[00:04] Hi, I'm trying to re-install nagios. I removed all traces of nagios from my server (including /etc/nagios3). However, when re-installing nagios with apt-get, it no longer creates the config files. How can I make it cleanly install nagios? [00:05] savid, you need to apt-get purge the packages [00:06] qman___, yeah, I did that.. probably didn't do things in the right order though :-( [00:07] if you manually removed files, you didn't purge the right package [00:07] install all the packages you need, then find them with dpkg -l, and purge all of them [00:07] then the next time you install them, it'll recreate the configs [00:08] qman___, I guess I'm not sure which are "all the packages I need". I think it's just "nagios3" [00:08] probably nagios-common or similar [00:08] dpkg -l | grep nagios [00:08] will show all currently installed and previously removed packages which contain 'nagios' [00:08] ahh, I see. [00:10] if that still doesn't show your culprit package, check the dependencies on packages.ubuntu.com [00:10] also, if you know which config files in specific are the problem, you can search what packages contain them [00:10] forget the command for that, I think it's apt-file or something [00:12] damn. shouldn't have manually removed the files. purge is failing for some packages. [01:03] I notice that there's no /etc/apt/apt.conf file in 10.04 LTS although I configured for automatic security updates at install. Can I still make one to define Install-Recommends & Install-Suggests? [01:03] NginUS, check /etc/apt/apt.conf.d/ [01:04] qman___: I see some files there, but none whose contents reflect what I expect to see in apt.conf. I'm confused & would love to see some docs but where? [01:04] add one in there [01:05] in the order you think it should be added [01:05] it works like rc.d, combines the files low to high to create the config [01:06] qman___ : So the prefix numbers are what matters, not the filename in /etc/apt/apt.conf.d? [01:06] yes [01:06] arosales, Daviey - we may do a respin to pick up some ubiquity fixes ( https://launchpad.net/bugs/942560, and possibly https://launchpad.net/bugs/940908) do you want to have the server images respun to pick it up? [01:06] Launchpad bug 942560 in ubiquity "keyboard layout screen - Keyboard navigation broken" [High,Triaged] [01:07] not having done it before, I'd create one called 49recommends [01:07] qman___ : Great to know, thanks [01:07] ok [01:07] that may or may not be the optimal spot to put it, but that's my guess [01:08] NCommander, smoser, ^^ any opinion? [01:09] * NCommander reviews [01:10] qman___ : Now that I'm looking at these files in /etc/apt/apt.conf.d with some understanding, it makes more sense of what I'm seeing, thanks again. [01:10] skaet: alternate images don't share the revelent code. Respin shouldn't be necessary [01:10] Thanks NCommander. :) [01:10] Not a problem [01:11] skaet: (we should respin all non-alternates though; thats a pretty delibating bug :-/) [01:12] NCommander, we'll be doing it for the Ubuntu Desktop - I've got a query out to the leads of the flavors if they have the testers able to cope or not - before doing the respins for them. [01:32] zul: https://code.launchpad.net/~gandelman-a/nova/libvirt_refresh_2/+merge/95295 [01:33] zul: we should really try to get that patch upstream, its going to be a nightmare to maintain going forward [01:33] adam_g: i was thinking that this week [01:47] zul: gonna send a merge for keystone, too, with all the stuff ive done this week [01:48] adam_g: ack [02:36] adam_g: you are going to do keystone tonight? [02:40] zul: just pushed lp:~gandelman-a/keystone/ksl-packaging [02:41] zul: that gets a working service, with everything that has an sql backend configured to use it (added to the sql patch thats there) [02:41] adam_g: so if i turn on the regular jenkins job tomorrow morning then we should be ok right? [02:41] zul: no [02:41] adam_g: what still needs to be done? [02:42] zul: the charm isn't 100% yet, once its done (i think its pretty much done), we need to change some of the utility scripts we use in prep'ing the deployment for testing [02:42] zul: and the nova and glance charms need a small update as well [02:42] adam_g: ok [02:42] adam_g: i totally intend to upload the e4 keystone on friday though [02:43] zul: also, need to figure out whats become of the service token stuff im hoping that straightens itself out over night [02:43] zul: ya, we will [02:45] zul: about the packaging. theres stuff in keystone-manage for migrating older keystone databases, im not sure if we want to attempt to use that from postinst on upgrade (assuming it works) [02:45] zul: i probably wont be able to test any of that before friday, so if you want to go for it [02:46] we dont do any user stuff in the postinst so i think we should put it in a readme [02:46] adam_g: i just added your changes [02:46] zul: well, i added the db_sync to postinst/configure like we do in nova [02:47] adam_g: i mean explicitly keystone users like administrator [02:47] zul: i guess the question is are we trying to support an upgrade path [02:47] ill ponder it [02:47] but its getting late for me [02:48] zul: from previous keystone version in the archive.. or, is it assumed that anyone using this stuff is going to be installing for the first time [02:48] yea. same, ttyl [03:04] New bug: #943748 in glance (main) "Glance is missing dep on python-dateutil ( 2012.1~e4~20120224.1290-0ubuntu1)" [Undecided,New] https://launchpad.net/bugs/943748 === Jasonn is now known as juicy [03:29] adam_g: still around? === sixstringsg|away is now known as sixstringsg [05:12] New bug: #943798 in samba (main) "package samba 2:3.5.8~dfsg-1ubuntu2.3 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/943798 [05:23] Curious, does anyone know if you can use a USB Serial adapter or Serial PCI Add-on card as a console port for a server? === sixstringsg is now known as sixstringsg|away [06:42] Curious, does anyone know if you can use a USB Serial adapter or Serial PCI Add-on card as a console port for a server? My on-board serial port just keeps dumping random letters like the settings are wrong but they match 100% the server and client machine. [07:04] gdeeble: in theory it should work... [07:09] I have installed sendmail on my ubuntu server. Im using Webmin GUI for server overview. I got 2 domains for my server. When I create a new user, its email adress is for the domain I dont need, but I cant seem to change