=== wedgwood is now known as wedgwood_away [00:06] sarnold, what is /etc/timezone used for? it contains the textual zone info on a single line. [00:07] copying over /etc/localtime worked fine and every new process gets the correct time, but this file still has the old zone mentioned. will this need to be overwritten as well? [00:13] sarnold: unfortunetly I do not have xterm on my working server install; I tried iconv " for i in 'iconv -l | less | sed 's/.\{2\}$//' | less'; do iconv -c -f $i -t utf8 /FILEPATH/" unfortunetly my bash skills are poor; could you help me to write a valid expression? [00:16] alllex: oh! silly me. I'd not noticed that /etc/timezone no longer has the zone data from /usr/share/zoneinfo/ itself. [00:16] alllex: I'm sorry I gave you bad advice. [00:17] sarnold, nevermind. i peeked into /var/lib/dpkg/info/tzdata.postinst [00:17] alllex: the zoneinfo files have a compiled description of offset from UTC, and at specific times what to change the offset to... [00:17] it is simply echoing the zoneinfo into /etc/timezone [00:18] sarnold, was that ^^ for me? if yes, i know what the zoneinfo files are :) [00:18] azbyin,alllex: *sigh* nickcollides in my brain :) sorry both. [00:19] haha [00:20] nevermind, :-) [00:20] alllex: got it? [00:21] also, the recommendation is to copy over /etc/localtime instead of a symlink as some progs apparently *may* try to modify /etc/localtime and they will end up modifying the olson tzdata [00:21] azbyin: yikes :) [00:21] :-) [00:22] indeed! anyway, the default is always to have that as a regular file. also noticed that the postinst actually creates a copy /etc/localtime.dpkg-new and then mv's the file to /etc/localtime [00:23] probably to ensure the tzdata file actually exists before making a symlink to a non-existent file [00:25] azbyin: probably also to avoid funny problems with filesystem races and loss-of-power.. a mv ought to be atomic, but opening, truncating, re-writing, and closing a file isn't atomic... [00:47] hi, after changing the modem eth0 doesn't come up [00:48] I looked at /etc/netwrork/interfaces, it's all normal eth0 is set to auto and dhcp [00:48] right after ubuntu starts to run the port on modem stops blinking as if this cable is disconnected [00:48] when I do ifconfig I don't get an IP, any ideas? [00:56] I asked my question here anybody care to answer? http://askubuntu.com/questions/261332/no-network-on-ubuntu-server-after-changing-the-modem === histo3 is now known as histo [01:15] hello - I cannot seem to be able to receive the 40976EAF437D05B5 apt key from any server - can someone point one that works pls? [01:16] Alysum: What archive/repo are you tyring to verify? [01:17] the official ubuntu one [01:18] basically when I run sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-key 40976EAF437D05B5 [01:18] it just hangs [01:20] Why are you trying to add a key for the official repo? [01:20] because I got a BADSIG and I deleted the key - need to readd it [01:21] I tried all the tips from google and Im stuck on receiving the key [01:23] Alysum: what did you delete exactly [01:24] histo: apt-kety del 437D05B5 [01:24] which is Ubuntu Archive Automatic Signing Key [01:24] !gpgerr | this may help [01:24] this may help: Getting GPG errors after adding custom repositories? Find the GPG keyword for the repository (it's 437D05B5 for the standard ones) and run « sudo apt-key adv --recv-keys --keyserver keyserver.ubuntu.com » [01:25] that's exactly what isn't working.... [01:25] now that you guys are all here do you mind taking a look at my Q? should I re-say it? [01:25] Alysum: No your command was different you are not using the keyword [01:26] Nicolas_Leonidas: what does your /etc/network/interfaces look like? [01:26] apt-key adv --recv-keys --keyserver keyserver.ubuntu.com 40976EAF437D05B5 [01:26] histo: It looks like it's the default, I pasted it here http://askubuntu.com/questions/261332/no-network-on-ubuntu-server-after-changing-the-modem [01:26] it just hangs [01:27] Alysum: try the command right from ubottu notice the keyword is 437D05B5 [01:28] Nicolas_Leonidas: is your router set to dhcp? [01:28] histo: It is [01:29] still hangs - could it be because I deleted it so it has nothing to compare it with ? [01:29] Nicolas_Leonidas: Does ifconfig show a link? [01:31] histo: it shows eth0 and Local Loopback, eth0 has HWaddr, but no inet addr, no mask, , it says UP BROADCAST MULTICAST MTU:1500 [01:31] I wish I could copy and paste but that server is not connected to network so I can't ssh [01:32] Nicolas_Leonidas: if you run dhclient does it obtain an ip or error? [01:32] let me try that [01:33] Nicolas_Leonidas: also what if you statically assign an ip within the routers network range and proper gateway etc... does it work? [01:34] Alysum: hrm. I was able to re-install the key in ~one second... [01:34] Alysum: 'host keyserver.ubuntu.com' shows two IPs.. I wonder if one or the other of them times out for you? [01:34] Alysum: can you ping keyserver.ubuntu.com ? [01:35] ok I need to find somewhere where I can download the 437D05B5 key then add it manually [01:35] sarnold: I can telnet keyserver.ubuntu.com 80 [01:35] histo: dhclient eth0 is still running no output yet,I haven't tried the static IP address, I will try it [01:37] Alysum: and when you try to interact with the server there, you get the results you expect? or errors? or _sometihng_ to show you've connected at least? :) [01:37] Nicolas_Leonidas: Did you change the cable when you changed the modem? Does the modem show a link light and the nic card show a link light? [01:37] histo: I did not change the modem, the link light goes dead after ubuntu starts and tries to configure virtual network devices [01:37] both on modem and server [01:38] I see an option to flush dhcp cache, maybe I should try that? [01:38] dhclient eth0 still running, is this normal? [01:38] Nicolas_Leonidas: With no output whatso ever? [01:38] histo: correct, no output [01:39] can I find the key online somewhere ? [01:40] Alysum: http://paste.ubuntu.com/5566546/ [01:40] Nicolas_Leonidas: No that is not normal [01:41] Nicolas_Leonidas: When you say you upgraded your internet connection what exactly does that mean? [01:41] Nicolas_Leonidas: You said you changed the modem also??? [01:41] histo: that means they took my modem, gave me a new one, and my connection is faster now, 50MBPS Down 10MBPS up baby [01:42] MBps and not Mbps? :) [01:42] sarnold: thanks - turns out it was a firewall issue....I thought I only needed port 80 open but hkp uses something else! [01:42] Alysum: aha! :D [01:42] sarnold: yes, THAT! [01:42] Nicolas_Leonidas: _drool_ :D [01:42] Nicolas_Leonidas: Do any other ports ont he modem work with other PCs? [01:42] histo: they do, PC works, Play Station works too [01:43] iphone and printers use the wifi as well [01:43] dhclient eth0 stopped with no output [01:44] histo: so see when I Restart the machine, lights start blinking then ubuntu starts and they both die [01:44] I'm looking at modem logs, it doesn't seem like that server has reached the modem for requesting a dhcp lease [01:44] Nicolas_Leonidas: try one of the cables and ports that is known working on the ubuntu machine [01:44] anyway thanks! [01:45] Nicolas_Leonidas: sounds like a bad cable to me. [01:45] histo: did that, just to make sure the cable is working, it was working with PS, now Ubuntu is using that cable, and PS is running fine with what Ubuntu was using [01:45] Nicolas_Leonidas: did you switch ports on the modem? [01:45] histo: yeah, it sounds kinda hardcore, I didn't even know you can turn off a light on a modem with software :) [01:46] histo: hmmm no, let me try that, I might have accidentally though [01:46] Nicolas_Leonidas: Was ubuntu working prior to this modem switch? [01:46] histo: it was [01:46] Okay well it has to be hardware then [01:47] If the software didn't change the only thing you changed was hardware [01:47] Nicolas_Leonidas: Are the TX and RX fields of ifconfig showing any traffic at all? [01:47] he took the old modem/routher away otherwise I would have used the same modem to see [01:47] genii-around: He has no link period [01:47] Nicolas_Leonidas: did