| === Jikan is now known as Jikai | ||
| === sv is now known as discopig | ||
| jasonmsp | hey all.. I have a strange occurance. I copied .bashrc from one of my servers to another one and in the old instance it works fine, but in the new one when I fire up screen all of the data I have at the prompt is underlined. | 01:20 |
|---|---|---|
| sarnold | jasonmsp: screen/tmux can modify how bold/underline/reverse/colors are interpreted.. those are exactly fun problems to track down :( | 01:44 |
| sarnold | jasonmsp: chances are your PS1 environment variable has control sequences in it that are handled funny in screen.. | 01:44 |
| jasonmsp | both servers are running the same version of screen. Screen version 4.00.03jw4 (FAU) 2-May-06 | 01:45 |
| jasonmsp | server a = 10.04 ; server b =12.04 (unbuntu) | 01:45 |
| sarnold | it could be a new terminfo file for screen on 12.04.. | 01:47 |
| === Jikai is now known as Jikan | ||
| jasonmsp | that sounds like it could be it.; i'll have to go back and rewrite the PS1 line another day but not tonight! Thanks! | 01:49 |
| sarnold | good luck :) | 01:50 |
| jasonmsp | 8-) | 01:50 |
| Legend2013 | Hello | 02:06 |
| Legend2013 | is there a difference between a LAN and setting up a local network using a server? | 02:10 |
| sarnold | Legend2013: what do you mean? | 02:10 |
| Legend2013 | The way LANs are explained, it fails to mention anything about a local server being used | 02:11 |
| sarnold | indeed, you do not need a "server" to have multiple machines connected on a LAN | 02:11 |
| sarnold | if you statically assign IPs to your machines, they'll just know how to talk with each other | 02:12 |
| sarnold | if you choose to nominate one computer a dhcp server so you can put the network information in one place, once, and let them allocate IPs on their own, that's fine too :) | 02:13 |
| Legend2013 | ok well that means I dont want a LAN. I want everything to route via a server. | 02:13 |
| Legend2013 | I want to have a master device overseeing all the other devices. What would that be called? | 02:13 |
| sarnold | Legend2013: do you really intend for the different hosts on the network to only communicate via a router? | 02:14 |
| sarnold | you might call that a star topology, with your server as router.. | 02:15 |
| Legend2013 | arnold have you heard of Novell's Open enterprise server? | 02:16 |
| sarnold | yes, but it's been a while, I've forgotten most details.. | 02:17 |
| sarnold | .. something like re-implemented netware services as linux daemons | 02:17 |
| sarnold | .. or netware running alongside linux using xen. I forget the details.. | 02:17 |
| Legend2013 | I want something like that. its damn hard to explain | 02:17 |
| Legend2013 | about 12 linux pcs connected to a linux server through ethernet | 02:18 |
| Legend2013 | they have full system power but the server can track their activities, and internet access is routed through the server | 02:18 |
| sarnold | chances are good you -do- want a LAN, and use that machine as a router between the LAN and the internet. (You probably don't want to buy 13 NICs for this server. :) | 02:21 |
| Legend2013 | I think im missing some vital piece of info that doesnt explain networking as it should | 02:21 |
| sarnold | could be could be | 02:21 |
| Legend2013 | do you know any reliable info sites? | 02:22 |
| Legend2013 | i found this, but not sure if it fits: | 02:22 |
| Legend2013 | http://www.benamy.info/guides/setting-up-a-small-linux-network | 02:22 |
| sarnold | this book is now a bit old (bad for services configuration details..) but ought to still be good for laying the tcp/ip groundwork.. http://shop.oreilly.com/product/9780596002978.do | 02:22 |
| sarnold | that looks like a decent enough 'cookbook' style copy-and-paste sort of source, but it misses a huge piece on the description of tcp/ip networking... | 02:24 |
| Legend2013 | but is the application appropriate? you mentioned the NICs which is a specific piece of redundant hardware for a local network | 02:25 |
| Legend2013 | I better read up on it. Thanks for the guidance arnold | 02:27 |
| Legend2013 | bye | 02:27 |
| === Masshuu_ is now known as masshuu | ||
| === ENOSLEEP is now known as greppy | ||
| jkyle | how can I get the gpg key/server for the ubuntu cloud archive so I can install directly instead of with the ubuntu-cloud-keyring package? | 05:17 |