it? Ive updated virtuser and aliases in sendmail for the account, but still when i try to compose a email for the account, it uses the wrong domain, how can I change this? [07:16] !webmin | znow [07:16] znow: webmin is no longer supported in Debian and Ubuntu. It is not compatible with the way that Ubuntu packages handle configuration files, and is likely to cause unexpected issues with your system. [07:18] znow: more info via Debian's bot http://paste.debian.net/158145/ [07:21] znow: check out zentyal [07:21] but... sendmail.. wow [07:22] IMO all WHCPs are crap and they all cause more problems than they ever solved [07:22] SpamapS Thank you. Maybe you can give me a little insite then, as It's not working right now. If I do the onboard, It gives me nothing but unlegible text(just random characters) like the client doesn't match the server but I've made sure it does. If I use the USB Serial Adapter, it works to show me the login but that's it. The Serial Card when booting causes the server to hang. It gets [07:22] to the point of saying "ttySx: detected caps 00000700 should be 00000100" x being what ever com port it sets to depending on if my OB is off or on. Any suggestions? [07:22] You should learn the damn CLI [07:23] gdeeble: try a different serial card [07:24] SpamapS: what *would* be interesting is a WHCP-like UI on top of something like puppet... [07:24] twb, It's only if the grub is directing to push console data out it. Otherwise it picks up fine. Still try another card? [07:24] gdeeble: oh, it's grub's fault *again*? Sigh. [07:25] gdeeble: have you tried telling grub to talk to serial, and not to use framebuffer, and not to talk to video console? [07:25] Not necessarily. I'm not exactly sure. I'm still new to all of this. [07:26] twb, do you mean by changing grub_terminal from console to serial? [07:26] TBH I don't know, my response to "grub is at fault" is to replace grub with syslinux [07:27] You'd want to look at /etc/default/grub and fiddle-fart around with that and /etc/update-grub.d/NN-script.sh or whatever it is, and then run update-grub to have that "compile" the grub config file in /boot/grub/grub.cfg [07:29] twb: Puppet is selling "puppet enterprise" for that. ;) [07:29] Right, I have fiddled with the grub file in default, but didn't know about the script. I don't think it's grub but no idea how to tell. I followed the console serial how to, to get this far. [07:29] @twb ^^^ [07:30] SpamapS: haha [07:30] SpamapS: and that landscape whatever stuff from canonical that I'm not interested in paying for [07:30] twb: Landscape is a little different really. [07:30] twb: more about supporting users than configuring services. [07:30] SpamapS: all I know is the server side is closed source so that's automatic fail for me [07:31] SpamapS: supporting users as in desktops? [07:31] twb: or remote servers [07:31] twb: but the point being, its not so much for a novice user with no skills.. its for streamlining an IT workflow [07:31] Ah, I get you [07:32] The right way to solve novice users is something like a rubber mallet and an actuator, and it triggers when you try to install X on the server [07:32] bops the user over the head and says "no! Bad user!" [07:34] haha [07:36] twb: So you're recommendation is if it isn't working replace it? Starting with the loader first? [07:36] gdeeble: I just hate grub; I'm a grub bigot [07:37] There are like twenty separate problems I regularly run into, like it deciding to install grub onto the USB key I'm running the installer off, not the HDD I just installed the OS to [07:37] Whereas extlinux I install and it just works forever [07:38] twb: LOL, see this is my first linux project for home, but have an understanding of some of the command line and want to use the console so I can just plug the laptop in and reboot it and see everything verses dragging a monitor over to see the boot. [07:39] Wait, console as in tty or console as in serial? [07:39] twb: tty over serial, if I'm not mistaking is what I'm trying to do. [07:40] Why don't you just use the laptop's screen? [07:40] https://help.ubuntu.com/community/SerialConsoleHowto <- I followed this. [07:40] twb: You can laugh and say anything you'd like if I'm explaining this wrong. LoL [07:41] serial console is the 9-pin port that you don't get on laptops anymore [07:41] gdeeble: hmm. That document doesn't appear to cover grub2 [07:41] It looks a bit like a D-Sub (VGA) port except it has two rows of pins instead of three. [07:42] twb: Right. I understand. I have a USB to Serial Adapter for my laptop [07:42] rbasak: I followed the Karmic and Newer options. [07:42] gdeeble: which release of ubunu server are you using on your server? [07:42] gdeeble: well, OK, if you realyl want to do that. Seems silly to me, compared to just using the onboard screen. [07:43] rbasak: 10.04 as far as Im aware. [07:43] gdeeble: since the BIOS will not work over serial, and once the OS boots you can just use ssh [07:43] gdeeble: oh now, I'm wrong. The document does cover it and you're using the right sectoin [07:43] If you want to get the dmesg output for debugging prior to SSH being available, you can use netconsole to spew it as UDP to the network. [07:43] twb: What do you mean? Just hooking a monitor up to it? Or am I missing something? Because I want it to be headless, but want to see the start up. [07:44] gdeeble: you said it was a laptop [07:44] gdeeble: oh sorry, I misread you [07:44] twb: I want to use my laptop to view it. [07:44] OK, right. That makes a lot more sense. [07:44] gdeeble: you've definitely got GRUB_TERMINAL and GRUB_SERIAL_COMMAND correct in /etc/default/grub, and you've run update-grub? [07:44] twb: It's fine. I may have miss said it. [07:44] The best way to have serial as the default console is to do the install over serial [07:44] Then d-i will just set everything up correctly out of the box [07:45] In the installer, hit F6 and add console=ttyS0 *after* the "--" argument [07:45] rbasak: Yes. I've enabled them. However, I am confused on the unit parameter, but I went with default like it showed. The main thing is the box hangs and doesn't show anything on the console when I have it going through the serial card. [07:46] If the server also is using a USB adapter you might be able to use console=ttyUSB0 or you might have to set it up post-install [07:46] gdeeble: unit corresponds to the port number I think [07:46] gdeeble: if you're using a serial card, that line may need