you see if the port is dead on the modem? [01:47] genii-around: 0 for eth0, some for loopback [01:47] histo: changed the port, changed to a port that was just working no difference [01:48] Nicolas_Leonidas: Does sudo ifconfig -a show an eth1 or eth2 that maybe it was plugged before? [01:49] genii-around: no just eth0 and lo [01:49] Hm. [01:49] is loopback a normal thing for an OS to make? windows does that too probably right? [01:49] that just makes localhost and 127.0.0.1 and such work correct? [01:50] Have you tried a crossover cable? [01:51] genii-around: no why would I try that? the cable is working I switched it with another computer [01:51] Nicolas_Leonidas: Because some older adapters can't automatically tell how the cable is wired [01:52] I had this issue before with many Ovislink modems [01:53] I'll try that after everything else fails because I'll have to move the server closer to the modem [01:53] Nicolas_Leonidas: does any other device plugged in at the server show a linklight? [01:54] ( eg: straight-through 568B - no go, x-over 568B - worked ) [01:54] sarnold: any device plugged in the server or router? [01:54] Nicolas_Leonidas: plugged into the router, but at the server.. [01:54] (does that entire run of cable and port on the router _work_) [01:55] sarnold: I'm sorry I don't understand the router has 4 ports, 2 of which are connected they have linklights blinking, this ubuntu server is also connected but not blinking [01:55] it stops blinking at boot [01:56] Nicolas_Leonidas: if you unplug the server and plug in a laptop, does the laptop work? [01:56] sarnold: I did that and it worked, but let me try again one more time [01:56] hunh. [01:57] I'm guessing bad connection on a cable end [01:57] Has to be [01:57] Is the modem bridged and using PPPOE login on each client? Bell for instance allows 3 MAC only this way, 4th fails. [01:57] Ubuntu didn't change the modem did [01:57] possibly the cable got bad from being moved etc... [01:57] it has a bad end most likely [01:59] hey guys. does anyone know what kind of hardware / driver issue would cause a permenant uninterruptible state on a process where, according to ps wchan, the process is in the syscall lock() ? [01:59] er I mean fork() [02:00] sad to report that it does in fact work with laptop [02:01] Nicolas_Leonidas: woo. that's a help. can you throw a crossover between the laptop and the server? [02:01] Nicolas_Leonidas: What are the IPs of the other boxes attached? eg: LAN type IP or outward facing type IP? [02:02] ( if outward facing types, likely the bridged situation and the servers MAC is not allowed because too many concurrent logins ) [02:02] genii-around: LAN type, 192.168.2.24 on this windows [02:02] Nicolas_Leonidas: OK. [02:02] sarnold: good idea let me see if I have one [02:03] Nicolas_Leonidas: (gigabit cards won't require crossover, they do that automatically..) [02:04] sarnold: wow that's cool [02:04] sarnold: I don't think that's a spec of gigabit cards, though most do, but I think what you're referring to is MDI/MDI-X [02:05] jetole: no kidding? I thought it was part of the spec of gigabit.. [02:06] sarnold: I said I don't think so I'm not sure but MDI/MDI-X is the term for auto detect and adjust for crossover [02:07] ok so laptop connected to server worked, they both get IPS, there is traffic in TX and RX [02:07] can we conclude it's not the cable, nor the NIC of the server? [02:07] jetole: gigabit uses all pairs in all directions I believe [02:07] jetole: http://en.wikipedia.org/wiki/1000BASE%E2%80%91T#1000BASE-T [02:07] Nicolas_Leonidas: do you get the 169 auto-config addresses? or is one host handing out dhcp? or .. [02:08] jetole: so MDI/MDI-X doesn't make any sense at all, thats 100Mbit and below. [02:08] lifeless and sarnold: it's a sad day when you don't learn something new. cool [02:09] jetole: indeed, and today I learned two new things :) gigabit does, but in a different way than 100. haha. :D [02:09] now we just need to apply this to USB micro so I can stick my android foot up the !@# of all my iPhone co-workers [02:09] sarnold: not quire sure what's happening one is a mac laptop, it says using DHCP, IP 169.254.178.29 the server has an ipv6 ip no ipv4 apparantly [02:09] Nicolas_Leonidas: 169.254 is a anycast [02:09] I think for bonjour [02:09] mDNS [02:09] er... not anycast [02:09] ... well maybe. don't know but yeah, it' for bonjour [02:09] It's fallback when no dhcp server [02:10] why didn't occur to me to restart the modem at all? [02:10] jetole: 169.254 is zeroconf, for DHCPless networks. [02:11] Nicolas_Leonidas: ok. I was worried that perhaps one of them was _also_ doing dhcp and making a royal mess. [02:11] jetole: http://en.wikipedia.org/wiki/Zero_configuration_networking#Address_selection [02:12] I have done some relatively advanced things such as bind9 on this ubuntu server, I wonder if there is another file like /etc/network/interface to look at to see if I have changed anything [02:13] lifeless: I'll take your word for it. I don't know the specifics or care though... isn't bonjour zeroconf? anyways. I know it's useful on those nets where they use some protocol for people who can't get a IP and don't know how to use packet capture [02:13] Nicolas_Leonidas: I don't work on those nets [02:16] the problem starts in boot with "starting configure virtual network devices" I can see the linklight going dead right then [02:16] what is a virtual network device do I need one? [02:17] Nicolas_Leonidas: ip -o link ls [02:17] what do you see? [02:18] 1: lo: [02:18] 2:eth0: [02:18] huh @ NO-CARRIER [02:19] curious :) and yet it link lights with the laptop.. [02:19] I missed your original question but with a up state and no carrier on eth0 it means one of a few small issues, cable unplugged, module not loaded... actually the nic shouldn't appear without the module so not that, broken module or broken hardware [02:20] Nicolas_Leonidas: sorry, what is the problem you're having? [02:20] jetole: I posted with more details here http://askubuntu.com/questions/261332/no-network-on-ubuntu-server-after-changing-the-modem Alysum it just hangs [02:21] btw, I know re-posting isn't the best etiquette but no reply on my original question yet. what can cause a D state on a process where it's the fork() syscall ? [02:21] Nicolas_Leonidas: let me look [02:21] I basically got a new modem today, and ubuntu server doesn't connect to the network, cables are all fine, the router's linklight turns off for the ubuntu server on boot [02:22] have you tried with another computer? The model connects via ethernet. Right? [02:22] jetole: tried the same cable with a laptop it worked [02:22] it does connect via ethernet right [02:23] and the computer connects fine to other systems that aren't the modem? [02:23] you mean like one computer to another computer on LAN? [02:24] yeah [02:24] well... stop [02:24] you're connecting eth0 directly to the modem. right? [02:24] jetole: I am right [02:24] ok. so eth0 works fine with other devices? [02:24] connected to a switch or cross connected to another system, etc? [02:25] jetole: correct [02:25] when I cross connected it worked fine [02:25] connect the system to the modem via Ethernet, unplug the power from the modem, count from 1 mississippi - 10 mississippi and when the modem is done booting run: ip link ls dev eth0 [02:25] and look for no-carrier [02:26] oh... after 10, plug the modem back in. I forgot that part but should be obvious [02:27] ok so I will be disconnected from here, see you guys in a bit [02:28] Nicolas_Leonidas: [02:28] if you don't come back I'll assume everything worked well :-) [02:28] good luck [02:28] By the way one time an ISP gave me a modem with 4 ports, that had a hard coded limit of one [02:28] heh [02:29] Nicolas_Leonidas: could have had the option to have more enabled via the ISP's end [02:29] customers who buy extra IP's, may have been limited to business class === jetole is now known as ^jetole === NomadJim_ is now known as NomadJim [05:20] psivaa: GAH. i've wasted a lot of time tonight forgetting that the lxc.run file points utah to a lp: even though i thought i was running from local :) [05:20] psivaa: