[00:26] hi guys :) [00:26] hi guys, im trying to upgrade a real old ubuntu server dist (6.10) and I encounter this error while trying to install libdevmapper: mkdir: cannot create directory `/dev/.static/dev/mapper': Read-only file system. Any suggestion welcome :) [00:43] T0aD: How are you doing this upgrade? [00:43] like explained on https://help.ubuntu.com/community/EOLUpgrades/Edgy [00:45] Where in that process are you? [00:45] at : sudo ./feisty --frontend DistUpgradeViewText --mode=server [00:45] it downloaded all packages, it was to suppose to remove 1 package, upgrade some, install a lot [00:46] it blocked at removing this single package libdevmapper [00:46] yeah not installing my bad [00:46] Setting up libdevmapper1.02 (1.02.08-1ubuntu10) ... [00:46] mkdir: cannot create directory `/dev/.static/dev/mapper': Read-only file system [00:47] damn that sucks [00:49] Are you running this off the machine itself, or did you boot from something like a live-cd? [00:50] of the machine itself [00:50] its in production [00:50] That sucks. [00:51] definitely [00:52] Is this an LVM volume, or a crypt one? [00:52] nope and nope [00:52] a dummy software raid 1 on 1 single and unique partition [00:53] i know nothing about this devmapper stuff [00:53] Is the software raid Linux software raid, or a funky motherboard software fake raid? [00:54] its mdadm [00:54] Is the raid module loaded? [00:54] its a monolithic kernel and the raid support is running fine it seems [00:55] Your own kernel or a ubuntu one? [00:55] my own [00:55] Hmm. [00:55] I *think* you might be missing a module that ubuntu expects. [00:55] FYI, this is a *WAG* [00:56] hmpf [00:56] wag ? [00:56] Wild Ass Guess [00:56] yeah I understand this is no easy issue [00:56] I thank you for your trial though [00:56] Im wondering if I shouldnt just buy a new box and migrate services but thats gonna be a pain in the ass [00:57] How much data is on this? [00:57] 200 gB [00:57] but it has a lot of stuff === _TechAway_ is now known as _Techie_ [00:57] Ouch [00:57] Got a spare drive? [00:58] a mysql server, a mysql proxy, a qmail install, a vpopmail dir, 10,000 websites of users, a hacked apache install, several CGIs binaries [00:58] This was never going to be easy was it :) [00:58] an imap server, a ftp server, tons of crons, a dns server, Im probably forgetting stuff :) [00:59] well it would be easier if I had more specialized servers separated in farms :) [00:59] but im at the starting point so... [00:59] what about the spare drive ? [00:59] to copy the data ? [01:00] You could do a fresh install on a spare drive. [01:00] on the same box you mean ? [01:00] Yup [01:01] while the server is running ? [01:01] No. [01:01] Hold up. [01:01] like how ? using virtualization or something ? i dont get you [01:01] When you built your kernel, did you start with the ubuntu .config file in /boot? [01:01] yep [01:02] ah you want me to spot the differences [01:02] ? [01:02] # diff -u /boot/config-2.6.24.2lescigales /boot/config-2.6.35.2-rsbac-lescigales | wc -l [01:02] 3590 [01:02] hmpf :) [01:02] That would be helpful. But I'm wondering if you didn't fsk around too much, it may have also compiled the appropriate modules, just not loaded them. [01:03] there are no more modules [01:03] modules are for the weak ! [01:05] Is lvm installed? [01:06] not sure, but it isnt used [01:06] dpkg -l | grep lvm [01:06] nothing [01:07] Does this help: http://ubuntuforums.org/showthread.php?t=620842 [01:09] Better still: http://ubuntuforums.org/showthread.php?t=704273 [01:10] well the first topic is very cryptic [01:10] I suppose the thing to take from it is, do you have any pinning set? [01:10] FYI: I'm going to have to run in 5 minutes. [01:10] "I've got lot work, and I've put lots of effort in this server. Reinstall and re-setup everything could take me longer than a month. For me it's not really a problem, however my boss don't share this opinion." [01:10] lol, sounds like me :) [01:11] pinning set ? [01:11] sorry english ain't my native language [01:12] Have a look in /etc/apt/preferences [01:13] nice it seems the release process is always overwriting logs in /var/log/dist-upgrade [01:13] ls: /etc/apt/preferences: No such file