| === megha is now known as baba | ||
| eltigre | hey, I tried to remove plesk from an ubuntu server and now I am stuck with broken dependencies | 07:20 |
| eltigre | http://pastebin.com/EreeGSkR | 07:22 |
| eltigre | most forum posts etc recommend doing apt-get install -f but that fails miserably... | 07:23 |
| eltigre | any idea what I can do short of wiping the server (which I can't do ...) | 07:23 |
| === akashj87_ is now known as akashj87 | ||
| === guntbert_ is now known as guntbert | ||
| RoyK | anyone seen the Current_Pending_Sector value on a disk been *reduced*? I thought that wasn't possible, but a drive in this server showed Current_Pending_Sector 41 yesterday, and now it's zero... :s | 12:25 |
| qman__ | while I haven't seen it before, that does match the definition of "pending" | 12:35 |
| patdk-wk_ | royk, your suppost to not have current_pending_sectors | 13:05 |
| patdk-wk_ | when it's >0, it means you had a read failure from that location | 13:06 |
| RoyK | I know | 13:06 |
| patdk-wk_ | it's removed when you write to that sector, and it relocates it | 13:06 |
| RoyK | but now it's reduced to zero, http://paste.ubuntu.com/5647819/ | 13:06 |
| patdk-wk_ | and your relocated count went up | 13:06 |
| RoyK | it didn't - see above | 13:06 |
| patdk-wk_ | then it must have gotten a good read then | 13:06 |
| RoyK | but - 41 to zero in a day? | 13:07 |
| RoyK | 41 pending sectors is rather a lot | 13:07 |
| RoyK | and the drive is still giving errors http://paste.ubuntu.com/5647827/ | 13:08 |
| RoyK | ata4 here | 13:08 |
| patdk-wk_ | maybe you just have bad firmware | 13:09 |
| patdk-wk_ | hmm, you have offline-uncorrectables too | 13:10 |
| RoyK | just waiting for the raid to rebuild to r6 - a friend's home server - told him to get new drives | 13:10 |
| RoyK | sda, sdd and sdf shows errors :s | 13:11 |
| patdk-wk_ | no idea why that count vanished though, it shouldn't have | 13:12 |
| patdk-wk_ | what I normally do, to fix them up | 13:12 |
| patdk-wk_ | is a long smart scan | 13:12 |
| patdk-wk_ | followed by writes to the lba_of_first_error locations it finds | 13:12 |
| patdk-wk_ | that normally remaps them to new locations | 13:12 |
| RoyK | I'll run a test once the raid is rebuilt... | 13:14 |
| patdk-wk_ | might have a cable issue too, but that would be a seperate issue | 13:14 |
| patdk-wk_ | the udma crc errors | 13:14 |
| RoyK | any idea what Multi_Zone_Error_Rate is about? | 13:18 |
| patdk-wk_ | multible sectors in a row had errors | 13:18 |
| RoyK | guess sdd is going out the window, then | 13:20 |
| RoyK | http://paste.ubuntu.com/5647847/ <-- do you see any particular bad drives here? | 13:20 |
| patdk-wk_ | WD-WCAVY0219138 looks like it could use a badblocks run on it | 13:22 |
| patdk-wk_ | same for WD-WCAVY0462201 | 13:22 |
| patdk-wk_ | probably the samsung also, thought badblocks on it could be considered overkill | 13:23 |
| patdk-wk_ | not sure what the abrt errors are though | 13:24 |
| RoyK | sd[ab] are IDE drives for the root, in a mirror, so I guess a few errors can be handled - probably possible to get an old drive from somewhere to replace sda | 13:24 |
| RoyK | for sd[bf], I guess new drives like wd red or something should do | 13:24 |
| LargePrime | My Fast Zombies are back and i caught them | 13:56 |
| LargePrime | http://paste.ubuntu.com/5647943/ | 13:56 |
| RoyK | I don't think minecraft zombies are well supported by ubuntu :P | 13:58 |
| Chris_W_ | can anyone help me with monitor resolution? | 14:46 |
| Chris_W_ | i have the monitor saying it is currently using 1600x900 resolution, i have added GRUB_CMDLINE_LINUX="video=VGA:1600x900" to /etc/defaults/grub but the monitor is still only using 3/5 of the screen | 14:47 |
| mdeslaur | Daviey: happy birthday :) | 14:48 |
| ivoks | smoser: https://bugs.launchpad.net/ubuntu/+source/ntp/+bug/604717 | 15:14 |
| uvirtbot | Launchpad bug 604717 in ntp "Please convert init script to upstart" [Wishlist,Triaged] | 15:14 |
| ivoks | jamespage: ^ | 15:15 |
| pythonirc1013 | I've 10 servers that I just created sudo user in. I would like to setup my laptop to passwordless ssh to these boxes and run commands automatically on these 10 boxes. What should I do? | 15:24 |
| ivoks | create ssh key | 15:25 |
