uvirtbot | New bug: #407379 in bind9 (main) "After an update the vpnc is not able to pick up dns information" [Undecided,Invalid] https://launchpad.net/bugs/407379 | 00:01 |
---|---|---|
=== funkyHat is now known as funkyCat | ||
=== funkyCat is now known as funkyHat | ||
=== lamont` is now known as lamont | ||
shadow98 | just trying to find the best option here for failover...i have one server currently and a 2nd server that i am setting up now to do a failover... | 02:49 |
shadow98 | i was planning to rysnc the directories and replicate mysql | 02:49 |
shadow98 | is that the best option... | 02:50 |
shadow98 | this is for my website and sql server | 02:50 |
shadow98 | i was going to replicate /var/www | 02:50 |
shadow98 | sorry rsync | 02:50 |
shadow98 | hello | 03:05 |
qman__ | shadow98, that sounds like a good plan to me | 03:24 |
shadow98 | awesome | 04:13 |
uvirtbot | New bug: #393792 in lighttpd (main) "lighttpd reload executes restart (bad on logrotate!)" [Medium,Fix released] https://launchpad.net/bugs/393792 | 04:21 |
psi-jack | Okay, so if X has a problem finding the 'fixed' font, what package would I need to reinstall to fix it? | 05:29 |
ScottK | One that's not related to Ubuntu Server | 05:33 |
TimReichhart | hi guys i am having problems compiling a drivers for my 56k modem on my server and here what I am getting http://pastebin.com/pastebin.php?dl=m76fb116d | 05:41 |
psi-jack | Do people actually still use modems? o.O | 05:42 |
jmarsden | FAXes are still fairly common in the business world, so yes, they do use modems for FAX transmission and reception. | 05:43 |
TimReichhart | that is correct jmarsden | 05:44 |
jmarsden | The more pertinent question here is why one would build a FAX server and use an unsupported internal modem rather than one known to work with the appropriate FAX server software... | 05:44 |
psi-jack | Okay.. So, I have a gigabit switch and gigabit NIC, but my switch is showing only 100mbit connection, not 1000mbit. Is there something I need to be doing to get it to go up to 1000? | 05:45 |
ScottK | How far apart are they? | 05:45 |
psi-jack | The computer and the switch? | 05:45 |
TimReichhart | jmarsden: its on a research that I am doing on and please I wanted to know how to build a fax server on linux | 05:45 |
psi-jack | ScottK: About... 2 inches.. Cable's a 6 inch. | 05:46 |
jmarsden | TimReichhart: Step one is choose appropriate hardware :) I'll take a look, but... | 05:46 |
psi-jack | Err, well, not 6 inch, but like, 2 foot. | 05:46 |
ScottK | OK. so much for that idea | 05:46 |
psi-jack | And it's cat5e :) | 05:46 |
TimReichhart | jmarsden: but what | 05:47 |
jmarsden | TimReichhart: But you make life difficult when you choose to ignore that step! I have built multiport FAX servers for clients... see http://www.hylafax.org/content/Hardware_Compatibility_List | 05:48 |
TimReichhart | jmarsden: yes I do have a external modem but I dont have a null cable for it so that is why I am trying to get this intel/internal card to work and I know the modem is working bc I can hear a dial tone and I can call out but I just cant get it to send out any faxes..... | 05:49 |
jmarsden | Are you sure the driver for it supports FAXing? What Class is is supposed to handle? 2.0 ? | 05:50 |
TimReichhart | from what I was told this driver is supported for faxing and I dont know what classs its handling on | 05:51 |
jmarsden | So... ask it... use minicom or whatever terminal emulatir you are using for debugging and try the AT+FCLASS=2.0 command, etc... what have you tried exactly? | 05:53 |
TimReichhart | I havent really tried anything except for trying to get this driver to work | 05:54 |
jmarsden | Define "work"? You got it to give you dial tone and dial out... how did you do that? What did you use to make the outgoing calls? | 05:54 |
TimReichhart | I dont remember how I got the modem to make a outgoing call | 05:55 |
TimReichhart | modem to make it to work to make the outgoing call* | 05:55 |
jmarsden | A research you are doing... but you made no notes as you did the research? Lesson #1: Take good notes. | 05:56 |
jmarsden | Looks like the thing you are trying to compile is a kernel module, is that correct? | 05:56 |