or directory [01:13] What about the second link I showed you? [01:14] unfortunately not useful [01:14] I still dont understand why its written post-installation in this log while it was telling me (I think now.. not so sure) it was about to remove it [01:15] anyway, maybe I should unpack deb file and read this post-installation script [01:15] I still think there is an issue with your kernel, with modules missing. Can you boot into a standard ubuntu kernel and do the upgrade from there? [01:15] any reboot / downtime is risky [01:15] Yeah [01:15] Reading the post-install script is a good idea. [01:15] yeah I might just hack it and blow this stuff [01:16] now i just need to find how to do that [01:16] dpkg -x it seems [01:16] I should be back in a couple of hours, there are others here who might be better able to help. [01:17] owh, I doubt it, but thanks for your time anyway :) [01:17] kirkland: nijaba, have you got anything to add for T0aD's problem? Gotta run. [01:19] well dpkg -x doesnt help but it seems scripts are in /var/lib/dpkg [01:20] lol all the postinst script does is that mkdir === _Techie_ is now known as _TechAway_ [01:30] alright hack successful === _TechAway_ is now known as _Techie_ === _TechAway_ is now known as _Techie_ === MTeck is now known as MTecknology [03:24] What was that fix released recently that helps protect Apache2 servers from clients that connect but don't send a request (perhaps with a timeout)? I have a custom Apache config that I think needs to have this applied quickly. [03:24] well apart from a firewall rule [03:24] I dont see what 'fix' would be realsed [03:24] released [03:25] iptables -I INPUT -p tcp --syn --dport 80 -m connlimit --connlimit-above 12 -j REJECT [03:25] try that one [03:26] Ah, looks like mod_reqtimeout [03:26] oh there is a module for that [03:27] though I can't find it in Karmic [03:27] thats nice, and whats about other services ? [03:27] this module wont help you I think [03:27] well suit yourself [03:28] T0aD: Thanks, but I'm not sure that the connections are coming from the same IP address and the number of available server connections can easily be overwhelmed by not that many parallel connections that take a long time to send the request. [03:28] well there is nothing you can do then [03:28] except configuring Timeout [03:28] If they don't send a request, I just want to time out on them and move on to the next request. [03:29] its in the basic apache configuration [03:29] http://httpd.apache.org/docs/2.2/mod/core.html#timeout [03:30] im personaly using Timeout 3 and KeepAliveTimeout 3 [03:33] I'll give that a try. I thought that perhaps Timeout did not kick in until the request started getting sent which was why the mod_reqtimeout patch was released. [03:34] you should read this module page and experiment [03:34] Timeout 3 rocks anyway [03:35] except for 14.400 bauds modems of course [03:35] :) [03:38] T0aD: Yep, "Timeout 3" does not start the countdown until the first "GET" line is sent by the client. [03:38] T0aD: A client that connects and sends nothing, uses up an Apache child. [03:39] mod_reqtimeout looks like it solves that. [03:39] "Available in Apache 2.2.15 and later" so not in Karmic. [03:40] I wanted to upgrade that server anyway. [03:40] T0aD: Yep, "Timeout 3" does not start the countdown until the first "GET" line is sent by the client. [03:40] hm yes it does [03:41] T0aD: I just tested it on my Apache 2.2.12 server and I can take much longer than 3 seconds to type in "GET /" and it's still listening. If I type "GET / HTTP/1.0" it hangs up 3 seconds later. [03:41] What Apache are you using? [03:42] hm maybe KeepAliveTimeout then ? [03:42] I have that low, too. [03:42] 2.2.10 [03:42] well I tried and its working fine here funny [03:42] but I have this setting for years [03:43] did you correctly restart your server ? [03:43] Since the 3 second timeout is now working after the "GET" (it had been set to the default of 300) I believe I did, yes. [03:43] This also appears to be how it is documented. [03:44] hm maybe Im confused with another directive.. [03:44] no that should be it [03:45] proceed to a full restart (not graceful) to experiment anyway [03:45] root@ns1:/tmp/tmpRktNYU# time