| chilicuil | pythonirc1013: after configuring the passwordless settings (with the ssh keys) you may want to use this quick & dirty script: https://pastee.org/m24yg , that's how I do it | 15:27 |
| pythonirc1013 | ivoks: I've to type the password 10 times for that | 15:27 |
| ivoks | pythonirc1013: for ssh key? no... no passwords | 15:34 |
| ivoks | ssh key + sudo NOPASSWD | 15:35 |
| ivoks | http://www.ducea.com/2006/06/18/linux-tips-password-usage-in-sudo-passwd-nopasswd/ | 15:35 |
| pythonirc1013 | ivoks: to place the ssh key on the 10 machines I've to type in passwords | 15:37 |
| pythonirc1013 | I actually have more than 10 machines right now. | 15:37 |
| ivoks | pythonirc1013: what, you'd like anyone to be able to put ssh keys on servers? | 15:40 |
| med_ | pythonirc1013, cry me a river... | 15:40 |
| pythonirc1013 | ivoks: I would like my script to type the password for me. Without learning expect please :) | 15:42 |
| ivoks | sigh | 15:42 |
| ivoks | have you tried typing 'expect ssh' into google? | 15:43 |
| ivoks | first links is 'Expect with SSH - Ubuntu Forums' | 15:43 |
| Pici | Its very bad practice to do that. | 15:43 |
| ivoks | i agree | 15:43 |
| pythonirc1013 | ivoks: looked painful | 15:44 |
| GrueMaster | pythonirc1013: You might also look at python's pexpect. They have good examples on http://www.noah.org/wiki/pexpect. | 15:54 |
| GrueMaster | passmass.py specifically on that site is a good eample of how to do this. | 15:56 |
| Quest | I have a dsl modem (192.168.1.1) -> server for QoS and monitoring (192.168.77.10) -> clients (192.168.77.1-255) . Now in the clients ip configs, I have to add 192.168.77.10 as gateway and 192.168.1.1 as DNS . I want the clients to make 77.10 as the DNS and not 1.1 as then clients would be independant and see only the server. I can change the dsl of 1.1 and say plug in the 2.1 and the client w | 16:08 |
| Quest | ont have to change their ip config. how it can be | 16:08 |
| Quest | done as right now I cant just use 77.10 as a DNS in clients. why? | 16:08 |
| shauno | Quest: would the sensible answer not be to just change the domain-name-servers option on your dhcp server? | 16:12 |
| GrueMaster | Quest: First, it looks like you have an overlap in your client ip range "server for QoS and monitoring (192.168.77.10) -> clients (192.168.77.1-255)". Your dns should be moved out of the way of the clients (i.e. move the clients to 77.15-255). Also, you are trying to span two separate subnets (77.* and 1.1). | 16:13 |
| Quest | shauno, I never installed a dhcp server | 16:13 |
| Quest | GrueMaster, yes. i agree. its a type . its actually 50.255 | 16:14 |
| Quest | shauno, I never installed a dhcp server. i mean i never installed a dhcp software specifically. iam just shairing internet between eth0 and eht1 of server | 16:15 |
| Quest | GrueMaster, ^ | 16:15 |
| GrueMaster | Quest: If you have the dsl connected to one system, and the clients have to go through that system to get outside, why not configure the dsl modem as a pass-through and have your firewall system do all the work? | 16:16 |
| Quest | sure. how | 16:16 |
| Quest | ? | 16:16 |
| Quest | nor i installed a DNS software | 16:16 |
| GrueMaster | That setup is exactly how i have my home network. | 16:16 |
| Quest | do i need a dhcp software or DNS software? | 16:16 |
| GrueMaster | Give me a sec and I'll pastebin my dhcp setup. | 16:16 |
| shauno | it sounds like you're going to need something on .10 to answer/relay those requests if you want it to answer dns. either a small proxy like dnsmasq to answer, or just that port on your internal if to your existing service | 16:17 |
| shauno | but personally, I'd move the clients onto dhcp. the problem of re-writing all the client configs was solved eons ago | 16:18 |
| Quest | shauno, ok. so what do i need. ? (i need simple most setup) | 16:18 |
| Quest | shauno, clients use static ips. not by dhcp | 16:18 |
| GrueMaster | Quest: http://paste.ubuntu.com/5648313 | 16:19 |
| Quest | GrueMaster, ok. do i need a DNS software? | 16:20 |
| GrueMaster | Quest: I use dhcp for every system, then assign them a static IP in my dhcp config (above link). | 16:20 |
| Quest | GrueMaster, i have setup the clients to use a static ip. (for each client) | 16:20 |
