is not there
[18:36] <rutri> I have added my users to the libvirtd group and reset the server several times with no luck.
[18:39] <SpamapS> wow.. booting a lucid t1.micro on ec2 takes about 45 seconds from run-instances to ssh being available
[18:40] <ph8> hi all, i'm having an ubuntu server nightmare, i seem to have lost my raid config between a hardy->lucid upgrade - i see no md devices on boot, any thoughts on how i can get them working again from busybox?
[18:40] <ph8> the disks don't appear to have uuid's, although they're all there as devices (sda*) which is odd
[18:40] <SpamapS> ph8: whats in /proc/mdstat ?
[18:40] <ph8> just the personalities: [linear] [multiple] [raid0] [raid1] etc... line and unused devices: <none>
[18:40] <ph8> the system drops to busybox because it can't find /dev/md0 to boot to
[18:41] <SpamapS> ph8: can you paste the output of 'fdisk -l sdX' on paste.ubuntu.com ?
[18:41] <ph8> it would appear i don't have fdisk on busybox
[18:41] <SpamapS> ph8: right.. hm
[18:41] <ph8> i see 3 sd partitions on each disk in /dev
[18:41] <ph8> which seems right
[18:41] <ph8> and /etc/mdadm/mdadm.conf looks ok
[18:41] <SpamapS> do you have mdadm?
[18:41] <ph8> well, it's configured by uuid
[18:41] <ph8> yes i have mdadm
[18:42] <SpamapS> ok.. hmm
[18:43] <ph8> so i guess i could try and recreate my arrays
[18:43] <ph8> with mdadm
[18:43] <ph8> but in /dev/disk
[18:43] <ph8> i don't have a by-uuid folder
[18:43] <ph8> but i do have a by-id
[18:43] <ph8> any thoughts on why that might be?
[18:44] <SpamapS> do mdadm --examine /dev/sdX
[18:44] <ph8> will do when it's back in a sec
[18:44] <SpamapS> you shouldn't need "uuids"
[18:44] <SpamapS> you just need raid superblocks
[18:44] <SpamapS> which you should most definitely have
[18:45] <SpamapS> the examine should show the UUID's of each raid member that it expects to see
[18:45] <SpamapS> but those are the raid superblock UUID's
[18:45] <ph8> ah ok, that's what's in mdadm then?
[18:45] <ph8> not the disk uuid's as i thought
[18:45] <ph8> * mdadm.conf
[18:46] <SpamapS> if the partitions aren't set to FD (raid autodetect) then the md driver won't detect them and try to assemble from them
[18:46] <ph8> would that have been changed in a dist-upgrade though?
[18:46] <SpamapS> no
[18:46] <SpamapS> but maybe they weren't created properly
[18:46] <ph8> hmm maybe, they have worked in the past
[18:46] <ph8> just waiting for someone in the dc to reboot the server
[18:46] <SpamapS> its entirely possible you've been relying on the device ordering or something in the initrd ..
[18:46] <SpamapS> ph8: you upgraded remotely?
[18:47] <ph8> with kvm access
[18:47] <ph8> still not my ideal
[18:47] <ph8> it actually went fine on 3 other boxes, this was the last one (typical)
[18:47] <SpamapS> ph8: yeah, thats a little scary. ;)
[18:47] <SpamapS> ph8: this has come up a few times, so I'm wondering if we need to do some fixing for 10.04.2
[18:47] <SpamapS> ph8: once this is figured out, it would be REALLY great if you could put together a bug report.
[18:48] <ph8> yeh i'd love to
[18:48] <ph8> if/when i get it fixed :-s
[18:48] <SpamapS> right :)
[18:48] <ph8> i should know better than doing an upgrade on a saturday afternoon when i'm expected at dinner in about 2 minutes (1 hour drive away) :p
[18:49] <SpamapS> heh.. been there, done that.
[18:49] <rutri> anyone here familiare with libvirt?
[18:49] <ph8> i have used it in the past
[18:49] <ph8> they've got a good support channel on irc.oftc.net in #virt as well
[18:50] <rutri> ok thanks
[18:50] <rutri> the #virt channel may be able to help me
[18:51] <rutri> how do i get into the #virt channel its password protected.
[18:52] <SpamapS> rutri: on OFTC or on Freenode?
[18:52] <rutri> im on freenode
[18:52] <ph8> SpamapS:  back in
[18:52] <SpamapS> ph8: mdadm --examine /dev/sdX
[18:52] <ph8> so on sda2
[18:52] <ph8> i have 2 active devices
[18:52] <ph8> and sda 5
[18:53] <ph8> and sda6
[18:53] <ph8> i have an sda1 but i don't think there should be any active devices there
[18:53] <ph8> so 3 partitions of raid superblocks
[18:53] <ph8> any thoughts on how to proceed/
[18:53] <ph8> some sort of mdadm --doallthemounting ?