nc 217.73.17.12 80 [03:45] real 0m6.020s [03:46] interesting, maybe its adding timeout+keepalivetimeout [03:49] T0aD: I'm off. Thanks for your help. [03:49] np === Pilif12p is now known as Pilif12p|afk [05:46] New bug: #666614 in samba (main) "package samba-common (not installed) failed to install/upgrade: subprocess installed post-installation script returned error exit status 10" [Undecided,New] https://launchpad.net/bugs/666614 [05:58] hey, I have a really dumb question about openldap in ubuntu server 10.10 [05:59] I know that the transition has been made to the cn=config style configuration [05:59] is there any way to set the default password to connect and configure things. dpkg-reconfigure is not doing that === erichammond1 is now known as erichammond [06:28] patdk-lap, why 100 for swappiness? other guides are telling me lower is better, like 10 or 0. [06:47] Hello, I'm having a strange issue with my server, when logon through the terminal at tty1 the "Welcome to Ubuntu!" message is there twice, once for 10.10 and once for 10.4 … any ideas what would cause this? [06:48] cat /etc/motd [06:48] or issue or issue.net [06:49] yep, that give me the welcome message [06:49] holy crap [06:49] upgrade to 8.04 breaks my klogd [06:51] well, the real issue is it says one use is logged into 10.4 and 0 users are logged into 10.10 [06:51] yet, eth0 has 10.10's ipaddress [06:51] so my network is confused [06:53] i really start to think do-release-upgrade is a bad idea [06:53] agree [06:53] im currently upgrading a big production server from 6.10 to at least 8.04 [06:53] and this is hell [06:54] ifconfig says eth0 is 192.168.1.118 but my /etc/motd says eth0 is 192.168.15.4 [06:55] well who cares about motd anyway [06:55] > /etc/motd [06:55] and maybe motd is speaking of your old address [06:55] i don't i'm just trouble shooting my network issue [06:56] well dont you worry [06:56] motd aint involved with networking [06:56] ping 127.0.01 doesn't even resolve [06:56] true, i think it gives me a clue as to what's going on [06:56] try w or who in that case [06:58] right, that tells me i'm logged in, but logged into what? [06:58] 10.10 or 10.4? [06:58] is it possible that they are both running? [06:59] motd says i'm logged into 10.4 [07:01] is it possible that they are both running? [07:01] lol [07:01] so now that [07:01] so now that's funny to you? [07:01] I havent laugh like that for a long time, thank you [07:02] check out your logs [07:02] Im not sure what your problem is exactly [07:04] problem is, i'm getting an IP address network seems to not be functioning properly [07:04] *but [07:05] seems seems pretty clear [07:06] which logs should i look at? [07:07] basically i get 100% packet loss no matter what ip i ping [07:08] even 127.0.01 [07:08] *127.0.0.1 [07:09] hmm [07:09] ifconfig should tell you which interfaces are up [07:09] check that [07:09] yep, all seems well there [07:10] check sysctl icmp_echo to see if you re not ignoring pings [07:10] it could be so many things, firewall, kernel [07:11] error: "icmp_echo" is an unknown key [07:11] sysctl -a | grep icmp_echo [07:12] thanks [07:13] …icmp_echo_ignore_all = 0 [07:13] …icmp_echo_ignore_broadcasts = [07:13] 1 === BlackZ is now known as Guest34369 [07:20] what is eth0:metedata from ifconfig? [07:22] hey guys any dovecot squirrelmail post fix experts in here === twister004_ is now known as twister004 [07:22] i followed the guidelines on the wiki for each of those and im having issues connecting to the imap server [07:22] i have dovecot configured for imaps [07:22] only [07:22] could that be the issue [07:34] morning [07:35] eagles0513875|2: with a self-signed certificate? [07:35] RoyK: think i have isolated my issue to squirrelmail [07:35] k [07:35] i have dovecot setup tp use imaps only how can i config squirrelmail to do the same thing [07:36] RoyK: ^ [07:42] New bug: #666625 in openldap (main) "slapd no longer installs default database since 10.04" [Undecided,New] https://launchpad.net/bugs/666625 [07:47] eagles0513875|2: don't remember - I haven't use squirrelmail for years - but then, if you run squirrelmail on the same host as dovecot, the only thing you'll need is a redirect from