| GrueMaster | I have bind setup as well, just for internal dns resolution. | 16:20 |
| Quest | GrueMaster, so i need bind? | 16:21 |
| === wedgwood_away is now known as wedgwood | ||
| Quest | GrueMaster, how hard is to configure bind on ubuntu. for the 77.10 server to pass on client requests to 1.1 ? then i can just put 77.10 for dns and gatway in clients config | 16:21 |
| GrueMaster | YOu oinly need bind if you need to connect between systems internally. | 16:21 |
| Quest | GrueMaster, i only need the clents to have gateway and dns as 77.10 (server ip) and the clients should not worry about the ip of the dsl modem. | 16:22 |
| GrueMaster | i.e. If I want my PS3 to resolve to my master media server. | 16:22 |
| GrueMaster | right. | 16:23 |
| GrueMaster | I assume 77.10 is configured as some sort of proxy for external dns resolution on the internet? | 16:23 |
| GrueMaster | I.e. each machine can ping google.com through it? | 16:24 |
| Quest | no | 16:24 |
| Quest | 77.10 has just traffic monitoring tools . i have shared its eth1 and eth0 with each other. eth1 is Wan. eth0 is Lan | 16:24 |
| GrueMaster | Ah. | 16:25 |
| GrueMaster | So no real firewall? | 16:25 |
| Quest | there is but i dont use it much | 16:25 |
| Quest | well i have done the sharing with the firefwall | 16:26 |
| Quest | firestarter... | 16:26 |
| GrueMaster | I.e. one system that blocks and routes incoming packets. | 16:26 |
| Quest | GrueMaster, yes i gues | 16:27 |
| GrueMaster | Most "firewall" programs for Linux are just a series of scripts that configure the Linux IPTables. | 16:27 |
| GrueMaster | But it sounds like you have something in place. | 16:27 |
| Quest | GrueMaster, either way. it works | 16:27 |
| Quest | it works with 1.1 .but i want to give it 77.10 as dns for clients. so if i change the dsl from 1.1 to 2.1 . the clients wont have to be reconfigured | 16:28 |
| GrueMaster | (I use shorewall). | 16:28 |
| GrueMaster | I honestly can't remember all the details that I did when I initially configured my system (Ubuntu 10.10, upgraded from Mandriva 9). It sits in a rack in my basement and just works. | 16:30 |
| GrueMaster | Looking at my configs, I have bind setup to forward dns requests to my isp. That gives me internet resolution to my home systems. | 16:32 |
| shauno | I honestly think insisting on static IPs is causing the problems you're trying to solve. you're trying to manage the clients instead of managing the network | 16:34 |
| GrueMaster | I agree. Easier to have a master ip list on the dhcp server. | 16:35 |
| Quest | shauno, GrueMaster i have to recognise clients by their ips..... | 16:36 |
| Quest | so i have to give static ips | 16:36 |
| shauno | you can still do this with dhcp, as GrueMaster's config is doing. but it means if you need to change your dns server, for example, you change one line in dhcpd.conf | 16:37 |
| GrueMaster | My dhcp server automagically updates dns, so the only time I need to change anything is when I want to move a new system into a defined range. This also makes it easier to spot and control leaches on my network (wifi drive-by, neighbors, etc). | 16:39 |
| shauno | I know it doesn't answer your question, but it is the fix that keeps on giving. like when you decide your phones need to be on their own subnet, you just match option vendor-class-identifier in dhcpd.conf and the phones will all move themselves | 16:40 |
| shauno | you can also hand out configs and firmware updates to your phones that way, it really is sanity++ | 16:43 |
| GrueMaster | In my case I was doing a lot of Ubuntu testing on arm, which required a LOT of reimaging (sometimes 10x/day). My setup required less configuration, as I could have a generic preseed that worked on all systems, regardless of system name/architecture. | 16:43 |
| shauno | the easiest way to do what you're actually trying to do, just so that I'm not preaching, is to put dnsmasq on .10 | 16:44 |
| jacobw | Preseed are hostname/architecture dependent? | 16:45 |
| GrueMaster | jacobw: Only if you have static ip's and arch specific kernels (in the case of omap4 testing). | 16:47 |
| jacobw | I see | 16:47 |