to be a bit different [07:47] twb: I have tried using the USB adapter on the server and got no errors nor any output but again it could be how I had it set up. [07:47] gdeeble: http://www.gnu.org/software/grub/manual/html_node/serial.html#serial - you may need to use the port switch instead of the unit switch if your serial card is using a non-standard port [07:47] gdeeble: I don't think that grub will work with a USB serial adapter [07:47] gdeeble: if you have done a non-console (i.e. VGA) install, then dmesg output does not go to the serial by default. [07:48] gdeeble: not sure about linux console output either. A login will work on it though [07:48] rbasak: AIUI his USB adapter is not in the server, it's in his laptop [07:48] rbasak: i.e. he has a normal 9-pin port on the back of his server, and is using a laptop and a USB adapter to talk to that [07:48] I see [07:48] ICBW though [07:49] What's the serial card on the server about? [07:49] NFI [07:49] gdeeble: do you have a normal serial port on the server, or is a serial card the only way of having one? And which port is the serial card set to emulate? COM1 or COM3? [07:49] rbasak: What do you mean? Why is it there? [07:49] (or something else?) [07:50] gdeeble: yeah, why is it there? Is it that you don't have a built-in serial port, or something else? [07:50] Rbasak: I have com1(ttyS0) on board which just gives me unlegible data, like I've set the client side wrong(which I've made it mirror), then the computer is puting the serial card as S1 and S2 [07:51] gdeeble: you could try --unit=1 in /etc/default/grub then, which will try to use COM2 (the equivalent of /dev/ttyS1 perhaps) [07:52] "unlegible data" is probably because you had the baud rate wrong or something [07:52] rbasak: To start, I didn't have the pigtail for the port as it only has the header for it. I got the serial card to replace that for the time being and finally got the header, but the header is kicking out just random characters like if the client side is set to different baud or something like that but I made sure it matched. [07:52] twb: I've confirmed both the server and client were matched at the same baud 115200. [07:52] Instead of --unit=1 you could also try turning off the onboard serial port in the BIOS [07:52] I've tried to drop it down as well. [07:53] gdeeble: fair enough [07:53] So what the serial board is doing at the point grub is running is the old DOS 16-bit thing. It'll listen on some fixed hardware port, and fire interrupts on some fixed number [07:53] (for each port) [07:54] twb: Tried that as well. Killed the OB and let it boot that way, which gave me no data and will detect the card and hang from there. [07:54] COM1 corresponds to port 0x3f8. COM2 corresponds to port 0x2f8. I suspect that grub won't use serial interrupts so that won't matter. [07:54] If you specify --unit=0 then grub will look for the port at 0x3f8 [07:55] If you specify --unit=1 then grub will look for the port at 0x2f8 [07:55] Or you can use --port then you can override that [07:55] I posted on a forum about this maybe if you look at the information posted there it can help? -> http://www.linuxforums.org/forum/ubuntu-linux/187114-console-via-pci-card-fails-boot.html [07:55] It isn't clear to me what address your serial board is using for which physical port. [07:56] rbasak: ttyS0(OB) ->03f8 ttyS1(SC1) ->EC00 ttyS2(SC2) ->EC08 === smb` is now known as smb [07:57] gdeeble: you got that from where? The kernel log output? [07:57] The post might be a little off as I was playing with setserial during that time. [07:58] rbasak: I ran -> setserial -ga /dev/ttyS? [07:58] gdeeble: that suggests to me that the board won't work with grub at all, since it has no PC serial port emulation [07:58] gdeeble: well I suppose ec00 might be a valid PC port number [07:59] rbasak: So that is the breaker for that. [07:59] gdeeble: you could try --port=0xec00 instead of --unit=0 in /etc/default/grub [07:59] gdeeble: it doesn't look right to me though [07:59] gdeeble: I'm also not sure if --port=0xec00 or --port=ec00 is correct [08:00] gdeeble: do you know what UART your board uses? [08:00] rbasak: so change unit to port? [08:01] gdeeble: yes, you might as well try it, though I'm not confident it'll work. [08:01] rbasak: 16c950/954 is what the card detects in kernel. [08:02] Hmm. 16c950 is compatible apparently. So it's down to whether the board maps that to a standard I/O port or not [08:03] I'm trying as we speak to see if it will break or work. [08:03] Try --port=0xec00 and --port=ec00 [08:03] (instead of --unit) [08:03] Trying ec00 first [08:06] Let me make sure I have the letters lower case. [08:06] It hung on boot so hopefully this will work though [08:06] The fact that they're uppercase in your output but 03f8 is lowercase suggests to me that they aren't actually I/O port numbers and so it won't work. [08:07] No that was me just typing them. The show lower in the kernel log [08:08] Sloppy typing on my part :-X [08:08] SpamapS: sendmail wow, why?.... [08:09] twb: what would you suggest instead of webmin then? [08:09] Learning to use the damn CLI [08:10] twb: erhh okay, not what I asked dude [08:11] SpamapS: I have changed my hostname for my server, to the correct domain, but when I create an user account, it still maps to the other domain? [08:11] znow: sendmail is a dinosaur in my mind. postfix, exim, and qmail are all better choices. [08:11] znow: I think few people have sendmail knowledge nowadays. It's not commonly used any more. [08:12] rbasak: oh damn... well im just using it for an application though... and I need to map an account to the right domain... so I just though sendmail would do it.. [08:12] * rbasak last used sendmail in around 1998. [08:13] SpamapS: qmail only looks good compared to sendmail [08:13] znow: it probably will do it, but if you need help then fewer people will be able to help you :) [08:13] It should be postfix unless you have to deal with Debian/Ubuntu weirdos that area already used to exim4 [08:14] twb: I said the same thing when it landed in my lap 7+ years ago.. but it was really impressive how stable it was, and how easy it was to configure. [08:14] znow: postfix and exim are common nowadays. I'd use postfix for a simple task, but I'd use whichever is more popular in the channels that you get help from, and I don't know which that is. [08:14] rbasak: When it started it said error unrecognized number with ec00, with 0xce00 no error, but