[18:54] <ph8> gargh, are there no text editors in busybox
[18:54] <ph8> it looks to me like mdadm.conf has the same section twice for raid config
[18:54] <ph8> so i need to just take out the bottom section and that's a start if nothing else
[18:54] <ph8> but how do i text edit!
[18:55] <SpamapS> ph8: does it show the UUID's of the other expected members?
[18:55] <SpamapS> mdadm.conf is sort of inconsequential if you have FD type partitions. I'm surprised you don't have fdisk.
[18:55] <ph8> yes, the uuids are the same in both sections
[18:56] <ph8> fdisk isn't there when i type 'help' and the command doesn't work :(
[18:56] <SpamapS> ph8: you should be able to have mdadm output the correct config file once you assemble the arrays properly
[18:56] <ph8> ok so i'll use mdadm to assemble my arrays?
[18:56] <SpamapS> ph8: you can use mdadm --assemble I believe
[18:56] <SpamapS> ph8: got backups?
[18:57] <ph8> no actually
[18:57] <ph8> :-s
[18:57] <ph8> i should know by now
[18:57] <ph8> i've only been doing this 10 years
[18:57] <ph8> i can pass assemble a uuid
[18:57] <ph8> worth a go?
[18:57] <ph8> do i actually risk dataloss doing an assemble? won't it reject mismatching arrays?
[18:58] <SpamapS> wtf, no backups?
[18:58] <SpamapS> yes you do risk fat fingering something and messing things up
[18:58] <ph8> yes i know :/
[18:58] <ph8> /dev/md0 is started with two drives!
[18:58] <SpamapS> just don't use --force ;)
[18:58] <ph8> using --assemble /dev/md0 --uuid <uuid>
[18:58] <SpamapS> yeah that should work. :)
[19:00] <ph8> then i just exit busybox and it boots?
[19:00] <SpamapS> dunno abt that
[19:00] <ph8> lets see
[19:00] <ph8> hurray
[19:00] <ph8> it's booting
[19:01] <ph8> so do you think it's just a malformed mdadm.conf?
[19:01] <rutri> ph8 no one is talking in the OFTC #virt channel, does it usually take them long to respond?
[19:01] <ph8> yes sorry they do take their time
[19:01] <ph8> but they do respond
[19:01] <ph8> what's your question maybe i can help
[19:02] <rutri> ph8  have followed this guide http://www.ideyatech.com/2010/05/virtualization-with-ubuntu-1004-lucid-lynx/ and tried to get libvirt working on my Ubuntu 10.4LTS server however I cannot connect to it.
[19:02] <SpamapS> ph8: I'd verify that your initrd is up to date and that mdadm.conf is right
[19:02] <SpamapS> ph8: yous hould also check out the partition types, make sure they're FD
[19:03] <SpamapS> For all I know, they've done something weird with the initrd and don't use FD type partitions anymore.
[19:03] <ph8> yes they are autodetect
[19:03] <SpamapS> ph8: I suppose thats why --assemble worked actually..
[19:03] <SpamapS> ph8: not sure why those weren't automatically started and run :-P
[19:04] <SpamapS> ph8: I know you're under the gun, but if you can save that bad mdadm.conf, and maybe attempt to recreate the problem.. it might save a few other people from having this problem as people start moving from hardy to lucid
[19:06] <ph8> indeed
[19:08] <SpamapS> ph8: my guess is that hardy saved a bad madadm.conf, which was never used before.. and now lucid uses it, and fails
[19:09] <ph8> the upgrade uninstalled mysql as well!
[19:09] <ph8> very odd
[19:13] <BauerUK> are there no alternate downloads (mirros/bittorrent) for Ubuntu server? the http download is running at ~50 KB/s
[19:14] <ph8> bittorrent will likely be much faster
[19:14] <ph8> i know i'm seeding on a 100mbit pipe for example
[19:14] <BauerUK> could you link to the torrent? i don't see it on the download page
[19:14] <ph8> sorry i'm a little busy
[19:14] <ph8> pretty sure it's there
[19:15] <rutri> ph8 did you have any suggestions for my problem?
[19:16] <ph8> what do you see when you run 'virsh list' rutri ?
[19:16] <rutri> do i run it on the client or the server?
[19:16] <rutri> oh wait
[19:17] <rutri> I ran virsh -c qemu:///system list on the server and got nothing running witch I would expect because I have not set any VM's up.
[19:17] <BauerUK> nevermind, i got it. the link isn't on the server download page itself, but available through an unrelated downloads page
[19:19] <rutri> ph8 it seems like its running however it just not listening on the management port?
[19:22] <ph8> is it all on one server?
[19:22] <ph8> or are you administering multiple libvirt daemons from one machine?