https://code.launchpad.net/~serge-hallyn/ubuntu-test-cases/server-lxc2/+merge/150491 should hopefully fix it all [06:04] Hello [06:04] Please view this : [06:04] http://paste.ubuntu.com/5566834/ [06:04] Why user can not connect to OpenVPN server ? [06:30] hello any best monitering server ???? [06:55] what? [07:08] Sup guys, how do I get Quantal to see OpenStack charms for Quantal? [07:09] I see some in the ~openstack-ubuntu-testing area but I can't seem to find documention on the 'openstack-origin' option [07:10] For instance, keystone is not in the main charm store repo for Quantal [07:16] got it [07:16] I used cs:~openstack-ubuntu-testing/keystone [07:17] Is there documentation for this anywhere? As far as where the official quantal openstack charms are? [07:17] Just want to make sure I'm not using some random outdated set of charms, etc [07:33] Hmm, seems rabbitmq-server from that repo for quantal actually just throws me a pleasant assertionerror with no string attached [07:39] Daviey, zul: review appreciated - horizon is broken in raring right now - https://code.launchpad.net/~james-page/horizon/g3-compress/+merge/150502 [07:40] trevorj, you can just use the charms for precise with quantal [07:40] trevorj, OR you can use the openstack-origin to select the folsom cloud archive and deploy openstack folsom on 12.04 [07:40] trevorj, there are some docs somewhere [07:42] trevorj, https://help.ubuntu.com/community/UbuntuCloudInfrastructure [07:45] jamespage: Yeah, I've seen those docs [07:45] jamespage: I meant ones specifying how to say, get it working on quantal [07:46] jamespage: cs:quantal/keystone does not exist [07:46] jamespage: same with mysql and others [07:46] trevorj, I'd branch the precise charms locally and the deploy them to quantal [07:47] trevorj, but you are quite right - that does need fixing up in the charm store as well [07:47] * jamespage makes a note to sort that out [07:47] jamespage: ty, I figured the precise charms wouldn't run on quantal, but that's fantastic to hear [07:48] trevorj, they have to because of the cloud-archive [07:48] which allows Folsom (and soon Grizzly) to be deployed in 12.04 [07:48] Wonderful [07:48] I would just use precise but the screen output on my blades is all black (radeondrmfb) [07:50] I can fix it manually with nomodeset or an updated kernel, but that's a pain for many servers. I'm just now getting into Juju and MaaS, so additional configuraton is one hurdle I need to handle after I get this dev stack up and running. [07:52] trevorj, please raise a bug for that - I've heard a few murmurings about the server boot experience and I feel it needs some attention - feel free to subscribe me to the bug (james-page in LP) [07:53] jamespage: will do; I'll put it in tomorrow. [07:53] Tis 3am here ;) [07:54] Something else to note, I had to disable avahi-daemon in the generic preseed for MaaS because my DNS domain is 'lan0.beast2.locsol.local', ie it ends in .local [07:55] I see a few bugs have been opened for this already, I'm just not sure if there's a better way I'm missing? [07:57] The problem is all nodes but the initial juju zk node cannot resolve the initial zk node's fqdn [07:58] I know I can remove [NOTFOUND=return] in nsswitch.conf, or use say 'files dns mdns4' in place, but as I said I still need to figure out a good way of doing additional configuration per node with MaaS+Juju [07:58] trevorj, hmm [07:59] trevorj, lemme think about that - I'll post something back in channel here (have to duck out for a bit) [07:59] jamespage: no prob, I'll check back. Thanks! [08:01] I was thinking of patching avahi-daemon to have a blacklist-domain option [08:01] where I specify a root domain (in my case, locsol.local) and any subdomains of that domain will not be handled by avahi [08:02] (subdomains and that domain I mean) [08:02] That way standard .local mDNS still works [08:08] can anyone suggest an easy-to-setup-and-deploy DNS server solution, one which I can configure via a GUI or web interface? [08:08] TheLordOfTime: powerdns [08:08] trevorj, part of the repos? [08:09] TheLordOfTime: no, http://www.powerdns.com/content/downloads.html [08:09] TheLordOfTime: it really just uses a mysql backend [08:09] TheLordOfTime: really nice for more 'dynamic' dns installs [08:10] TheLordOfTime: I haven't used it for years, but it had a nice web interface ~5 years ago [08:10] ideally, ease-of-configuration is what i really need... [08:10] whether CLI or gui/web based [08:10] TheLordOfTime: so, not bind? ;) [08:11] :P [08:11] TheLordOfTime: if it's a real small install, I use dnsmasq [08:12] TheLordOfTime: just depends if you just need to host a few A records or need something more elaborate [08:12] jamespage: looking [08:12] multi-zone A, CNAME, etc. [08:12] trevorj, ^ [08:12] then check out powerdns ;) [08:12] they've got a deb on the download page [08:27] Wow, folsom is much nicer than essex. Wow. [08:29] I still have to catch up on using quantum/cinder, lots of work ahead === smb` is now known as smb [08:34] trevorj, quantum is supported in the charms from folsom onwards [08:34] trevorj, quantum-gateway is the extra bit [08:58] jamespage: awesome, ty [08:59] jamespage: lots to learn, last I played with openstack was when cinder and quantum were broken git repos heavy in development [09:02] This is also my first ceph install, I hope to get all of this playing happily with each other eventually [09:02] first worthwhile ceph install that is [09:15] hi guys when i boot my Ubuntu Server 12.04 it freeze and stock on the midlle of both becasue of soem NFS share error, any workaround to skip this.? [09:18] <_KaszpiR_> boot into single user mode [09:25] _KaszpiR_: how..? [09:26] <_KaszpiR_> whe your system reboots, you got for a brief moment a grub menu, to choose kernels [09:26] <_KaszpiR_> etc [09:26] <_KaszpiR_> adn there is afair single user mode/recovery [09:27] <_KaszpiR_> that needs physical access to server, or any other solution like ILO [09:27] _KaszpiR_ do i need to repss escapre or press e..? [09:27] <_KaszpiR_> it should show either way [09:28] <_KaszpiR_> by default menu is there for 5 seconds [09:28] <_KaszpiR_> pressing any arrow stops the timer [09:30] <_KaszpiR_> your screen can get blank during restart and it may be just too slow to show that menu [09:38] hallyn: thanks for that [09:49] _KaszpiR_ still there..? [09:50] <_KaszpiR_> no [09:50] <_KaszpiR_> you're talking with the robot [09:51] _KaszpiR_: i see recovery mode , should i select this..? [09:51] Ubuntu, with linux 2.6.32-40-server ( recovery mode) [09:52] ruben231: yes [09:52] still the same freezed during boot when i select recovery mode [09:53] trevorj: still freezed during boot up with recovery mode [09:53] What's the exist output [09:53] exact* [09:54] the NFS share error during boot and its mamped storage.. [09:54] Are you booting from NFS? [09:54] I doubt you are, ignore that [09:55] Can you take a picture? [09:55] wait taking picture now [10:06] ruben231: When it's locked up [10:06] ruben231: Does your caps lock key work? [10:06] ruben231: If you type text, does it appear on the screen? [10:07] ruben231: If it does not, try adding 'nomodeset' to the boot args in grub by pressing 'e' in grub on a boot option, go to the line that starts with 'linux', and add ' nomodeset' without quotes to the end of that line === shanemeyers_ is now known as shanemeyers [10:16] <_KaszpiR_> I'd rhater try boot from another device (usb/cd/dvd) and chroot to system and check out what's going on [10:19] !nomodeset [10:19] A common kernel (boot)parameter is nomodeset, which is needed for some graphic cards that otherwise boot into a black screen or show corrupted splash screen. See http://ubuntuforums.org/showthread.php?t=1613132 on how to use this parameter === Graet is now known as Guest96986 === Guest96986 is now known as Graet [10:48] Why is "mailutils" package depending on Postfix? I just want /bin/mail to forward outgoing mails to the real mail server and not install postfix locally. [10:48] any ideas ? [10:50] trapniii: it's not, it's depending on any MTA [10:57] trapniii: you can use nullmailer for delivery to a smarthost with a queue, or msmtp for immediate