| GrueMaster | The whole point of preseeding is to automate installation without interaction. I could just nuke the bootloader on a system, reboot, and it would do a new install via pxe. After 10-20 minutes (depending on system speed and package selection), the system would reboot to a login, complete with user ssh keys preinstalled and everything was fresh. | 16:49 |
| GrueMaster | My preseeds were extremely generic, and I had scripts that would customize them as needed on the fly for testing. So if I wanted to build a glusterfs cluster with all my panda boards (7 at the time), 1 would be imaged as the controller, one as the client, and 5 as teh redundant fs cluster. | 16:51 |
| GrueMaster | 20 minutes after launching the job in Jenkins, I had a GlusterFS cluster comprised of (essentially) a cluster of cell phones. :P | 16:55 |
| * GrueMaster enjoys the geek moment in that. | 16:55 | |
| jacobw | GrueMaster: :) | 16:57 |
| === wedgwood is now known as wedgwood_away | ||
| === wedgwood_away is now known as wedgwood | ||
| === wedgwood is now known as wedgwood_away | ||
| pythonirc1013 | how can I ask apt-get upgrade not to ask a "Y/n"? | 17:44 |
| pythonirc1013 | -y , thanks | 17:44 |
| Pici | np | 17:44 |
| Chris_W_ | anyone here? | 17:55 |
| sarnold | Chris_W_: a few hundred people.. | 17:56 |
| Chris_W_ | huh. i only saw you respond. | 17:57 |
| shauno | if 'anyone' got 400 replies each time, it'd be quite unbarable | 17:58 |
| Chris_W_ | true | 17:59 |
| Chris_W_ | im having video issues and i could use some help. | 17:59 |
| Chris_W_ | im using a laptop with a broken screen to run the server, but the vga output is shrunk to like 2/3 of the screen | 18:00 |
| Chris_W_ | if i booted into windows it would think i have two monitor, but im only using the VGA output, so i think ubuntu is doing the same thing. | 18:01 |
| Chris_W_ | is there any way to disable one of the 'outputs' ? | 18:01 |
| sarnold | Chris_W_: many laptops have a key that let you swap between video outputs for easy use with projectors | 18:03 |
| justizin | Chris_W_: also worth a shot, plug in a keyboard and mouse and close the laptop lid. also try booting that way, so it never initializes the LCD. | 18:05 |
| Chris_W_ | ive tried using the fn-f8 combo which would change between lcd/vga output, but id doesnt do anything now. but the lcd is removed and there are no lcd cables connected to the motherboard. it still thinks its there, but it forces video out through vga without me pressing the button combo | 18:10 |
| Chris_W_ | is there a command to see what displays or outputs it is currently using? | 18:12 |
| sarnold | Chris_W_: xrandr without arguments dumps a huge pile of display information | 18:14 |
| Chris_W_ | sarnold: im not using x. im on ubuntu server, with only command line | 18:24 |
| sarnold | Chris_W_: ah :) dmesg may have details on the video devices being used.. | 18:26 |
| === Ursinha-afk is now known as Ursinha | ||
| mathaino | hi all, i am new, so sorry if my question is stupid...how can i install html5 on ubuntu server? | 18:48 |
| sarnold | mathaino: html5 is a specification, not a specific thing that you can install | 18:49 |
| sarnold | mathaino: if you want html5, hire html programmers and tell them what to write :) hehe | 18:49 |
| mathaino | i see, lets say i want to learn html5, do i have to install something on the ubuntu server? | 18:50 |
| shauno | a text editor would be handy | 18:51 |
| === wedgwood_away is now known as wedgwood | ||
| shauno | try it the other way around. just starting learning, and go looking for tools if & when you find something you're lacking. this isn't something that requires any kind of toolbox to get started | 18:52 |
| === wedgwood is now known as wedgwood_away | ||
| sarnold | mathaino: you'll want firefox and chromium-browser, so you can test what you write; you should test your code via a webserver, rather than just accessing files on the filesystem, to prevent silly mistakes from making it through to the code; apache or nginx are useful there, but setting up a full-featured webserver can take some effort | 18:54 |
| mathaino | sarnold: i have already an apache2 running so that is not such a big deal, just want to start playing with html5, that's why i asked if i need to install any package for that. thanks for the hints :) | 18:57 |
| sarnold | mathaino: aha, good :) | 18:57 |