[19:24] <rutri> ph8: I want to set up multiple VMs one one machine, its going to be a virtual server.  but as far as i know there is only one libvirt daemon running.  How do I know if the daemon is running properly?
[19:24] <ph8> i'm sorry i'm due out
[19:24] <ph8> apologies for not being more helpful
[19:24] <rutri> meh its ok
[19:25] <rutri> I may just go back to a windows server, its way easier to set things up there.
[19:36] <rutri> I have a n ssh question that my be causing the problem.
[19:38] <rutri> if my ca certificate has changed how do i update it on my desktop computer?
[19:42] <RoyK> what?
[19:42] <RoyK> you mean if ssh complains about something bad?
[19:42] <rutri> yes
[19:42] <rutri> because my key is outdated
[19:42] <RoyK> vi $HOME/.ssh/known_hosts
[19:43] <rutri> how do i force ssh to update the certificate?
[19:43] <RoyK> just remove the old key
[19:43] <rutri> cool thanks one second
[19:43] <RoyK> vi $HOME/.ssh/known_hosts +linenumber
[19:43] <RoyK> dd
[19:43] <RoyK> :wq
[19:44] <rutri> ok that worked, I figured out my problem, now I am able to connect to my virtual host
[19:44] <rutri> thank you very much Royk and ph8
[19:44]  * RoyK opens a beer
[19:44] <RoyK> rutri: np
[19:45] <rutri> Ryok I have been trying to get this VM server up an running the past week lol i didnt know changing my CA key would break it.
[19:46] <rutri> Ryok but it makes sense since it tunnels through ssh.
[19:48] <RoyK> I guess you'd want to know if your server's ssh key suddenly changed
[20:01] <bobobob> I am trying to learn how the LAMP system is put together....It seems that /var/www is setup for root 755 and if you add something like wordpress, how do you change permissions for wordpress to be able to write to its cache directories?  Seems like there should be a user associated with the running of wordpress, how do I find that?
[20:01] <bobobob> is there a wordpress user that it runs under?
[20:03] <rutri> RoyK: yea but Linux nubs like me don't think of these things lol.
[20:09] <RoyK> hehe
[20:09] <RoyK> perhaps you'll think of it next time
[20:10] <RoyK> bobobob: it all runs under user www-data
[20:10] <RoyK> give that user write access to the specific directories it needs to write to
[20:10] <RoyK> not all of the wordpress install
[20:12] <RoyK> allowing www-data to write to the whole wordpress install might compromise your system when a security issue comes up
[20:12] <RoyK> note, not if, when
[20:12] <bobobob> okay...RoyK...thanks
[20:12] <rutri> Royk: Oh believe me, if it uses and SSH tunnel I will be making sure the certificates are good.
[20:14] <RoyK> bobobob: it's not about wordpress being insecure, it's about _all_ systems being insecure - don't take chances
[20:59] <atm0s> Question, I made a copy of my ISCSI image on which Ubuntu Lucide Server was installed and created a new ISCSI target with that copy, what files in that filesystem should I change, to make Ubuntu know to boot from the NEW target, not the OLD..? /etc/iscsi didn work..
[21:00] <guntbert> atm0s: but here too: you will only get an answer if someone has to say anything about that - you are not being ignored
[21:00] <atm0s> great :)
[22:08] <Qix> I am having uber issues -- my linux box recognizes ethernet devices but wont connect at all.
[22:08] <Qix> /etc/network/interfaces was configured like a guide told me but still no beans
[22:08] <Qix> any help? >.>
[22:09] <Qix> lawl wrong button.
[22:09] <Qix> any help?
[22:09] <guntbert> Qix: what guide was that?
[22:09] <Qix> err lemme find it.
[22:09] <Qix> http://www.cyberciti.biz/faq/setting-up-an-network-interfaces-file/
[22:09] <Qix> ^
[22:11] <guntbert> !serverguide | Qix look at that please
[22:14] <ecelis> ~/win 12
[22:14] <ecelis> oops
[22:19] <Qix> The problem is is that DHCLIENT fails.
[22:19] <JasonMSP> im installing postfix and the first screensays, "Please select the mail server configuration type that best meets your needs."  with <ok> at the bottom.  Im connected SSH and I cant click, select ok.  i've tried enter.  That didn't work.  and I don't have a cursor inside the window either.
[22:20] <JasonMSP> ctrl c, q or x doesn't work either.
[22:20] <guntbert> Qix: do you have a DHCP server running in the network?
[22:20] <Qix> yes.
[22:20] <Qix> All of my windows boxes connect just fine.
[22:21] <guntbert> JasonMSP: <tab> should move the highlight
[22:21] <JasonMSP> THANK YOU!
[22:21] <Qix> and the server now running ubuntu server, the box in question, had windows server 2008 on before it, and it connected just fine.
[22:21] <JasonMSP> that was nerve racking!