blocking delivery to a smarthost without a queue. If you install one of them at the same time as instlaling mailutils, then postfix won't be required as they both provide "mail-transport-agent" as maxb explained [10:59] Although I've had problems with nullmailer in the past (possibly fixed in current versions, not sure). I wonder if a minimal postfix isn't the better option. [11:05] rbasak: Have you looked at http://www.monkey-project.com/ ? [11:05] * rbasak looks [11:06] rbasak: i wondered if it had hyperscale value over apache2 or nginx [11:06] Interesting [11:06] I'm wondering what it gives over nginx [11:06] It says it's optimised for Linux, which I suppose is something [11:07] It's certainly interesting, but I don't feel that their website gives me anything compelling. I don't know of anyone feeling nginx pain, so not sure what they resolve. [11:08] rbasak: yeah, that sound accurate [11:08] I do appreciate the existence of multiple options though. I dislike monocultures. So good for them, and I hope they are around if/when we do feel nginx pain [11:10] Then again, even a small performance improvement would make a big difference at hyperscale. But I don't think we're at a point where that would make a difference quite yet. [11:11] It might be interesting for some of our hyperscale performance guys [11:12] dannf, mahmoh, cmagina: ^^ [11:41] instead of adding myself to a group, I've removed myself from all the groups, now I'm only in 1 group, I don't have any other user (apart root), I don't know root password (apparently it's an auto-generated hash), is there a way to sort this out? [11:43] i am trying to configure a PXE server so i can install two other servers with that [11:43] but i am having a PXE-E11: ARP timeout [11:43] <_KaszpiR_> bilou23 haha [11:43] and PXE-E32: TFTP cannot open connection [11:43] <_KaszpiR_> bilou23 try sudo su - [11:43] <_KaszpiR_> but i doubh it's gonna work [11:44] the PXE server is giving IP to the other servers and i was going through this tutorial https://help.ubuntu.com/community/PXEInstallMultiDistro#Getting_Started to configure it [11:45] <_KaszpiR_> I suggest rebooting machine, boot from the removable media, mount the system/chroot, add yourself to sudoers [11:50] guys, how to decrease Ubuntu Server boot time? [11:54] _KaszpiR_: it says I'm not in the sudoers file [11:55] <_KaszpiR_> boot in recovery or boot from live distro [11:55] <_KaszpiR_> the only option [11:58] _KaszpiR_: ok thx [12:20] anyone knows about PXE server? [12:40] !pxe [12:45] !ipsec [12:47] subnet 192.168.5.0 netmask 255.255.255.0 { range 192.0.0.10 192.0.0.20 } <--- is this possible to implement in ubuntu? meaning i had different range of ip's from its subnet [12:52] mae_tae: Your question needs more context [12:56] maxb: subnet 192.168.5.0 netmask 255.255.255.0 { range 192.0.0.10 192.0.0.20 } <--- i know that this doesnt work because of the subnet 192.168.5.0 but is this possible to implement in ubuntu? [12:57] mae_tae: You've given us a fragment of some sort of configuration file, but not explained what kind of thing you're trying to configure [12:57] mae_tae: I believe so. But not that it's not really a question of Ubuntu. All Linux-based systems will have the same capabilities here [12:57] I'm assuming you're talking about a DHCP server here [12:58] rbasak: yes you're right, i really wanted to give a range of ip different from its network [12:58] if that is possible in linux [12:59] I'm sure Linux can do it if it's possible. But I wonder how DHCP renewals would work. It might violate the DHCP protocol itself. [13:01] If the machines are able to broadcast DHCP discovery packets and have the server hear and respond to them, why not just keep things simple and give the DHCP server an IP in the appropriate subnet? [13:01] subnet 192.168.5.0 netmask 255.255.255.0 { range 192.168.5.10 192.168.5.20 } <---- i know this works because those ranges belong to 192.168.5.0 network, but would it possible to assign range of ip's different from its network, if not how can i implement like that, because in pfsense this is really possible [13:18] if there is request for address which doesn't have subnet defined in ISC DHCPD server I believe server wont offer anything [13:20] e.g. there is client that wants to renew or extend lease for address 192.0.0.11, and there is not subnet defined for 192.0.0.0/24, DHCP server wont offer anything [13:22] DHCP server is authoritative only for subnets you have explicitly defined in DHCPD server's config [13:30] ugh. https://bugs.launchpad.net/nova/+bug/1125378 [13:30] Launchpad bug 1125378 in nova/folsom "VNC proxy can be made to connect to wrong VM" [High,In progress] [13:46] jamespage: *sigh* ill fix quantum and cinder [13:53] zul, ftbfs again? [13:53] jamespage: new binary [14:19] jamespage/yolanda: https://code.launchpad.net/~zulcss/cinder/cinder-backup-service/+merge/150568 [14:21] zul, looking at it [14:21] they created a new service? [14:21] yolanda: yep [14:23] why are tests like that? PYTHONPATH=$(CURDIR) bash run_tests.sh -N [14:26] Daviey, rbasak thx: will keep an eye on monkey; esp. once they've some benchmarks to share though for hyperscale, low mem is becoming less of a constraint. i wonder if optimized for linux == uses epoll? [14:27] zul, why are test run like that? PYTHONPATH=$(CURDIR) bash run_tests.sh -N [14:28] yolanda: because cinder requires it to be installed in order for the tests to be run...i didnt get a chance to fix it properly for g3 [14:29] dannf: pass :) [14:30] Good morning people. [14:31] but not good morning to the bots...lol [14:32] dannf, Daviey: $ grep -r epoll .|wc -l [14:32] 180 [14:32] I'll take that as a yes :) [14:33] rbasak: *nod* - then might be interesting to see if apache's event mpm gives the same improvements [14:35] jamespage/yolanda: last one i swear https://code.launchpad.net/~zulcss/quantum/quantum-config-refresh/+merge/150570 [14:44] good morning people === wedgwood_away is now known as wedgwood [15:00] zul: d'oh, think i finally figured out what my precise->raring upgrade problem was. will keep you updated if im' right :) [15:00] how? [15:01] i had dropped a breaks/replaces from qemu-system-common -> qemu-common [15:02] but it showed itself in very weird ways [15:02] i dunno, we'll see if it re-adding it fixes it [15:21] zul, what are we going todo about versioning in the lab? quite a few projects are being rejected from the PPA's now due to <= current version checking [15:22] jamespage: im not sure add an epoch? [15:22] jamespage: have you had any ideas? [15:22] zul, thats not a solution [15:22] zul, how does the version numbering bit work now? [15:23] 2013.1.a11.g59514bf+git201302260501~precise-0ubuntu1 [15:23] 2013.1.a11.g59514bf comes from upstream right? [15:23] yeah [15:23] and what does .a11 mean? I see the last is the commit [15:23] err, derived from the git commit - upstream hasn;t standardised on that [15:24] jamespage: im not sure about the .al1 [15:24] er.. .a11 [15:25] jamespage: i know why its changing but the a11 means nothing to me [15:26] zul, why does it change? [15:26] jamespage: because the projects are standardizing on oslo changes [15:27] jamespage: if i do a `git describe` in the tree for quantum i get the following: 2013.1.g3-38-g511ac76 [15:30] jamespage: so when we run python setup.py sdist the version is grabbed from the git it [15:30] jamespage: however if we set the OSLO_VERSION before running python setup.py sdist we should be able to get a sane version number [15:32] zul, OSLO_PACKAGE_VERSION might work [15:32] jamespage: checking [15:32] it overrides everything else in setup.get_version [15:35] jamespage: yeah that works [15:35] zul, I think that way we can move back to the 2013.1+gitXXXXXX approach [15:35] jamespage: looks good to me [15:36] zul, lemme take a look at o-u-t [15:36] jamespage: do you want to add it to the or did...nm [15:41] zul, or we could just move to using git archive to create the tarballs.... [15:41] adam_g, any opinions on the above [15:41] ? [15:42] jamespage: i think it would be easier just to export OSLO_PACKAGE_VERSION wouldnt it? [15:42] zul, well it would if the sdist was not run a schroot [15:42] jamespage: good point [15:42] zul, what differences do you get between sdist and git archive? [15:43] jamespage: the shouldnt be any differences lemme check [15:44] zul, egg stuff, AUTHORS Changelog [15:45] nothing that looks like a problem [15:46] zul, if the upstream version number includes the previous commit reference this would all go away [15:46] 2013.1.a38.g511ac76 [15:46] -> 2013.1.g3.a38.g511ac76 [15:46] as the a38 is the # of revs since last tag [15:47] agreed lemme poke around [15:47] rbasak: thx, looks interesting + like the name [15:48] zul, OK _ lemme know - we can always switch to git archive if need be [15:52] rbasak: I might even test it out this week and let you know how it performs [15:53] mahmoh: cool! [15:54] jamespage: so if use git archvie you dont get the egg info [15:57] zul, which means? [15:57] which means its not complete :( [15:58] jamespage: however when you run python setup.py sdist it does do a git describe --exact-match but replaces the '-' with '.' [15:59] zul, and drops a few bits [15:59] right [16:01] zul, might wanna check those syncs as well [16:02] jamespage: for warlock? [16:02] zul, and deps - some ftbfs [16:02] jamespage: ack [16:03] jamespage: lovely [16:08] zul, thats a bug in openstack-pkg-tools [16:08] jamespage: can you pastebin the ftbfs [16:09] zul, https://launchpad.net/ubuntu/+source/python-json-pointer/0.6-1/+build/4327765 [16:09] https://launchpad.net/ubuntu/+source/python-json-patch/0.12-1/+build/4327764 [16:10] jamespage: *sigh* there is absolutely no point to have that as a dependency [16:10] zul, agreed :-) [16:12] so we should have upload to ubuntu1 yes? [16:13] zul, yes [16:20] jamespage: do you want me to fix it? :) [16:20] zul, I can do it right now [16:21] jamespage: k ill just finish off this bug report [16:22] i'm new to ubuntu, if i install syslog-ng with apt-get will it replace the current syslog stuff? [16:23] zul, done [16:24] cool thanks [16:24] ill do the MIR this afternoon [16:25] zul, lemme talk to zigo first - might be able to persuade him to drop that (if I offer todo the work) [16:25] ack [16:27] jamespage: https://bugs.launchpad.net/oslo/+bug/1133433 [16:27] Launchpad bug 1133433 in oslo "Running python setup.py sdist for quantum creates the wrong version." [Undecided,New] === mdeslaur is now known as random_nickname === random_nickname is now known as mdeslaur [16:56] since i am new to ubuntu, i just laid out 12.04 LTS. is there any reason not to upgrade it to 12.10, is there a big difference between them? [16:57] thegoat: 5 years of support for 12.04.. 18 months for 12.10.. thats the biggest reason not to [16:58] holstein: is there a way to tell it up grade just the packages for the 12.04 train, when i did the apt-get upgrade it sends it to 12.10 [16:59] thegoat: try again.. i didnt get that [17:00] so when i did an apt-get upgrade, it upgraded it to 12.10, is there a way to update the packages leaving the release at 12.04 [17:00] oh. you just want to upgrade 12.04 and not go to 12.10... 'sudo apt-get update' 'sudo apt-get dist-upgrade' [17:00] yeah ;-) [17:00] thegoat: if you upgraded to 12.10, then you are on 12.10 [17:00] my bad. [17:01] i am just experimenting with it right now. looking to switch distributions, centos has been making me mad lately [17:02] zul, that fix is uploaded to deb exp as well - I'll sync over when LP notices [17:05] holstein: thanks for the info [17:06] thegoat: there are several reasons not to like rhel-based distros, but there are some reasons to like it as well ;) [17:08] thegoat: but you must have done a do-release-upgrade to upgrade 12.04 to 12.10, and even edited /etc/update-manager/release-upgrades and changed Prompt from lts to normal before doing so... [17:09] all i did was apt-get upgrade [17:09] that won't upgrade to 12.10 [17:09] apt-get (dist-)upgrade won't update the distro, just the packages for your current distro [17:09] i did it as root, could that have caused it? [17:10] no [17:10] do-release-upgrade updates the distro [17:10] but with LTS releases, only to next LTS release unless the file I mentioned is changed [17:10] Did you modify your sources.list file with the new release name? [17:10] so I guess you installed 12.10 in the first place :) [17:11] Pici: heh - that's the debian hack ;) [17:11] could have sworn it was 12.04 rolling out another vm now to verify [17:11] RoyK: Yeah, seen it way too often in Ubuntu though. :/ [17:12] hallyn: did you send to ubuntu-server froma different email address? [17:12] * RoyK confesses he may have used that method some years back as well when he first started using ubuntu [17:12] Daviey: ? [17:13] hallyn: your mail was held for moderation [17:13] Daviey: i used @ubuntu.com. i think my default used to be @canonical.com (not as good for community work) [17:14] hallyn: oh, ok.. approved [17:14] Daviey: thanks. [17:20] oh !@#$ i did dl 12.10..my bad..sorry all [17:20] hehe === andygraybeal is now known as nullbytes === nullbytes is now known as andygraybeal [17:25] stgraber: you probably haven't done much with utah :) but a merge proposal to fix the lxc testcases there was sent your way. [17:25] fwiw, all tests passed locally under utah for me, i just figured you could find anything i was doing stupid lxc-wise [17:26] psivaa: ^ https://code.launchpad.net/~serge-hallyn/ubuntu-test-cases/server-lxc2/+merge/150491 lets lxc testsuite pass under utah for me, 100% fixed here. [17:26] hallyn: ok :) [17:27] thx [17:28] hallyn: ok thank you. hope that gets run in tomorrows set :) [17:58] ok, wish me luck.....getting stupid with fresh installs [17:58] CallingPanther: what up? [17:59] going to go fresh installs onall machines and attenpt to set up some kind of LDAP+NFS situation [18:01] RoyK: ok, how i have a 12.04 install :-D [18:01] trying to make it where no matter which machine I login to I have access to same files and browser bookmarks [18:02] thegoat: goodie [18:02] tired of having to remember which machine has what [18:02] going 12.10 across the board [18:04] this old laptop I am on now will be the only exception since it can only run 8.04 servwer [18:13] CallingPanther: for servers, I'd recommend LTS releases [18:13] 12.04 being the latest [18:15] CallingPanther: since if you install a bunch of machines on 12.10 now, they'll need an upgrade in a year or so, and servers should be left to do their work for years without upgrades needed all the time... [18:16] in "upgrade" I mean "dist-upgrade" [18:16] or "do-release-upgrade", really [18:16] which may have a few impacts [18:18] I'm having an issue with the Ubuntu Server 12.04.2 installation on a Supermicro X9SCi-LN4-B motherboard. I cannot get video/output past Grub - I've tried various VGA settings, etc, to no avail - any suggestions? [18:19] SynSynWork: oh, i didnt notice you had posted this in #ubuntu ..you can repond here [18:20] Awesome. [18:20] SynSynWork: you dont see the text installer? [18:20] Its a 1U server, I could try, but I don't have a card that would work in it on hand. [18:20] Nope. [18:20] http://fridge.ubuntu.com/2013/02/26/ubuntu-developer-summits-now-online-and-every-three-months/ [18:20] SynSynWork: nomodeset is what i would try [18:20] Grub comes up, but nothing after that. [18:20] Tried it. [18:20] Didn't work either. [18:21] SynSynWork: do you have an installation? [18:21] vga=vga16fb ? [18:21] holstein, No, this is during the installation. [18:22] SynSynWork: you can always take the hard drive to another machine, install, and setup ssh [18:22] RoyK, Just tried it, not working either. [18:22] SynSynWork: if you're grasping at straws, maybe try 12.04.1 -- it's got a different kernel and Xorg stack. (I'm not very optimistic. But it _is_ different..) [18:22] vga16fb.modeset=0 [18:22] try that [18:22] sarnold, Decent idea, I'll start that download now. [18:22] RoyK, standby [18:24] RoyK, no go. [18:25] I used that with some hyper-v setups that were dead slow with fb [18:25] stgraber: lxc.postinst, you restart dnsmasq so long as /etc/dnsmasq.d-available/lxc exists, should that only be done if the symlink gets created into /etc/dnsmasq.d? [18:26] low prio, but... [18:26] hallyn: no, we also want to restart dnsmasq