| justizin | are there any plans to make ufw transactional, like iptables-save and iptabes-restore? we're using the opscode's community firewall cookbook for chef, which currently only supports the ufw provider, and when chef runs and reconciles the firewall, all connectivity fails for a brief period. obviously, this is not our preference on prod boxen. ;) | 19:09 |
| jdstrand | justizin: ufw runs iptables-restore under the hood | 19:11 |
| justizin | hm. | 19:11 |
| justizin | i thought it might, i'll have to dig a bit deeper before casting blame ;) | 19:11 |
| justizin | i'm basically just trying to figure out if there's a bug somewhere that i can fix, rather than abandoning ufw | 19:12 |
| jdstrand | justizin: so you could supply your various *.rules file and then just run /lib/ufw/ufw-init reload at the end | 19:12 |
| justizin | could be opscode's cookbook | 19:12 |
| justizin | hm.. | 19:12 |
| === wedgwood_away is now known as wedgwood | ||
| === wedgwood is now known as wedgwood_away | ||
| Blinkiz | Hello. I have a libvirt+kvm server with bridge interface. I do not want the bridge to forward tagged VLANs to every machine (vnetX), how can I filter this? | 20:11 |
| === VD is now known as Guest73519 | ||
| bekks | Blinkiz: Using a bridge, thats not possible. | 20:19 |
| bekks | Blinkiz: A bridges bridges all traffic. | 20:19 |
| Blinkiz | bekks, I found the solution. It is to use ebtables to control the bridge. Like dropping all frames, "ebtables -t broute -A BROUTING -i eth0 -p 802_1Q -j DROP " | 20:22 |
| Blinkiz | I can control which virtual machine gets what VLAN without having to set up a bunch of virtual vlan interfaces in the /etc/network/interface file. | 20:24 |
| Blinkiz | It was a nice feature.. Did not know about ebtables. Am used to setting up everything in the interface file and the default br0 bridge do not forward tagged VLANs. That was changed some kernel ago.. | 20:25 |
| === megha is now known as baba | ||
| bekks | Good to know. :) | 20:26 |
| === smb` is now known as smb | ||
| pythonirc1012 | ubuntu 10.04 LTS was supposed to e supported till 2015 or so, right? | 21:22 |
| sarnold | parts of it, yes | 21:22 |
| pythonirc1012 | sarnold: so I can run it safely in production? | 21:23 |
| pythonirc1012 | sarnold: or do I need a reinstall? | 21:23 |
| sarnold | pythonirc1012: yes, you can run it in production | 21:23 |
| pythonirc1012 | I just updated the 10.04 lts. nginx got updated. And now it won't start | 21:24 |
| sarnold | pythonirc1012: see here for details on how long which pieces are going to be supported: https://wiki.ubuntu.com/LucidLynx/ReleaseManifest | 21:24 |
| pythonirc1012 | anyone uses nginx+ubuntu here that could help? | 21:24 |
| sarnold | pythonirc1012: nginx-extras or something? | 21:24 |
| pythonirc1012 | no just pure nginx - It won't start after the latest update | 21:25 |
| sarnold | pythonirc1012: hrm, this page suggests that last update for nginx on 10.04 was in June, 2012: https://launchpad.net/ubuntu/+source/nginx | 21:27 |
| pythonirc1012 | sarnold: I had not run updates for a long time. I ran update today and got to 1.4.1. now when I try to start nginx - nginx: invalid option: "/var/run/nginx.pid" | 21:27 |
| pythonirc1012 | Even though that option is commented out in nginx.conf | 21:27 |
| sarnold | pythonirc1012: that makes me think perhaps you've got a local configuration error of some sort.. | 21:27 |
| pythonirc1012 | ngingx configtest -> says OK | 21:28 |
| pythonirc1012 | the same config used to work well before the upgrade | 21:28 |
| sarnold | 1.4.1?? lucid's nginx was version 0.7.65-1ubuntu2.3 .. | 21:28 |
| sarnold | even our devel release just has 1.2.6-1ubuntu4 | 21:29 |
| pythonirc1012 | mine says nginx version: nginx/1.4.1 | 21:29 |
| TheLordOfTime | pythonirc1012: are you using the PPA | 21:39 |
| TheLordOfTime | sarnold: they're probably on the PPA, it has 1.4.1-1 | 21:39 |
| sarnold | TheLordOfTime: ahhhhh. | 21:40 |
| TheLordOfTime | see, weird nginx questions like that are why I'm here. | 21:40 |
| sarnold | well, I guess it doesn't much matter one way or another, it's not like it's in main.. | 21:40 |
| TheLordOfTime | sarnold: no it's not in main :P | 21:40 |
| sarnold | (probably should be, but that's another matter :) | 21:40 |
| pythonirc1012 | seems like long time back I added the ppa and then removed it :( | 21:40 |
| pythonirc1012 | readded and then trying to update everything again | 21:40 |