[22:21] <guntbert> JasonMSP: :-)
[22:22] <Qix> dhclient fails with "No DHCPOFFERS received"
[22:23] <guntbert> Qix: have a look at /var/log/syslog, and are you sure that the interface is "up"?
[22:23] <Qix> Ill take a look, and yes ifup returns that the interface is up.
[22:24] <Qix> guntbert - syslog shows DHCLIENT says : ADDRCONF (NETDEV_IP): eth1: link is not ready
[22:24] <Qix> seems to be the first error.
[22:25] <Qix> eth1 is the NIC card Im using.
[22:25] <Qix> wanting to use, anyway.
[22:25] <guntbert> Qix: link is not ready: that could mean there is no cable in the nic
[22:26] <Qix> just checked -- its in there, and all four lights are on
[22:26] <Qix> and my router's light for that port is on.
[22:26] <Qix> which is better than before when I was using the on-board nic port
[22:26] <Qix> my router didn't even acknowledge it was connected.
[22:27] <Qix> and I tried plugging the same eth cable into a working, tested laptop and I got a link right away
[22:29] <Qix> would plugging my modem directly into the box help?
[22:29] <Qix> it may be my router refusing the connection.
[22:31] <Qix> guntbert?
[22:34] <guntbert> Qix: are you certain its eth1? you could try: tailf /var/log/syslog     and then plug/pull the cable and watch the log
[22:35] <Qix> weird
[22:35] <Qix> it freaks out when I plug it into the default port.
[22:36] <Qix> instead of the PCI nic
[22:36] <Qix> ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready
[22:36] <Qix> >.>
[22:36] <Qix> lemme take a look
[22:37] <Qix> k so it looks like my onboard is actually eth1
[22:39] <Qix> alright well now
[22:39] <Qix> when i plugged it into eth1, and ran ifconfig it showed it assigned an IP
[22:40] <Qix> let me look at my router...
[22:40] <guntbert> Qix: all ok now?
[22:40] <Qix> alright sorta
[22:40] <Qix> router is now showing the computer's hostname, etc.
[22:40] <Qix> but when i try to ping it shows host is unreachable
[22:40] <Qix> ping google.com
[22:40] <Qix> doesn't work
[22:41] <Qix> same with pinging my gateway IP
[22:41] <Qix> Destination Host Unreachable
[22:41] <guntbert> and Qix for the next time: please use the nick of your partner in most/every line  -- I'm like following about 10 channels right now and without highlight it might take a long time till I see your answer
[22:41] <Qix> ah kk
[22:42] <guntbert> Qix: start pinging the own IP address
[22:42] <Qix> guntbert: the computer's IP address?
[22:42] <guntbert> Qix: yes
[22:42] <Qix> guntbert: kk thats working.
[22:43] <Qix> guntbert: but pinging the router or google.com or my other computer is not working.
[22:43] <guntbert> Qix: good, can you ping any other host on your home network?
[22:43] <Qix> guntbert: nope
[22:43] <guntbert> Qix: then I suspect the switch might make troubles
[22:44] <marc__> hi all
[22:44] <Qix> guntbert: probably. Should I try connecting the modem directly to the box?
[22:45] <Qix> ifconfig
[22:45] <Qix> blah
[22:45] <Qix> wrong keyboard
[22:45] <guntbert> Qix: not sure, but restting the switch might be a good idea (or are all your computers connected to the router?)
[22:46] <Qix> guntbert: they are all connected to the one router
[22:47] <guntbert> Qix: then look into that router - config, firewall, reset it....
[22:47] <guntbert> !hi | marc__
[22:47] <marc__> where should I put things I want all website to have access?  I want all virtual host to be able to access files like PHP library (ie: zend framework, cakephp) for read only.
[22:48] <Qix> guntbert: alright, Ill play around with it. Ill try connecting directly to the internet to see if that works. Brb...
[22:49] <guntbert> marc__: that looks like the perfect case for nfs...
[22:50] <marc__> for now it will be on the same server as the virtual hosts.  (got only 1 server)
[22:52] <marc__> I'm thinking /usr/share/php maybe?  or /var/www/shared ....  what I want is to be able to create a new virtual host for a website and give access to the libraries without having to install them in each folders...
[22:52] <Qix> guntbert: kk, I connected it directly to the internet, and I could ping google.com no problem
[22:52] <guntbert> marc__: but nfs is independant of the virt product you use
[22:52] <Qix> its my router.
[22:52] <guntbert> Qix: seems so
[22:53] <Qix> guntbert: I'll reboot my unbuntu box to make sure its all working by default, and then play around with the router settings.
[22:53] <guntbert> Qix: good way to take
[22:53] <Qix> :]
[23:10] <Qix> guntbert: Everythings working :] just needed a reboot it looks like
[23:10] <Qix> thanks for the help!