still hanging. [08:14] I like qmail, and used it heavily in the past. But it doesn't play well with others, and it gets really obtuse when dealing with modern issues such as spam and backscatter. So I shun it nowadays. [08:15] rbasak: those issues are largely handled in some of the alternative distributions of it, since the license allows distributing patched versions. [08:15] rbasak: but postfix is just more broadly known.. so I too shun qmail. :p [08:17] gdeeble: OK so 0xec00 (NOT 0xce00) is the right way to specify it to grub. I don't think this is a grub problem. Serial access in the 16-bit DOS world is really simple, and that's all that grub is using. It's whether your board supports this or not, and what port it maps the UART to. That's all there is to it, but I don't know the answer - sorry! [08:18] The djb dance alone is enough to fail qmail IMO [08:18] SpamapS: the last I knew about it, qmail specifically did not permit the distribution of patched versions. You had to distribute the pristine source and a patch. Has this changed? And I realise that patches have fixed a lot of the issues, but knowing which alternative to use really confuses the matter. [08:19] And at that stage, are we talking about qmail, or qmail-? [08:19] NFI [08:19] rbasak: yes the license was changed [08:19] SpamapS: ah. I think I left qmail for postfix before that happened. Haven't looked back since :-) [08:20] rbasak: I really appreciate you helping with this. Maybe you can come up with a solution for the ob port. :-P [08:21] gdeeble: no problem [08:25] rbasak: Do you have any tips on what to check why I keep getting garbage output from the OB port? [08:26] gdeeble: pinouts for motherboard serial connector pins vary. You might try and figure out which scheme your board is using and examine your pigtail carefully to verify. [08:26] gdeeble: usually the sheet that comes with the motherboard tells you the pinout, and you should be able to google a DB9 serial pinout [08:26] rbasak: SpamapS: okay - ill have a look at those, thanks. [08:26] but for now.. [08:27] I just need the user to map to the correct domain [08:27] znow: I have absolutely no idea how to do that with sendmail. I have much more of an idea with postfix and exim. I suspect others are the same, but you're welcome to keep asking for sendmail-specific help :) [08:29] rbasak: well, maybe its not sendmail ... its just when I create a user, and I "read user mail", when I try to compose a mail it sets the sender to be the wrong domain? [08:30] znow: these are all trivial to solve in postfix [08:30] SpamapS: trivial = easy? [08:30] I'm sure its trivial in sendmail.. if you know sendmail [08:30] znow: yes [08:30] SpamapS: lovely, can I have postfix installed with sendmail? [08:30] or do I need to get rid of sendmaiL? [08:31] rbasak: I feel like an idiot and almost guarantee I have the wrong adapter. [08:33] SpamapS: ill just need to get some breakfast, then im back [08:34] znow: you can only have one of them. Actually I need to get more *sleep* .. so I leave you in the capable hands of #ubuntu-server. ;) [08:34] * SpamapS passes out [08:51] Hi everyone, I've setup SPF record from this guide https://help.ubuntu.com/community/Postfix/SPF and when I check the header in Gmail it's strill neutral. How do I get it to say "Pass"? [09:06] What is a prefered free Web GUI for Ubuntu server? other than webmin? [09:07] hey hey! [09:07] morning folks. [09:08] znow: http://en.wikipedia.org/wiki/Comparison_of_web_hosting_control_panels [09:08] znow: the only real free option is zentyal. [09:08] hex20dec: cheers ill have a look [09:08] Daviey: okay, 2 sek [09:08] znow: TBH, i'd recommend NOT using a webui TBH [09:08] Daviey: not what i am seeking [09:08] znow: what are you looking for? [09:08] Daviey: yes, but CLI instead yeah? well... [09:09] Daviey: as I asked, looking for a webui similiar to webmin but better, and free [09:09] znow: the console is not as scary as people think... [09:09] Daviey: I know the console [09:09] I just want a webui to overview stuff [09:09] Daviey: agree. [09:09] I have dual boot ubuntu and windows7. on laptop. I deleted my ubuntu partitions from windows7 and now it didn't boot ended with Grub error. my CD drive is also not working. how could I fix with which command to make windows7 boot [09:09] znow: I hear you, it's good to have a webui regardless. [09:10] znow: Imho. [09:10] I have dual boot ubuntu and windows7. on laptop. I deleted my ubuntu partitions from windows7 and now it didn't boot ended with Grub error. my CD drive is also not working. how could I fix with which command to make windows7 boot [09:10] ? [09:11] linocisco: Get hiren's cd. [09:11] linocisco tough luck... [09:11] cd drive is not working [09:11] Well, you can mount it on USB. [09:11] linocisco, sad [09:12] You can fix it with ease with Hirens. [09:12] Download it. [09:12] Hirens on USB ? how to? [09:13] They have a tool to mount it to USB. [09:13] I'll link you. [09:13] One sec. [09:15] http://thepiratebay.se/torrent/7018340/Hirens.Boot.DVD.15.1.Restored.Edition-PROTEUS [09:15] Actually. [09:15] Hold on. [09:16] linocisco: You can try this as well, http://www.ultimatebootcd.com/ [09:16] linocisco: But the link I gave you before is not good, don't use it. [09:17] where could I download Hirens on USB ?ok [09:17] http://www.hiren.info/pages/bootcd-on-usb-disk [09:17] They have a special tool for it, as I stated before. [09:19] Hi everyone, I've setup SPF record from this guide https://help.ubuntu.com/community/Postfix/SPF and when I check the header in Gmail it's still neutral. How do I get it to say "Pass"? [09:26] hex20dec, there are so many tools under hirens. which tool and which command I should run? [09:27] linocisco: That's the beauty of it, when you download it, you have so many tools for the long run in case your computer breaks down. [09:27] Well, you go into partitions. [09:28] And you will see a bunch of options. [09:28] hex20dec, yes. like Norton Partition magic. with which i can see partitions. and then what? I do not wanna lose my data [09:28] ok bro. let me try fixing mbr first. [09:28] Yeah. [09:28] Uhmm.. I actually tackled the same exact problem not long ago. [09:28] I used easus partition manager. [09:28] Fixed it. [09:30] And by the way, if losing your data was an option I'd tell you to go the easy way and format, lol. [09:30] So, don't worry. [09:31] hex20dec: does your SPF record end in "-all"? [09:31] hex20dec: actually