| TheLordOfTime | (oh there's an active merge request to merge 1.4.1 into saucy, sarnold) | 21:40 |
| pythonirc1012 | hopefully I don't break anything | 21:40 |
| sarnold | TheLordOfTime: woot | 21:41 |
| TheLordOfTime | pythonirc1012: if you want something "stable" for production you might want 12.04 | 21:41 |
| TheLordOfTime | but nginx will be 1.4.1 regardless if you stick to the PPA | 21:41 |
| TheLordOfTime | (it's supported through Saucy) | 21:41 |
| sarnold | TheLordOfTime: he's on 10.04, he's still got two years for most of it.. | 21:41 |
| TheLordOfTime | s/PPA/stable PPA/ | 21:41 |
| pythonirc1012 | TheLordOfTime: I'll stick to 10.04 for another year | 21:41 |
| TheLordOfTime | sarnold: he's OK then but there won't be any updates past 1.4.1 for nginx | 21:42 |
| * TheLordOfTime manages the nginx stable PPA and is not going to update past nginx 1.4.1 for Lucid | 21:42 | |
| sarnold | TheLordOfTime: ahhhhhh :) | 21:42 |
| TheLordOfTime | just for their reference :) | 21:42 |
| pythonirc1012 | TheLordOfTime: as long as if there are no security hazards that are patched. | 21:42 |
| TheLordOfTime | pythonirc1012: if there's any new CVEs I may or may not patch them | 21:42 |
| pythonirc1012 | TheLordOfTime: when was the PPA last updated for lucid? | 21:43 |
| TheLordOfTime | pythonirc1012: yesterday | 21:43 |
| TheLordOfTime | with 1.4.1-1 from Debian | 21:43 |
| pythonirc1012 | that is what broke my machine I think | 21:43 |
| TheLordOfTime | pythonirc1012: define "broke" | 21:43 |
| pythonirc1012 | k | 21:43 |
| pythonirc1012 | after update - nginx won't start at boot | 21:43 |
| TheLordOfTime | check the logs | 21:43 |
| TheLordOfTime | it should throw errors about why | 21:43 |
| TheLordOfTime | probably a conf thing | 21:44 |
| sarnold | is that the ipv6_only=on bug or whatever that was? | 21:44 |
| TheLordOfTime | or me forgetting to apply a patch to the default conf... | 21:44 |
| TheLordOfTime | ... crap. | 21:44 |
| pythonirc1012 | when I try to start /etc/init.d/nginx start, it gives me: nginx: invalid option: "/var/run/nginx.pid" | 21:44 |
| * TheLordOfTime facedesks | 21:44 | |
| TheLordOfTime | sarnold: probably | 21:44 |
| * TheLordOfTime checks something | 21:44 | |
| pythonirc1012 | I commented out the line in my conf --> pid /var/run/nginx.pid -- still there was a problem | 21:45 |
| TheLordOfTime | pythonirc1012: you can't comment that out | 21:45 |
| TheLordOfTime | that's a required line | 21:45 |
| TheLordOfTime | but... | 21:45 |
| pythonirc1012 | k | 21:45 |
| pythonirc1012 | I'll uncomment after reboot | 21:45 |
| TheLordOfTime | pythonirc1012: can you do `sudo nginx restart` and pastebin the EXACT ERROR MESSAGE | 21:45 |
| pythonirc1012 | yes, will do in a moment, brb | 21:46 |
| TheLordOfTime | sarnold: probably that bug and the fact that the lucid one doesn't check the right pid location | 21:47 |
| pythonirc1012 | TheLordOfTime: http://bpaste.net/show/FpJHJwUaM4CInWpo04v4/ | 21:47 |
| pythonirc1012 | uncommenting: pid /var/run/nginx.pid; | 21:48 |
| pythonirc1012 | exact same error before uncommenting | 21:48 |
| TheLordOfTime | hmm... | 21:50 |
| TheLordOfTime | that shouldn't happen even on normal updates and upgrades | 21:51 |
| TheLordOfTime | ... and my last lucid VM died a week ago... | 21:51 |
| TheLordOfTime | hmm............................... | 21:51 |
| TheLordOfTime | pythonirc1012: and this is in a production environment? | 21:51 |
| pythonirc1012 | TheLordOfTime: yes :) Well it doesn't have that much traffic yet. | 21:52 |
| pythonirc1012 | I still have my job :) | 21:52 |
| TheLordOfTime | pythonirc1012: the only thing I can think is you're using an invalid conf file | 21:54 |
| TheLordOfTime | but that file there... | 21:54 |
| pythonirc1012 | TheLordOfTime: configtest passes | 21:54 |
| TheLordOfTime | or rather the error | 21:54 |
| TheLordOfTime | the error doesn't say WHERE it's erroring | 21:54 |
| TheLordOfTime | and there's no reason for it to be erroring | 21:54 |
| TheLordOfTime | the program is direct from debian | 21:55 |
| TheLordOfTime | no other modifications | 21:55 |
| TheLordOfTime | and it works on every other release | 21:55 |
| blackthor | hmm... is anyone running a file system with >1 billion of files ? | 21:55 |