the sqm running http, right? [07:47] as in [07:48] ya i have it mostly setup i needed to do some configuration of squirrel mail a bit [07:48] erm [07:48] using sudo squirrelmail-configure [07:48] to allow tls connection as well as use the imaps port [07:48] as in [07:50] Hello ! I have switched from Debian to Ubuntu and I would say that I'm quite happy, good work ! [07:52] When I put service squid stop on a script I have this error : exec: 129: stop: not found, how can I stop and start squid in a script ? [07:53] where are the squid init scripts ? [07:53] dubphil: it would have to be somethign like this [07:53] sudo /etc/init.d/squid stop [07:53] all startup shutdown and restart scripts are in /etc/init.s [07:53] init.d [07:53] eagles0513875|2: thats was the debian way, but there no more squid scripts in /etc/init.d [07:54] dubphil: try sudo squid stop [07:54] hey dubphil im gonna need your expertise with getting squid setup [07:54] eagles0513875|2: it is a script launched by root in a crontab [07:55] yes tell me [07:55] dubphil: where to begin lol [07:56] dubphil: mind if i pm you [07:56] dubphil: you won't notice much difference :) [07:57] dubphil: basically is squid easy to setup [07:57] im going to need it as im hosting a few sites and my server connectivity wise is quite fast, but sometimes it gets bogged down [08:02] T0aD: eucaluptus-cloud was causing the trouble [08:03] eagles0513875|2: yes it is quite easy but more complex scheme you need the more it will be difficult to setup ;) [08:03] sudo apt-get remove eucaluptus-cloud did the trick [08:03] strange [08:04] eth0:metadata went away [08:04] from my ifconfig [08:04] RoyK: yes for sure, the lonely difference for now is the localisation of the squid init scripts [08:04] where damn are they ? [08:05] squid-langpack - Localized error pages for Squid [08:07] RoyK: squid is working well, my problem is just to stop or start it in a root cronjob [08:08] eh I understand Iwould have not use "localisation" I ment "the place they stand" [08:12] sudo apt-get remove eucaluptus-cloud did the trick [08:12] lol [08:14] apt-get remove --purge \* # :D [08:14] * RoyK does not recommend that [08:15] * T0aD wonders if RoyK thinks rm -fr / is funny too ? [08:15] T0aD: that's far worse [08:16] that apt-get gives you a pretty good warning [08:16] its still as funny as eddie murphy [08:16] that rm of yours doesn't [08:16] * z3cka wonders if T0aD is really a troll [08:16] * T0aD hides his hair [08:17] z3cka, I just dont see why that is 'funny' :) [08:17] T0aD: i just don't see why you have to quote me and lol [08:18] well that was funny [08:18] installing a cloud and having networking issues [08:18] and removing it being called 'a fix' [08:18] you got me [08:18] that's true [08:18] well… i wasn't using it [08:18] yeah, that was truly funny :) [08:19] i'm sure is i installed it again it would probably work [08:19] *if [08:19] no idea, im still a cloud virgin [08:20] * T0aD is the innocent little flower [08:20] played with it a bit on the local lan but didn't get far [08:21] and then the update seemed to break the rest of the network [08:21] more specifically added eth0:metatdata entry to ifconfig === _Techie_ is now known as _TechAway_ [08:22] so i searched eth0:metadat and all i got was entries about problems with eucalyptus [08:22] so i removed it [08:23] hense, fixed :-) [08:23] you cant trust software written by koalas [08:23] errr *hence [08:23] that's true [08:24] or just Australians in general... === _TechAway_ is now known as _Techie_ [08:33] adios [08:40] hi there, i am at my devt machine. I ssh into my server. From my server, i am trying to do a git pull from a public repo like projectlocker. I am always prompted for my ssh passphrase. i followed this but i am still prompted. please advise.http://help.github.com/working-with-key-passphrases/ [08:42] IMO agent forwarding is evil and wrong [08:42] I prefer -oProxyCommand [08:42] Admittedly it wouldn't help in your case... [08:43] Zeu5: that article doesn't seem to mention agent forwarding. Do you have it on, or off? [08:44] Zeu5: btw, if you're in an Ubuntu desktop environment, ssh-agent is started automatically when you log in, so you can skip that