never mind. That's probably not what you're asking. [09:32] hex20dec: sorry for the delay, yeah [09:32] hex20dec: what would you suggest mate? [09:32] I use webmin, it does it all. [09:32] But I use console more. [09:32] Webmin is jsut there. [09:33] hex20dec: okay, but Im being told that webmin is outdated and not as good as others? [09:34] znow, Don't take my word for these things, as I'm new to the scene as well. For me, it helped me with my stuff. [09:34] hex20dec: cool okay [09:36] rbasak: Thanks again. I repinned the header and low and behold, Serial port on board working. Apparently, I grabbed 1 from my really old pile ;) and it pins 6-9 are even numbered wires on the adapter ;) haha [09:36] gdeeble: excellent! [09:38] rbasak: I would have never thought to check that, as even with dealing with computers for as long as I have, never knew serial cables came with different pinouts hehehe. Problem resolved, this box can now go back to it's corner again :-P [09:38] gdeeble: headless servers FTW :-) [09:41] Si. Now as my next project, convert the old WRT54G to a Serial 2 Network adapter that I can port forward to play with out of network hahaha [09:49] gdeeble, dunno if you're aware of these devices, which do roughly the same thing: http://tibbo.com/ [09:49] damn [09:49] first time in all these years i almost had an opportunity to discuss serial-to-ethernet, but then he timed out :p [10:00] onre: I'm still here. [10:00] No I didn't know about that. [10:04] morning all [10:37] hello jamespage [10:47] morning o/ [10:48] morning [10:53] any recommendations for tools for reading database dumps? normal text editors are bit cloumsy [10:56] Tm_T: grep... [10:56] ...or perhaps you're looking for something like Percona Toolkit? (mysql, that is) [10:56] OZ8AAZ: yeah, I've been grepping (: [10:57] perhaps I'll do some sed magic too [10:57] Tm_T: and awk too? :) [10:57] ye [10:57] :) [10:59] looks like all db browser tools expect having running database server, which I've been trying to avoid [11:19] Tm_T: but u wanna inspect a dump, yes? [11:22] OZ8AAZ: yeah, did find what I was looking for with grep (: [11:22] -o switch saved the day [11:23] do tell... ;) [11:24] I rarely have to care what kind of data db contains as long as it's working and uptodate so takes some time to find a way to work with these when I need to === gema_ is now known as gema [12:09] nijaba: just came to know your coming for ubuntucloud day!! :) [12:09] SpamapS, reviewing/testing your MP for iso-testing now [12:12] in banglore, india :D === Ursinha-bbl is now known as Ursinha [13:06] how do I correctly specify a template for lxc-create on Precise? sudo lxc-create -t lxc-ubuntu and sudo lxc-create -t ubuntu both fail for me [13:27] TeTeT, don't you need a -n XXX as well? [13:28] the template name is def ubuntu tho [13:28] or ubuntu-cloud which uses the cloud tarball instead of debootstrap [13:29] SpamapS, we need to discuss next steps for putting reboot testing live [13:36] jamespage: weird, works fine now: $ sudo lxc-create -n test -t ubuntu -f /etc/lxc/lxc.conf - thanks [13:58] New bug: #813400 in apt (main) "lxc-create -t ubuntu complains about /dev/pts" [Low,Confirmed] https://launchpad.net/bugs/813400 [14:11] #linux-india [14:11] Sorry typo. [14:27] Daviey: I have proposed a fix for 943000, need a review and sponsorship (if applicable, of course0 [14:28] bug 943000 [14:28] Launchpad bug 943000 in cobbler "update to system fails with Exception value: 'NoneType' object has no attribute 'os_version'" [Critical,Confirmed] https://launchpad.net/bugs/943000 [14:29] hggdh: why are you using images without a distro? [14:32] jamespage: How long would it take you to do the iscis test for beta? [14:33] Daviey: why not? it is a valid usage [14:33] hggdh: right, just wondered why.. [14:33] and it is not so much that I am using images without distro... [14:34] the *code* itself makes a call to the method with distro=None [14:34] ah [14:34] hggdh: Were is the patch? [14:34] Daviey: proposed for merge in bug 943000 [14:34] Launchpad bug 943000 in cobbler "update to system fails with Exception value: 'NoneType' object has no attribute 'os_version'" [Critical,Confirmed] https://launchpad.net/bugs/943000 [14:35] hggdh: you don't need to delete MP's to make an amendment :) [14:35] Daviey: about 1 hour I think [14:36] Daviey: I did not think I needed, but this early in the morning I am not reliable to find solutions for dumb problems [14:37] hggdh: missing a bzr add ? [14:37] hggdh: heh [14:38] jamespage: is that something you could crack out today? [14:38] good morning. [14:39] Daviey: heh [14:39] smoser: good afternoon [14:39] silly time zone [14:40] smoser: we invented time, respect us for it.. kkthnx [14:40] morning [14:44] Hi all, I am using LTS and I just updated apache2 via aptitude. The 'deflate' module now breaks some of my hosted glassfish applications! Is this a known bug (workaround is to turn off deflate, but that hits my bandwidth) [14:57] roaksoax, ping [14:57] smoser: pong [14:57] pm [14:58] /query [15:00] jamespage: ping === cwillu_also_at_w is now known as cwillu_at_work [15:30] ivoks: ping you have a typo in your merge proposal [15:31] lynxman, ping - sorry was testing iscsi [15:32] jamespage: no worries, was just wondering if you could give me some tips in launchpad daily builds :) [15:32] jamespage: I see how I can create the recipe for it, I don't see how I can pull upstream each day and build a new version [15:33] lynxman, so the upstream needs to be in a bzr branch [15:33] you can do this even if they are not on launchpad [15:33] create the project on launchpad and add the upstream vcs source - it will then get imported regularly [15:33] see jenkins for an example [15:33] or hadoop or zookeeper [15:34] Daviey, lynxman: I see ipxe can iscsi boot - that might help with iscsi testing [15:34] jamespage: will have a look at any of those then, thank you very very much [15:35] jamespage: good thinking! I didn't thnk the current framework could support two kvm instances tho? [15:35] Daviey, it does not need to [15:35] jamespage: I'm intrigued... [15:35] I don't do it with two kvm instances [15:36] I run tgt on my laptop and setup the targets so they can be seen on the libvirt network that the kvms are on [15:37] jamespage: right.. but for jenkins jobs? [15:37] Daviey: BTW just marked the JeOS on KVM test cases as passed - they get auto tested [15:37] Daviey, the server running the test cases can act