| TheLordOfTime | sarnold: now you know why I don't like working with really old releases :/ | 21:55 |
| TheLordOfTime | lol | 21:55 |
| TheLordOfTime | pythonirc1012: which nginx did you install, nginx-full, nginx-light, or what? | 21:56 |
| pythonirc1012 | I don't remember | 21:56 |
| pythonirc1012 | how can I find out? | 21:56 |
| TheLordOfTime | someone else might be able to answer that one better than I | 21:57 |
| sarnold | dpkg -l 'nginx*' | 21:57 |
| pythonirc1012 | TheLordOfTime: http://bpaste.net/show/TPAbbFKnkBAjGvJrksvk/ | 21:58 |
| pythonirc1012 | sarnold: thanks | 21:58 |
| TheLordOfTime | so you're on nginx-full i guess | 21:58 |
| pythonirc1012 | I guess so | 21:58 |
| TheLordOfTime | did you try to reinstall the nginx web server software after backing up your site configs? | 21:59 |
| TheLordOfTime | actually better question have you modified the default config files at all, for the sites-available and such | 21:59 |
| pythonirc1012 | yes | 21:59 |
| TheLordOfTime | back those up | 21:59 |
| pythonirc1012 | k | 21:59 |
| TheLordOfTime | save them somewhere, because i'm going to walk you through reinstalling the nginx webserver software | 22:00 |
| TheLordOfTime | because that looks like a runtime error I've never seen before | 22:00 |
| pythonirc1012 | you want me to backup /etc/nginx? | 22:00 |
| TheLordOfTime | (also did I mention how much I hate working with old releases?) | 22:00 |
| TheLordOfTime | pythonirc1012: i want you to back up what you've modified substantially from there yes. | 22:00 |
| TheLordOfTime | in fact back up everything :p | 22:00 |
| TheLordOfTime | from /etc/nginx | 22:00 |
| pythonirc1012 | done | 22:01 |
| TheLordOfTime | `sudo apt-get purge nginx-full nginx-common nginx` | 22:01 |
| TheLordOfTime | without the ` of course | 22:01 |
| pythonirc1012 | done | 22:02 |
| TheLordOfTime | pythonirc1012: sudo apt-get install nginx-full nginx-common nginx | 22:03 |
| pythonirc1012 | done | 22:03 |
| TheLordOfTime | now restore your configuration files | 22:03 |
| TheLordOfTime | the ones i had you back up | 22:03 |
| TheLordOfTime | (you may need to overwrite the default ones) | 22:04 |
| pythonirc1012 | done | 22:04 |
| TheLordOfTime | sudo /etc/init.d/nginx stop | 22:05 |
| TheLordOfTime | sudo /etc/init.d/nginx start | 22:05 |
| pythonirc1012 | same error: * Starting nginx nginx nginx: invalid option: "/var/run/nginx.pid" | 22:06 |
| sarnold | TheLordOfTime: I wonder if it would help to see the entire config.. | 22:07 |
| pythonirc1012 | Lemme do one thing. Lemme just run it with the default config and see if that starts it? | 22:08 |
| TheLordOfTime | whoopsies | 22:08 |
| * TheLordOfTime exploded | 22:08 | |
| TheLordOfTime | pythonirc1012: yeah see if it can run with the defaults | 22:08 |
| TheLordOfTime | it SHOULD | 22:08 |
| pythonirc1012 | yes it runs with the defaults | 22:09 |
| pythonirc1012 | ah, no it does not | 22:09 |
| TheLordOfTime | hmm... | 22:09 |
| TheLordOfTime | i wonder if this is a lucid problem | 22:09 |
| pythonirc1012 | nginx stop; nginx start; nginx status -> nginx is not running | 22:09 |
| TheLordOfTime | and JUST a lucid problem | 22:09 |
| TheLordOfTime | oh i know that one | 22:09 |
| TheLordOfTime | pythonirc1012: that one's my fault and a known bug | 22:09 |
| TheLordOfTime | because the pid isn't in the right locatoin | 22:09 |
| TheLordOfTime | pythonirc1012: lemme push an update to fix that one | 22:11 |
| pythonirc1012 | k | 22:11 |
| TheLordOfTime | pythonirc1012: do me a favor: pidof nginx | 22:12 |
| pythonirc1012 | TheLordOfTime: no output | 22:13 |
| TheLordOfTime | okay yeah that's a KNOWN bug | 22:13 |
| TheLordOfTime | lemme push an update, but we'll have to wait for it to build | 22:13 |
| * TheLordOfTime shall return after food | 22:14 | |
| pythonirc1012 | I'll leave for food as well | 22:14 |
| pythonirc1012 | will be back later | 22:15 |
| markthomas | Anybody here proficient on LVM storage underneath KVM volumes? | 22:18 |
| TheLordOfTime | pythonirc1012: try again, `sudo apt-get update; sudo apt-get upgrade` | 22:31 |
| TheLordOfTime | then try and run again | 22:31 |
| === Endafy is now known as Guest63387 | ||
| === Guest63387 is now known as Endafy` | ||
| TheLordOfTime | whoops hang on | 23:01 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!