if we change /etc/dnsmasq.d-available [18:26] hallyn: it's certainly useless if the symlink doesn't exist, but we shouldn't only do it when we first create the symlink [18:26] Man this sucks. Pulling the drive is an option, but the drives are on a raid card - I'd have to set that up on a working system. [18:27] stgraber: ok, there's no downside to restarting right? [18:27] SynSynWork: what raid card? [18:27] hallyn: right [18:27] zul, just figured out the cinder scheduler problem with jgriffith [18:28] MFI MegaRAID - Don't know the exact model [18:28] jamespage: oh? [18:28] stgraber: kthx (pushing analogous fix for libvirt) [18:28] SynSynWork: I just use software raid ;) [18:28] zul, yeah - it seems that if any of the requires for cinder are missing/broken, the stevedore extension manager can't load the endpoints for the filters/weights correctly [18:28] easier to handle if something goes bad [18:28] zul, python-babel was required [18:28] RoyK, I hear you. This is a custom 1U server though, so I don't have much say ;-) [18:29] s/if/when/ [18:29] jamespage: o...k [18:29] RoyK, Plus, I'm not even to the disk section yet [18:29] ok first machine is installed now for next one [18:29] zul, plus rtslib has no PKG-INFO or egg-info so pkg_resources can't tell which version it is! [18:29] jamespage: awesome [18:29] SynSynWork: megaraid sounds like LSI [18:30] zul, the package is based on upstream git not released tarballs so is a bit broken [18:31] jamespage: heh.. [18:31] RoyK, Yeah, typo/brainfsck - LFI MegaRAID [18:31] jamespage: wanna me have a look at it? [18:31] zul, yes please - https://pypi.python.org/pypi/rtslib/ should be the canonical source [18:31] SynSynWork: yes, LSI ;) [18:31] there is a new version - maybe pick that up and make the switch over. [18:31] * SynSynWork bangs head on keyboardmore [18:32] RoyK, yeah =/ lol [18:32] RoyK, Alright, according to MB docs, the board has a Matrox® G200eW for integrated video. [18:33] SynSynWork: https://help.ubuntu.com/8.04/installation-guide/sparc/boot-parms.html [18:33] look for framebuffer in there [18:33] 8.04 is a bit old, but it might work [18:34] zul, I added detail to bug 1126378 [18:34] Launchpad bug 1126378 in rtslib "SchedulerHostFilterNotFound: Scheduler Host Filter AvailabilityZoneFilter, CapacityFilter, CapabilitiesFilter could not be found." [Undecided,New] https://launchpad.net/bugs/1126378 [18:34] zul, we should aim to fix this for what we have in g3 [18:34] jamespage: agreed [18:35] zul, OK _ I'll leave it in you hands! [18:35] ttfn [18:35] toodles [18:35] RoyK, Yeah, was just over talking to an co-worker about going back to 9.04 and upgrading up. [18:35] I'd prefer not doing that though. [18:36] SynSynWork: if so, start with 8.04 [18:36] Alright. [18:36] direct upgrade path to 10.04 and then to 12.04 [18:36] Let me see if I can get a 08.04 installer to load, I should have a CD around here... [18:37] or try 10.04 first [18:38] i am runnin 8.04 server on this laptop [18:38] * RoyK has a vm on hardy running zimbra [18:38] too bad support for that goes out in a month or so [18:39] my laptop is so old it won't runn any thin newer and can't runwith gui [18:39] Ha. [18:40] so CLI only [18:40] Yeah, my laptops only 11.10 but I waited a long time to upgrade that. [18:40] 40Gb hdd and 256Mb Ram [18:40] I have a 333MHz running Gentoo, still. [18:41] I can't remember how much memory it has, but it took weeks to compile everything. Those were the times. [18:41] i forgot what this cpu is.... think like 266 [18:41] Nice [18:42] goodie ;) [18:42] old toshiba was given to me with winblows 2k [18:42] a raspberry pi will kick that out [18:42] Yeah, it's crazy. My phone is quad-core (1GHz), server is 333MHz. [18:42] quickly found new use for it...lol and new OS.....windows free home [18:43] * RoyK has a quad-core 1,4GHz in his phone [18:43] I have one Windblows box for testing, everything else is *nix [18:44] I ran a webserver for a bit but could only handle 2 connections or it crashed [18:44] wintendo is nice for certain things [18:44] I do miss games but hey thats the price you pay for security [18:45] OpenArena will do in a pinch...lol [18:45] RoyK, I have the Nexus 4, so it's actually 1.5GHz Quad. [18:46] * RoyK only has the S3 [18:46] I'll admit I play games on the windblown box from time to time, but with Steam's official port to Linux, no more fscking with Wine…I'm happy. [18:47] my internet is not good enough for Steam [18:47] lol [18:47] out in tha boonies [18:48] Then you don't want to hear about my 105Mbps connection. [18:48] * SynSynWork laughs [18:48] no offence..... but bite me [18:48] lmfao [18:48] I have a friend who lives in the sticks, he had dial up until only a few years ago. [18:49] my mother has 25mbps she is in town... only 45 minutes away [18:49] His only other option the entire time was Sat. Sat for down, and still dial-up for up. Or ISDN (which just isn't worth the price). [18:50] I remember ISDN days..... [18:50] then T1s and T3s [18:51] Thought we were the bees knees back then [18:51] lol [18:51] now everyone and their mother is faster than me [18:51] DSL does suck. [18:51] All around, its just bad news. [18:52] on DSL but speed test shows .08/.05 [18:52] Yup. [18:52] almost dialup [18:52] Damn, do you have the 3.0 or 1.5M plan? [18:53] I _paid_ half of that (~$30) for 25M. Now I pay $70 for 105M [18:53] 1.5 not paying 3 times moe when they can't even give me what I am paying for [18:54] Do you know how far you are from the CO? [18:54] ummm by their reaction.....lightyears.....lol [18:55] lol [18:55] Further you are, the worse service/signal you'll get. [18:55] I thihnk it is in town 3 miles away [18:56] my nid to their box is 200yards [18:56] 3 miles is generally considered the upper limit for acceptable service. [18:56] Sounds like your on the edge there [18:57] yeah I know [18:57] Which wold explain part of the problme [18:57] I used to do tech support for a DSL company [18:58] does not help that I am shooting wireless 100 ft from where the router is [18:58] I have a jacked up setup [18:59] phone to modem to wireless PC that passes it off to another router in here for the other PCs [19:01] and here I was thinking of adding another router to the mix....lol [19:02] for a webserver and email server [19:03] when you have machines with no purpose you think .....what else can i do here [19:04] RoyK, All of our CDs were desktop (and i386) so downloading now... [19:06] will return when all OSs are installed and updated......cya then [19:09] using 32bit installs for a server isn't exactly recommended [19:09] (nor are desktop installs) [19:10] RoyK, Got it working… vga=711 nomodeset (which I tried earlier). The difference this time? I switched out the monitor, as a last ditch effort. [19:13] Yeah those desktop CDs were for other various projects. [19:13] SynSynWork: working with 12.04? [19:14] 12.04.2 [19:14] ok [19:14] RoyK, Same option I had tried earlier. Just swapped the monitor out. [19:14] So clearly it was an issue with the monitor - go figure. [19:14] seems monitor handshake in fb is somewhat dodgy [19:14] I've seen that earlier [19:15] there should be a "use text mode!" setting [19:15] No kidding [19:15] especially if you're through a kvm switch with oldish monitors [19:16] ubuntu can't read from the monitor what to do [19:16] Was a standard new-ish Dell monitor, nothing fancy or special. [19:16] I'm not using a cheap, old ass 13incher [19:16] imho using framebuffer on a server doesn't make sense [19:17] what would you need it for? [19:17] Agreed [19:17] Serial Console or GTFO [19:17] can you file I bug? [19:18] "enable text mode by default, please" [19:18] it shouldn't be very hard [19:19] jamespage: sigh the python-rtslib debian package is not shipping any egginfo [19:19] zul, indeed [19:20] jamespage: i have it fixed locally but its minor surgery [19:20] zul, hmm - why not just switch the upstream source and fix it properly [19:20] jamespage: because that is using the upstream source (ie from pypi) [19:23] zul, hmm - I see [19:25] zul, debian/rules is doing to much IMHO [19:26] jamespage: agreed [19:26] jamespage: ive simplified it locally [19:38] jamespage: babel