as the iscsi target [15:37] in the same way my laptop does [15:38] jamespage: good thinking! [15:39] If I could inject iPXE DHCP boot options into the libvirt dnsmasq instances it would JUST work.... [15:39] an be minimal effort to implement [15:39] * jamespage scratches his chin... [15:39] do we have ipxe on oneiric? [15:47] jamespage: yes [15:48] jamespage: you could also use a ipxe rom, and declare it as a boot param. [15:49] I need to pass it options - just trying to figure out how! [15:54] oh - that actually works! [15:54] iscsi boot with no TFTP! [15:55] \o/ [15:56] authenticated and unauthenticated! [15:58] jamespage: so can we test the initiator ? [15:58] Daviey, at the moment you have to go through this painful step of shuting down the KVM instance [15:59] mounting its target; grabbing the kernel + initrd (which has the target details in it) [15:59] and then restarting again [15:59] if we could get it to iscsi boot using iPXE that step is no longer require [15:59] ipxe just grub boots off the iscsi volume [16:00] jamespage: ok, sounds like you have it inhand [16:00] :) [16:00] hallyn: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/769503 [16:00] Launchpad bug 769503 in linux "KVM network I/O degraded using virtio and freezes guest sometimes" [Medium,Fix committed] [16:01] it says fix commited means its fixed and we are not going to face it in precise [16:02] koolhead17: fix is in kernel [16:02] awesome cheers!1 :) [16:02] (i.e. pls ask on #ubuntu-kernel if there's a problem like it got dropped, i have no control over that) [16:03] agreed it looks like it's been sitting a long time [16:03] i suspect it snuck in through upstream since then? [16:03] but then, you're probably pinging me bc you ran into it? :) [16:03] its on maverick a friend [16:04] ah [16:04] hi guys is it possible two ubuntu server and sync a directory on the other server directory also, and whatever chnages on each other directory it will be the same for both [16:05] you mean something other than nfs/cifs/sshfs? [16:05] hallyn:the easiest way to implement..? you cna suggest [16:06] it depends. if it's just you, i'd say sshfs. But probably i'd say nfs is easiest [16:06] just install nfs-kernel-server on the one, nfs-client on the other, edit /etc/exports on the first, mount on the second [16:07] hallyn:hmm havent tried it do you have some how t link on it.>? [16:09] should be on the server guide [16:09] !serverguide [16:09] The Ubuntu server guide may be found at http://help.ubuntu.com/10.04/serverguide/C/ [16:09] (i dnt' have any other links offhand) [16:12] jamespage: sure, our time zone skew messes with that quite a bit, so we should plan ahead. Are you available now? [16:13] jamespage: (re next steps for getting reboot testing live) [16:14] SpamapS, sure [16:15] I am setting up 10 computers in a lab to print to an IPP printer. My question is this, if i don't care about restricting how/when people can print, what is the win of making the clients use a cups printer on ubuntu server (and having ubuntu submit the jobs) as opposed to simply attaching them to the IPP printer url directly? [16:15] SpamapS, so; now that the code changes are in the main project [16:16] I need a branch containing the reboot tests; I think that you already have this (I branched something from launchpad this morning todo some testing for the reboot bits) [16:16] but it would be good to get it tidied up [16:18] jamespage: right, lp:~clint-fewbar/+junk/server-tests-reboot [16:18] jamespage: and it definitely needs a bit of review [16:18] thats the one - I had to tweak the run_test script to make it work [16:19] jamespage: I was able to use it as-is locally, but perhaps I was skipping steps.. ;) [16:19] hmm [16:20] my KVM instances struggled to talk with couchdb - the python kept borking [16:20] SpamapS, it was fairly minimal - http://paste.ubuntu.com/863660/ [16:21] Can someone suggest the latest Amazon Ubuntu AMI that support FUSE and mounting remote directories with SSHFS? I currently have ebs/ubuntu-images/ubuntu-maverick-10.10-amd64-server-20101007.1 (ami-548c783d) which does not have/support FUSE. [16:22] jamespage: oi.. I had that problem too.. and I have almost the same diff here uncommitted.. *whoops* [16:22] SpamapS, lol [16:22] plural confusion FTW [16:23] jamespage: so ideally that can be merged into the precise tests, and then we'll keep doing things basically the same way, except on the lamp-reboot test, which will install lamp *after* installation, and then reboot, and verify that all services are still running [16:23] SpamapS, ah - so this is just an extra server test - thats fine. [16:23] makes sense [16:24] I which case can you raise a MP against: lp:~ubuntu-server-iso-testing-dev/+junk/server-tests-precise [16:25] jamespage: well the run_test change is also necessary [16:25] SpamapS, yep - that sits in the branch as well [16:25] its specialised for each set of tests [16:26] desktop does different stuff [16:26] jamespage: ahh ok [16:26] dch4pm4n: hang on I'm checking .. [16:26] also means we can do different things with libvirt configuration etc... [16:28] SpamapS: LOL please don't say 7.04 [16:28] dch4pm4n: HAH no I am checking my 11.10 EC2 instance to see if fuse works there [16:29] CONFIG_FUSE_FS=y [16:29] dch4pm4n: 10.10 support ends in < 2 months.. perhaps try 11.10 ? [16:29] what config is that from? [16:29] Yes. I will try 11.10 [16:30] dch4pm4n: thats from 11.10's kernel [16:30] kk, that's what I thought, but didn't want to assume [16:30] Thanks, SpamapS [16:31] SpamapS: One more thing. What's your AMI ID? :-) [16:32] dch4pm4n: you can get it using ubuntu-cloudimg-query on 11.10 or later, or if you don't have that utility, http://cloud-images.ubuntu.com [16:32] well, nvm I'll just use the latest ones clound.ubuntu.com/ami... setting up east and west slave and master failover [16:33] lol.. that defaults to 'maverick' .. [16:33] * SpamapS opens an RT ticket to get that fixed [16:33] What defaults to maverick? [16:33] Ooo it has a launchpad project [16:33] Ohhh [16:33] dch4pm4n: the search on that page [16:33] that link [16:55] jamespage: ok, actually there were 12 revisions I hadn't pushed yet [16:56] jamespage: and junk branches don't allow merge proposals [16:57] jamespage: http://paste.ubuntu.com/863718/ [17:04] jamespage: http://chzb.gr/wXDqbe [17:08] SpamapS, lol - looking now [17:10] SpamapS, testing now for regressions.... [17:11] * jamespage is slightly worried by the fact