step. [08:49] twb: this is my config file [08:49] this is my config file in ~/.ssh http://gist.github.com/646426 [08:49] twb: does that indicate whether i have agent-forwrading on? [08:50] And you're running "ssh staging"? [08:50] twb: i am actually trying to get capistrano working. but somebody told me to do this. so i did so. if it s wrong i will gladly remove it [08:50] capistrano is a hostname? [08:51] capistrano is a deployment tool [08:51] twb: let me go remove the config file first [08:51] Bleh, more ruby gank [08:51] twb: i am a ruby noob. more so than being a ubuntu noob. ha [08:52] Unfortunately I have no idea what that tool does [08:52] its okie twb [08:52] right now i just want this ssh forwarding to work. [08:52] Rather than sshing into a server and doing a "git pull" there, I guess what I would do is "git push" to the server directly [08:53] You can try "ssh-add -l" to see what keys are enabled in your agent on each host [08:53] Note that you need to run ssh-add [~/.ssh/id_rsa] to add a key to your agent. [08:54] is it safe for me to paste my output when i run ssh-add -ls [08:54] is it safe for me to paste my output when i run ssh-add -l ? [08:55] A line like this is safe to publish: 2048 4d:cd:2c:5b:9f:a5:d1:cd:f6:f1:2f:30:f8:74:8b:47 /home/twb/.ssh/id_cyber (RSA) [08:55] It's basically the same as your public key [08:55] (In terms of information disclosure.) [08:56] username@server:/var/www/abc.biz$ ssh-add -l [08:56] 2048 bf:8a:cf:e1:97:24:86:6d:8f:8a:f0:7b:86:17:5c:7f /home/ubuntu/.ssh/id_rsa (RSA) [08:56] so am i doing everything correctly so far? === _Techie_ is now known as _TechAway_ [08:57] That means "server" has your key usable in it [08:57] So it should work without prompting for a passphrase [08:58] twb: i am still prompted when i do a git pull from server [08:58] What is the command you're running? [09:00] twb: http://gist.github.com/646508 [09:03] OK, for some reason it is ignoring /home/ubuntu/.ssh/id_rsa and trying to use /home/username/.ssh/id_rsa instead (or first). [09:03] You could try moving that key out of the way, but other than that you're going to have to ask #openssh, since this is where my experience with agent forwarding runs out. [09:03] (You can also try just hitting ^D which will make it try the next key.) [09:04] twb: i apologise beforehand. i wanted to keep my username secret. its all ubuntu [09:04] twb: sorry for the misunderstanding [09:07] twb: i have to go off. will be back in an hour's time to try my luck again. thank you for your assistance. appreciate it. sorry for misleading you about my username. will definitely try #openssh as well [09:08] Which, kids, is why we you don't try to elide useful, non-compromising information from your tech support staff. [09:13] any recommendation for resolving the following apt-get upgrade errors? http://pastebin.com/WByrVRjp === _TechAway_ is now known as _Techie_ [09:19] MACscr: turn -security back on? [09:19] huh? [09:20] libssl-dev: Depends: libssl0.9.8 (= 0.9.8k-7ubuntu8.1) but 0.9.8k-7ubuntu8.3 is installed [09:21] 0.9.8k-7ubuntu8.3 comes from lucid-security and lucid-updates. [09:21] The fact that it's trying to install an older version suggests that either those components are disabled, or you haven't run "apt-get update" lately. [09:22] I suppose it could also happen if you're using a binary package that needs a binNMU (i.e. it links against the obsolete version). lucid-security should fix that, too. [09:27] its enabled [09:27] all the defaults are [09:27] and i ran apt-get update right before i attempted the upgrade [09:29] nvm, its working now. Weird [09:29] simple reboot seemed to fix it [09:49] hi there, i am at my devt machine. I ssh into my server. From my server, i am trying to do a git pull from a public repo like projectlocker. I am always prompted for my ssh passphrase. i followed this but i am still prompted. please advise.http://help.github.com/working-with-key-passphrases/ [10:18] Zeu5: do you start ssh-agent ? [10:18] mgolisch: how do i do that? [10:18] i did notice that when i run "ps", there is no ssh-agent [10:19] It may be called "gpg-agent" (it can do both jobs). [10:21] twb: when i run ps all i see is ps and bash [10:21] is