was in pydist-overrides for cinder as well fyi [19:39] jamespage: http://paste.ubuntu.com/5568534/ [19:52] anyone know a channel where people might be able to answer something about webmaster tools for different search engines [19:52] #web ? [19:53] k I'll try it, thanks [20:37] jamespage: can you try https://launchpad.net/~zulcss/+archive/rtslib/+packages please === main is now known as threesome === main is now known as threesome === main is now known as threesome [21:08] zul, so babel should be in the deps right? [21:09] jamespage: yeah its not explicitly there though [21:10] zul, " * debian/control: Drop python-babel." in quantal? [21:11] jamespage: right in quantal but im looking in tools/pip-requres its not there anymore [21:12] zul, tis - Babel>=0.9.6 [21:12] jamespage: *sigh* [21:13] zul, maybe it should not be? [21:13] no i think it is but its for internantilizing string [21:14] translating strings [21:16] zul, translations etc. [21:17] jamespage: right [22:17] Is this an appropriate place to ask a question about SMTP servers in Ubuntu? [22:24] br0xen: sure, whats the question [22:27] Ok, and please excuse my ignorance, I just haven't done much with email servers... I built a website for a company that was previously relaying email through their exchange server, everything worked fine that way, but then they upped their security and are now demanding that I run my own email server using a subdomain of their main domain as the 'from' address... [22:28] So, previously we would send email from something like "payments@domain.com", now they say they want me to switch it to "payments@my.domain.com", and they claim that I should be able to authenticate everything correctly... My thought is that since I'm not relaying through the 'official' domain.com, this email will be caught in more spam filters and stuff. [22:30] Just trying to find anyone that might have experience with something like this... If not, I'll keep searching. :-) [22:33] br0xen: I don't think you need to worry about that, relaying though the "official" domain doesn't make it less likely to get blocked by spam :) [22:33] filters [22:33] Heh, ok. [22:35] So, I suppose my next question would be, what route is recommended for setting up an email server, piggy-backed on a web server? It would only need to send email (in fact, preventing incoming email would be a plus, that way we wouldn't need to worry about mailboxes filling up.) [22:36] I think people recommend Postfix https://help.ubuntu.com/community/Postfix [22:36] I've looked into postfix and exim a little bit. [22:36] Yeah, that's what I was thinking. [22:36] I tend to do everything with postfix, but then, I'm heavily biased. [22:37] * blkperl is stuck with sendmail :( [22:37] lamont: me too [22:37] I cannot stand exim's config syntax [22:37] Reminds me of m4 macros [22:37] blkperl: I remember when I accidentally switched a production mailing list from sendmail to postfix... 2 weeks later someone noticed. [22:37] sendmail? in the year 2000? wow. [22:38] the "melissa virus" was my postfix win at HP, back when [22:38] configs were written in the 90s not going to touch them :) [22:38] ah :) [22:38] jamespage: I filed that bug report, lp:1133683 about radeondrmfb booting to a black screen on my blades on precise kernel [22:38] jamespage: Added you to it, not sure if I selected the right package though I must say, I just entered the full linux-image including version [22:43] postfix is fantastic with multiple instances for large chunks of mail [22:49] Hola. I'm looking for some advice on setting up my web server - I've ran into an odd issue. Is this the righ place? [22:49] yup [22:49] :) [22:50] What's your question, sir? [22:52] So, I've created a virtual host for a site, installed the software (Lamp stack) on Ubuntu 12.04 LTS, moved site files over, and when I navigate to to the domain.com/install.php page it's blank. However, a file with the SAME ownership, grouping, and permissions containing a simple works fine… Any thoughts? [22:52] joel___: you're prolly missing a php module [22:52] I was thinking my vhosts file is bad, but not sure why one file would show up fine and others not… (seems to be some PHP files, as .txt files show up fine) [22:52] Mmm.... [22:52] joel___: enable php startup logging [22:53] joel___: in /etc/php5/apache2/php.ini enable error_log = syslog [22:53] joel___: also enable showing of errors and startup errors [22:53] joel___: restart apache2, then try it [22:54] error_log = syslog is already set. [22:54] joel___: you can also try tail -f /var/log/apache2/error.log or whatever error log you set your vhost to use [22:54] joel___: great [22:54] joel___: install ccze [22:54] joel___: tail -qF /var/log/syslog | ccze -A [22:55] joel___: tail -qF /var/log/syslog /var/log/apache/*.log /whatever/other/vhost/logs/you/got | ccze -A [22:55] joel___: I mean that ^ [22:55] joel___: then go to it, see if you see anything interesting [22:55] lol. Right. Ok. [22:55] ccze? [22:55] sarnold: makes logs readable, colorizes ;) [22:55] ah :) [22:56] sarnold: it's got a memleak though so you can't run it for months (tried that) [22:57] trevorj: hahahaha [22:57] Was great when production servers went down due to my own colorized log viewing [22:57] lol [22:58] So, hitting my domain.com/butt gives me a not found in my error log (awesome), but it doesn't throw any error when tailing the syslog and navigating to a file that IS in existence (domain.com/install.php) [22:59] lol @ CoLoRz [22:59] joel___: kinda nice isn't it? [22:59] Totally [22:59] :) [22:59] joel___: enable showing of startup errors and showing errors [22:59] joel___: that way it'll actually output something to your browser as well [22:59] Mmm. [23:00] joel___: display_errors and display_startup_errors [23:00] joel___: they're both off by default in the config [23:00] Hi there, I have an issue with updating my 12.10 ubuntu server with fetching error, can anyone help? [23:00] Sure enough, startup_errors was Off (displaying I turned on already). I'll restart. [23:01] tyson_: what do you mean fetching error? [23:01] tyson_: like with apt-get ? [23:01] trevorj: yes, sorry [23:01] tyson_: pastebin the error [23:01] tyson_: er, full output I mean [23:01] trevorj: ok [23:03] trovrj: http://imagebin.org/248250 [23:04] tyson_: sounds like your dns isn't working [23:04] tyson_: can you 'dig google.com' [23:04] tyson_: er, dig +short google.com [23:05] tyson_: cat /etc/resolv.conf [23:05] tyson_: do you see any nameserver entries? [23:05] trevorj: I haven't tried as I installed it as a server (black and white) [23:05] tyson_: are they correct? [23:05] trevorj: let me try [23:05] tyson_: hey, it's got colors too ;) [23:06] So, apache2 error.log is showing 'caught SIGTERM, shutting down'. It's crashing, then? I'm a bit confused as text files and a phpinfo(); is running fine, but no other *.php files are working. [23:07] trevorj: lol... i meant, I only have the shell sript and I only access it remotly, anyway, I think it's empty [23:07] tyson_: are you using DHCP or static IP [23:07] # Dynamic resolv.conf(5) file for glibc resolver(3) generated by resolvconf(8) [23:07] # DO NOT EDIT THIS FILE BY HAND -- YOUR CHANGES WILL BE OVERWRITTEN [23:07] tyson_: two lines? [23:07] tyson_: are you specifying network config via /etc/network/interfaces or via network-manager [23:07] trevorj: good question... I was on DHCP but messed about with it the other day to get a static IP, so it is now on static [23:08] tyson_: rm -f /etc/resolv.conf; echo -e 'nameserver 8.8.8.8\nnameserver 8.8.4.4' > /etc/resolv.conf [23:08] tyson_: try that ;) [23:09] trevorj: knowledge is power.... UP and running, thank you... what was the problem? [23:09] tyson_: you were missing DNS servers as you were originally setup DHCP [23:09] tyson_: when you switched to static you didn't specify any nameservers so after you rebooted you lost your generated dns server list [23:10] tyson_: I asked you to rm the existing copy as if you don't it will auto generate it via resolvconf [23:10] tyson_: it's originally a symlink to a tmpfs file [23:11] trevorj: thank you very much [23:11] tyson_: you can also specify them in /etc/network/interfaces but I always forget if the syntax is plural or not so I just do it the old way [23:11] tyson_: anytime, sir [23:11] trevorj: that is the file I edited to give it a static Ip in the first place [23:11] tyson_: yes it is [23:12] tyson_: you can use dns-nameserver or dns-nameservers in there under the iface if you don't remove the symlink as I asked you to do [23:12] tyson_: I just never remember which of those two it is [23:12] tyson_: so I do it this way [23:12] tyson_: tis all preference [23:13] trevorj: thank you for sorting the problem out but most of all for the explanation.... VALUABLE [23:16] joel___: sorry, I missed your message [23:16] NO, you're good. :) [23:16] joel___: thats just from you restarting apache2 [23:17] joel___: it's also possible your app is die()'ing [23:17] joel___: you may be in for some php troubleshooting [23:17] (how do you guys keep addressing people like trevorj__: ? Typing that manually?) [23:17] joel___: tab complete ;) [23:17] trevorj: BAm [23:17] yeah, IRC is awesome. [23:18] trevorj: Yeah, I'm able to drop some PHP in a test file (http://dev2.concrete.ly/test2.php) [23:18] But the install.php is just blank… [23:18] joel___: what app is this? [23:18] ClientExec [23:19] ah, paid app [23:19] does it use some nasty obscurity methods? [23:19] Could you tell me if my permissions are correct? [23:19] Oh… maybe? [23:19] most paid apps do [23:19] thought it was a free download [23:19] they usually use something like, crap I forget the name [23:19] Aaaah. Interesting. [23:20] joel___: ioncube is the one most widely used [23:20] joel___: If you're just looking for helpdesk software and you're just testing one out, I like OTRS [23:21] joel___: in access.log when you go to the install.php link [23:21] trevorj: Figured Ioncube was installed already. Like some sort of 'encryption', but not? Also, I'll check that out. It's more for the billing and server integration aspect [23:21] trevorj: Nothing. doesn't show a thing, as the file is there. It's just blank... [23:21] joel___: can you paste the single line that outputs in access.log? [23:21] joel___: access.log has to output something, it always does upon http connection [23:21] Mmm. One sec (was thinking error.log) [23:22] trevorj: 76.27.233.224 - - [26/Feb/2013:23:22:16 +0000] "GET /install.php HTTP/1.1" 200 293 "-" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_8_2) AppleWebKit/536.26.17 (KHTML, like Gecko) Version/6.0.2 Safari/536.26.17" [23:22] oh wow, it's a 200! [23:22] your script is not dying with an error [23:22] your script is showing you nothing with a success code [23:22] this is an application problem [23:23] joel___: yes, it requires ioncube [23:23] joel___: https://www.clientexec.com/members/index.php?fuse=files&view=FileList it says _ioncube.zip [23:23] trevorj: Awesome! That's really helpful trevor. I assumed I botched something in the vhost setup. I'll pay more attention to access log. [23:23] trevorj: aaah. K. I thought I had that enabled, but I'll double check. [23:24] joel___: make sure you enable it in the right place in /etc/php/, cli and apache2 have separate configs [23:24] trevorj: Mmm [23:24] joel___: ioncube comes with a php file to test it out [23:25] joel___: toss that in your webroot and try going to that [23:25] joel___: also, keep in mind you're running untrusted non open source code ;) [23:26] joel___: it may require a license, they seem to on their site, but who knows [23:26] trevorj: This is making sense. Yeah, they gave me a trial code. So there is some encrypted 'ion cube' stuff i with the php files. [23:27] trevorj: Yup - there it is. Ok. [23:27] joel___: yeah, ioncube instantly makes code undebuggable [23:30] Anyone know where the most up to date cinder charms are for precise-folsom? [23:31] oh weird, it found it this time, nevermind [23:31] trevorj, lp:charms/cinder should be. [23:31] adam_g: yeah, it gave me a 404 the first time, my fault, thanks ;) [23:31] ah [23:32] oh, nevermind, I was in the wrong terminal [23:32] sigh [23:42] How does everyone here handle post-install configuration when using MaaS+Juju? I used to use scripts specified in the preseed, but is there any better way these days? [23:43] I wrote some scripts to handle iLO configuration like setting server name to hostname, setting network in boot order, setting up standard user accounts, things like that [23:44] If there's a nice place to put such things in MaaS I'd be happy to submit it in [23:45] I see commissioning-user-data would be a great place for alot of this, besides setting the server name to hostname [23:46] do you really care about servernames in the maas+juju lifestyle? [23:46] yes, these are iLO server names [23:46] otherwise it's a freaking guess as to which blade is which [23:46] hehe [23:46] I have to look it up by mac address [23:48] there's more post install stuff I want to do as well, and I'm assuming the best place for that would be in the preseed [23:49] I've been using salt for deployments in the past [23:49] trevorj, you can modify the base preseed that is used for all machines. not sure if there is a better way to do it [23:49] adam_g: yeah, I've already done that, it's a tad finicky for my taste though ;) [23:49] trevorj, by post-install stuff, do you mean things that can be done post-first boot? or during initial provisioning? [23:49] during post-first boot [23:49] er after [23:50] Is cloud-init a good place for such things? [23:50] trevorj: you may wish to also ask in #juju or #maas -- both are a bit more dedicated to the tools at hand [23:50] great, ty [23:50] sarnold: I didn't know such channels existed [23:50] sarnold: much appreciated [23:50] trevorj: woot :) [23:50] trevorj, cloud-init would be, but when using juju you have no ability to modify cloud-config that is sent to new nodes [23:50] adam_g: right! [23:51] adam_g: what I'd love to do is setup a salt minion upon first boot [23:52] adam_g: as it seems juju is more meant for services and not deploying configurations [23:52] adam_g: it just sems as if there may be (or should be) a better way to modify the cloud-config that juju sends [23:52] trevorj, is setting that up something that could be contained in a juju charm? i'm not familiar with the salt architecture. [23:52] adam_g: you know, it could. [23:53] adam_g: and that is what I was looking for, the revelation [23:53] trevorj, you may want to look at subordinate charms. you may be able to put together a subordinate charm that is deployed alongside the rest of your serivces, that takes care of hooking it up into your salt stack, and have it take care of the lower-level config bits that the principle service probably doesn't care about [23:54] adam_g: wonderful [23:54] trevorj, theres been discussion about doing similar with subordinates that relate services to existing puppet or chef servers. [23:54] trevorj: a quick question for you... what do I need to access my server from outside my "home" network? [23:55] adam_g: I suppose I should start looking into how juju really works then! [23:55] adam_g: thank you [23:55] tyson_: port forward for ssh [23:55] trevorj, np [23:56] adam_g: since you seem to know some intricate details on things like this, maybe you'll know an answer to this as well [23:56] trevorj: do I have to do that on my router and modem? [23:56] tyson_: router only probably, unless you are doing some nasty double natting [23:56] adam_g: I had to disable avahi-daemon in my preseed because my internal domains are subdomains of locsol.local [23:57] adam_g: ie, it ends in .local [23:57] adam_g: since my dns domain is lan0.beast2.locsol.local, avahi takes it that it owns the entire .local root domain and returns NOTFOUND [23:58] trevorj: nothing dodgy. I have fiber and it seem weird but I have a modem from BT on the wal connected to a router from Talk talk on it... [23:58] adam_g: is there an existing way to get avahi to not manage subdomains of .local? I was thinking of patching avahi-daemon to have a blacklist-domain option for anything below a certain domain [23:58] adam_g: that way .local would still be handled, just not anything below the blacklisted domain (in my case locsol.local)