that there are two people in the channel who's nick starts with Spanky [17:11] jamespage: Spanky99, and Spanky, son of Spanky99 [17:12] * SpamapS gets all emotional thinking that he'll get to see Timmet son of Timmet and Chagga wandering King's Landing in just a month when HBO starts running GoT again [17:19] Hi. I'm using Paralells to install Ubuntu Server 11.10 and I can't get the install work with my DHCP Server. I've set up network to bridget and selected proper network interface. Any ideas? [17:22] hi [17:22] i just reboot my server [17:22] ok [17:22] i got this error [17:22] The disk drive for /tmp is not ready yet or not present. [17:22] is not ready /home /boot /bin /tmp is not yet ready [17:22] I installed a new driver on my server, it appears in /lib/modules/kernelversion/updates, trouble is the kernel is not loading the new driver when it boots. HOw can I check lsmod which module was loaded and how can I change which one it loads ? [17:22] everything is not yet ready [17:23] how i can troubleshoot this error ? * [17:26] anyone there ? [17:26] azertyu: do you have all of those on separate disks from the root? [17:26] azertyu: please be patient, it may take *hours* to get a good answer. Don't expect instant answers. [17:27] no [17:27] there is no separate disks from root [17:27] after googling i can see its an major using ubuntu on server [17:27] problem [17:30] hello [17:30] can't understand ? === sixstringsg|away is now known as sixstringsg [17:33] azertyu: *PATIENCE* [17:34] azertyu: what version of Ubuntu Server? [17:34] lucid ? [17:35] azertyu: ok, can you boot into a recovery console? [17:37] smoser, around? [17:37] reboot recovery console in progress [17:37] smb, here. [17:38] smoser, Just to get back about that stuck cpu you were pinging me yesterday. Has that happened more often? [17:39] what i have to do once after reboot from recovery mode ? [17:39] smb, i only booted the thing once. [17:39] i can launch another and let you poke at it if you'd like. [17:39] that was a cc2.8xlarge [17:39] wait. no, that happened on the cg1.4xlarge. [17:39] the cc2.8xlarge came up fine. [17:40] (only one boot, though, not much for a sample) [17:40] and thank you for following up, smb [17:40] azertyu: you should check that your filesystems have the right UUID's by running 'blkid /dev/xxxx' and comparing to /etc/fstab [17:41] smoser, At least more vcpus than I usually see. I think we should see how that comes out with more boots. The strange thing I see is that vcpu#0 seems stuck in processing softirqs [17:41] smoser, The ip does not change, but the place it is is nothing that could get stuck really [17:41] smb, i'll launch you an instance. [17:42] smoser, I am wondering whether this just was one weird case where one vcpu just was not getting cycles from the hv [17:42] well, i suspect that a cg1.4xlarge is pretty much single occupant. [17:42] at most it would be 2 occupant. [17:42] it has 2 tesla graphics cards and 64G memory. [17:43] smoser, Depending on how many phyisical cores/hts there are it would be shared possibly only with dom0 [17:44] smoser, It is strange though to be not progressing for about a minute [17:44] anyone around? [17:45] smoser, Another case where I wished I could not only look at the guest... :/ [17:45] smb, you can poke. i'll aunch an instance. [17:45] do you want me to ? [17:45] so i can't figure out how to login to my nodes... :( [17:46] smoser, No, it is ok. I am not sure I will see much I fear... :( But let me know when this happens again in testing [17:46] yeah. [17:46] smb, well, to be honest, we don't launch those that often. [17:46] or really any of the hvm. [17:47] i just did it yesterday to test something ,a nd also to at least sniff the beta-1 candidate. [17:47] what key i have to hold to display the grub menu ? [17:47] smoser, It tends to get scary on the credit card bill :) Well I will in parallel try to set up my test box with a precise hvm using all available cores [17:48] azertyu, left shift (at the right time) [17:48] smb, well, its $0.60 right now as a spot instance [17:48] that is what i did yesterday [17:49] smoser, Ah ok, not as bad as huge memory ones then [17:49] I can't ssh to my nodes... well pardon me, I can, but user pwd, and for some reason the master didn't setup login creds. I can't login to nodes. not even from node directly. what am i doing wrong. [17:49] well, in the spot instance prices, everything is reasonable. [17:49] the cc2.8xlarge is $0.54 [17:50] SpamapS: i got the same error even choosing the recovery mode [17:50] press s to skip [17:51] m to manual recovey [17:51] what to do ? [17:52] smoser, Ok, so not really a price issue. Guess I will try a few boot attempts of my own then, tomorrow. And see to have a decent setup locally [17:53] i keep dreaming that someday you'll have something even in that price range that is bare metal [17:53] that woudl really rock. [17:55] smoser, We'd miss so many nice issues... :-P [17:55] SpamapS, that all looks good - but I'm going to let beta-1 out of the door before I push :-) [17:58] Daviey: I just discovered ipxe scripting.... anything is possible now! [17:59] azertyu: m should let you do the same thing.. verify the uuids [17:59] jamespage: sounds good. :) [17:59] SpamapS, it rocks - no more filename="pxelinux.0" [17:59] where would you like to very the uids ? [17:59] when everything not working proerly [18:00] i can't even access to /etc/fstab [18:00] it says file missing === sixstringsg is now known as sixstringsg|noms [18:18] jamespage: Hmm, it might now work in Oneiric... [18:18] jamespage: I encountered an issue before precise current [18:18] azertyu: sounds like your system is having trouble even finding the root partition. [18:18] (which was why he had a new upstream version) [18:19] jamespage: The really neat thing is that you don't need to bundle the script into the binary.. that is quite new. [18:20] zul: still the same error with those changes [18:21] wtf [18:31] adam_g: ok doing another build now [18:34] zul: where are you installing the packages you build? [18:34] adam_g:locally [18:35] adam_g: should be fine now [18:49] adam_g: holy crap that takes its time [18:53] zul: what does? [18:53] the essex deploy-test [18:55] zul: 16 min not so long considering what its doing :) [18:55] true [18:56] zul: however, pointing my browser at the dashboard node gets a Internal Server Error [18:59] adam_g: grrr....gimme a sec [19:03] adam_g: once the horizon builds