ssh-agent installed at all? [10:22] New bug: #666682 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 5.1.37-1ubuntu5.1 failed to install/upgrade: subprocess new pre-removal script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/666682 [10:22] this page you linked seems to add stuff to your shellconfig to start ssh-agent automaticaly [10:22] that's keychain's job, yeah [10:23] Except that if you do a GUI login to an ubuntu desktop, ssh-agent is already started automatically, and (I assume) agent forwarding means THAT agent is the one used on the remote server. [10:23] twb: i apologise. i am really not very good with this whole agent forwarding business. so are you saying tat i made a mistake in the first place? [10:24] Damned if I know. [10:24] I'm just saying that (AFAIK) if you login with gdm, you don't need keychain [10:30] twb: since i am sshing into server, i guess tat means i am not login using gdm yes? [10:32] You're using agent forwarding, so your local agent (started by gdm) will talk over the ssh link to *something* [10:32] Whether that something is ssh (directly) or a remote ssh-agent (indirectly), I don't know. [10:32] As I said, I'm not a fan of agent/x11 forwarding. [10:34] twb: i see. thank you. === _Techie_ is now known as _TechAway_ [11:04] I have ubuntu server 10.10 64 bit installed and have to install Conexant HSF softmodem driver for making hylafax work with my softmodem PCI card. It is a Conexant System modem. I am following this article http://www.linuxant.com/drivers/hsf/downloads-installer.php but when I run the script I get this error: Package not compatible with your system. Kernet module can't be compiled [11:04] what do I have to do, please? [11:05] check what distros are supported [11:05] 10.10 obviosly isn't [11:05] or at least 10.10/64 [11:05] http://www.linuxant.com/drivers/hsf/full/downloads.php [11:06] not a very promising list [11:06] no x64 at all [11:07] I guess perhaps an x86 install of 10.04 or even 10.10 might work with the bottom one [11:07] x64 will not work [11:11] RoyK, I have to find out if I can make hylafax work with my modem. I am not able to find any distro or a Hylafax IRC Chanel to ask. How can I make this work? [11:12] Any advice? [11:12] even by reinstalling ubuntu or another distro [11:14] RoyK, what if I install the 32 bit version of ubuntu 10.10? If not can I install a earlier verion of ubuntu server that wil be compatile with ? [11:22] anyone can reply, pleaze [11:23] reply [11:24] SAngeli: i would go for 32bit [11:26] of 10.4 [11:27] Apart from legacy / proprietary programs and specialist compilers, I see no reason to install i386 on an amd64 system. [11:28] ("specialist" meaning that it has one developer, and he's too overworked to port it to x86-64) [11:37] adamk-pl, can you please tell me where to download the 32 ver or ubuntu server 10.4, please [11:37] and thank you for your reply [11:38] adamk-pl, is it this one? Ubuntu 10.04.1 LTS (Lucid Lynx) [11:39] adamk-pl, found it: http://mylayn.blogspot.com/2010/05/download-ubuntu-104.html [11:39] SAngeli: uh, maybe you should try ubuntu.com; it probably has a download link on the front page [11:40] http://www.ubuntu.com/server/get-ubuntu/download [11:41] twb: exactly ! [11:42] mylayn links point to torrent files, so if you prefer torrent go for torrent, i think i still seed LTS [11:42] :D [11:43] Hi everybody! On my server I get dependency problems when calling apt-get upgrade: http://hpaste.org/40892/upgrade ... can anybody tell me how I can fix these? [11:44] adamk-pl: I would trust any link I found on [11:45] evident: please run "export LC_ALL=C" and re-run the command [11:45] twb: sure [11:45] evident: that will give it in English :-) [11:46] evident: also, paste the *whole* output. You can do this by running "script", which will make a copy into the file "typescript" [11:46] http://hpaste.org/paste/40892/upgrade_2#p40893 [11:46] i annotated the paste with the full output (in english) [11:47] OK, thanks. [11:47] (other question: Will my outputs be in english forever now? If so: Great... been trying to do that for a while) [11:47] The procps issue shouldn't be happening; have you got bad entries in /etc/sysctl.conf? [11:47]