try that one please [19:05] zul: just fire off another deploy test and point your browser to wherever it got deployed (usually test-10/192.168.20.10) [20:07] hi guys, i need to send a remote reboot command to one server, just one command like ssh user@host sudo reboot, but i'm not getting to do this - can u help me with that? [20:07] zul: yeah, i've noticed [20:08] tyska ~# you cannot reboot without loggin in as root [20:08] tyska ~# or having passwordless sudo or something for the user you login as [20:08] tyska ~# or having the user you login as in the power or powerdev group...or some other important group [20:09] uhhhgg why am i in ubuntu-server....i hate ubuntu [20:09] ya no doubt [20:09] been trying to use their new 11.10 with deploy clusters. [20:09] and ya k, i seem to beable to do what they try to auto mate faster just by hand. [20:11] and get lockouts , like zero answers to why i may not be authenticating properly with pxe booting, i guess I'll break down and do disk install on them once to see if it's just ub bugging out like normal. [20:11] 3days i been looking for info / answer to what issue i am having and all you see is how to install this or that. and not really anything pertaining to finalization of server. perhaps these are odd questions. sry. [20:28] SpamapS: FYI, I've pushed the mysql security updates to -proposed, and have sent out a call for testing [20:32] mdeslaur: thanks I'll echo your call for testing! [20:32] SpamapS: cool, thanks! [20:33] SpamapS: I patiently await the hate mail and the social network beatdown :) [20:37] mdeslaur: I think we've done enough due dilligence at this point that we are rubber and Oracle is glue. [20:38] hehe [20:38] SpamapS: orabuntu? [20:40] lifeless: Linux for Unbreakable Human Beings [20:40] SpamapS: thats a movie, right ? === sixstringsg|noms is now known as sixstringsg|away [20:49] zul: ping [20:49] roaksoax: pong? [20:49] zul: so quick question... I was looking into the horizon packaging you did, and I'm just wondering what should go on python-django-whatever and what in the whatever package? [20:50] roaksoax: python-django-whatever has the py stuff eveything else is in the whatever package [20:54] jcastro: http://nerdapproved.com/wp-content/uploads/2011/09/winter-is-coming.jpg?cb5e28 === bladernr_ is now known as bladernr_afk [20:57] zul: right, but I see in the horizon source, that openstack-dashboard has py stuff too [20:57] urls.py views.py etc [20:57] roaksoax: horizon is basically two apps [20:58] roaksoax: better example might be django-mumble [20:58] zul: yeah looked at it and the same thing :) [21:23] New bug: #829880 in juju "object store doesn't like key with '/' " [Medium,Fix released] https://launchpad.net/bugs/829880 [21:33] in order to ssh to a remote machine ... to copy files ... do I need to copy over .pem s first or should SCP/rsync work right out of the box? [21:33] * delinquentme has a ec2 ubuntu instance hes trying to move local files up to [21:48] hallyn, hi. We found another overlayfs issue yesterday that means xvfb does not work. The xvfb error comes down to the fact that, in an overlayfs, you can't hard link a file with a permission of 0444. See line 17 of http://pastebin.ubuntu.com/864198/ [21:49] hallyn, (1) where did you say that aufs question was going to be sent? :-) I didn't see it [21:49] i think it went to ubuntu-devel [21:49] (2) Do you happen to know anything about that error, or should I just file a bug and hope someone catches it? [21:50] ubuntu-devel: ack. Not subscribed/not a devel. :-/ [21:50] jinkeys [21:50] gary_poster: ping apw on that, i think he's collecting bugs [21:51] (he emailed linux-kernel yesterday or so about it) [21:51] hallyn, ack, thanks [21:51] gary_poster: np. if you want to file it as a bug, i guess that'd be to pad.lv/u/linux [21:51] thanks, ttyl [21:52] ack hallyn will do [22:04] Trying to test my ldap restore procedure. After installing the databases to the barebones test box, I tried to start slapd and got this "Mar 1 16:20:06 fakekdc slapd[2326]: invalid directory /etc/ldap/slapd.d, error 13" message [22:52] so im used to setting path vars within .bashrc .. where should I set them if I'm accessing the remote node over ssh .. and I want those vars to be available? [22:55] delinquentme: environment variables are stripped out by SSH for various security reasons. You need to copy them over to the other side's .bashrc if you want them there too. [22:56] SpamapS, so if i set it in the ~/.bashrc and source that ... theyll b e available? === RoAkSoAx_ is now known as roaksoax === RoAkSoAx is now known as roaksoax [23:28] delinquentme: yes [23:29] hello [23:44] SpamapS: saw the rabbitmq update, we need to update the plugins as well (rabbitmq-stomp rabbitmq-erlang-client), I'll propose merges tomorrow morning [23:46] lynxman: rabbitmq-stomp is included in 2.7.1 [23:46] SpamapS: oh... great news then :) [23:46] SpamapS: then we need to change mcollective dependencies :o) [23:46] lynxman: as is erlang client [23:46] lynxman: well, the packages are still useful, in that they setup configuration for the plugins [23:47] lynxman: we *should* deprecate the source packages, and move those configs/binary packages into the rabbitmq-server package [23:47] SpamapS: hmm whatever we do needs to be done fast then [23:47] lynxman: nothing is "broken" right now [23:47] lynxman: just "weird" [23:48] lynxman: actually I suspect the plugins, on rebuild, may produce binary packages that conflict [23:48] SpamapS: I reckon they would [23:48] lynxman: and agreed, we need to do that this week. [23:48] SpamapS: will get my hands onto it tomorrow morning then [23:48] or rather [23:48] in the next 7 days [23:48] SpamapS: just to get all the ducks in line [23:49] lynxman: I leave it in your highly capable hands.. let me know if you need sponsorship. [23:49] SpamapS: thank you! highly appreciated :) [23:49] lynxman: the main reason I rushed it in is that OpenStack upstream requested it be updated for some HA work they're starting to do. [23:50] SpamapS: yeah HA support in 2.7.1 is way better [23:50] and I think they also wanted a few of the other plugins available [23:50] SpamapS: I've done a bunch of them but just had stomp and erlang-client in the archive [23:52] lynxman: nice that they're all in the official rmq releases now [23:52] SpamapS: a welcome addition indeed [23:54] SpamapS: will check it all tom. morning and see how it fits together and what needs adding/removing