TimReichhart | that is what I have been told to do | 05:57 |
TimReichhart | but when I tried to search for the modem there isnt anything showing for the modem | 05:57 |
jmarsden | Been told? This is a research project, you should be the one deciding what to do. Following what other people tell you do to is not research. Do you have the kernel sources installed? | 05:57 |
TimReichhart | how do I know if I got the kernal sources installed? | 05:58 |
jmarsden | Because you installed them as part of your work on this research project? :) | 05:58 |
TimReichhart | i dont think I installed them but how can I make sure I got them installed | 05:59 |
jmarsden | Same way you tell if you have any other package installed... dpkg -l PACKAGENAME -- in this case the PACKAGENAME is probably linux-generic | 06:00 |
psi-jack | Hmmmm | 06:01 |
TimReichhart | see I tried the linux-generic it was telling me to re-install the server part of it when I did the updates | 06:02 |
jmarsden | Be specific... you used what command and got what exact error message back from it? | 06:02 |
TimReichhart | when I did the sudo apt-get update | 06:03 |
TimReichhart | I dont have the linux-generic anymore I got the server part of re-installed | 06:03 |
psi-jack | Bleh. Still no 1gbit. | 06:03 |
jmarsden | TimReichhart: So try installing linux-generic now and then try compiling your kernel module again. | 06:04 |
artillerytx | If i wanted to add a subdomain pointing at a url in bind wouldn't that be a server alias | 06:09 |
artillerytx | or i mean an alias record | 06:09 |
psi-jack | Okay, so it's an on-board 3c940 10/100/1000 NIC that's not going 1000 mbit mode.. On another system, I have it with a nForce 2 10/100/1000 and Linksys 10/100/1000, the Linksys will go 1000 no problem, but neither on-board 1000-capable NIC is going 1000. | 06:15 |
jmarsden | artillerytx: DNS records can only point to an IP address, not to a URL. | 06:30 |
artillerytx | so i would create a new virtual host | 06:30 |
artillerytx | with that subdomain | 06:30 |
jmarsden | Sure, you could do that. | 06:31 |
artillerytx | and then point the dns record to the same ip | 06:31 |
jmarsden | Yes. | 06:31 |
artillerytx | does this work for you http://invoice.longhornpcrepair.com | 06:31 |
nick125 | Doesn't look like it | 06:31 |
jmarsden | Nope, that is not in the DNS here. | 06:31 |
artillerytx | alright oh well no big deal | 06:31 |
artillerytx | i will figure it out later | 06:31 |
qman__ | psi-jack, it is possible that your cable is too short | 06:37 |
qman__ | the specification requires 3' minimum length | 06:37 |
qman__ | shorter ones usually work but not always | 06:39 |
psi-jack | qman__: Oh no, it's not too short. | 06:42 |
psi-jack | The Linksys 10/100/1000 connected to my switch at 1000, but neither the nForce2 or 3c940 on-board NIC's of two computers will do it at 1000. | 06:44 |
qman__ | nforce 2? | 06:44 |
psi-jack | The Linksys one is PCI card. | 06:44 |
psi-jack | qman__: Yeah. | 06:44 |
qman__ | those aren't gigabit | 06:44 |
qman__ | so there's your problem | 06:44 |
psi-jack | Yes, actually, it is. | 06:44 |
qman__ | nforce 4 is gigabit | 06:45 |
psi-jack | Hey, the mobo manual says the NIC is 10/100/1000, so I'm trusting the manufacturer here. ;) | 06:45 |
qman__ | must be a pretty unusual setup then | 06:45 |
qman__ | nforce 2 is pretty old, before gigabit was common or cheap enough for most people | 06:46 |
psi-jack | The new NVIDIA nForce2 Gigabit MCP delivers industry-leading Gigabit Ethernet performance/ | 06:46 |
psi-jack | Not uncommon at all, nVidia even says it. | 06:47 |
psi-jack | And the nforce2 was well after gigabit was out. | 06:48 |
psi-jack | But the 3c940 is also having the same problem, it won't go 1000mbit. | 06:48 |
qman__ | if the cable is tested good, then that's pretty strange | 06:49 |
psi-jack | So far, the only one I've managed to actually go 1000mbit, is the bloody Linksys off-board one. | 06:49 |
jmarsden | psi-jack: Trying a longer and different cable would be worth it at this point, IMO. | 06:49 |
psi-jack | Yep. Same cable that went to the nforce2 that wouldn't go 1000, is hooked into the linksys, and connecting at 1000 | 06:49 |
psi-jack | jmarsden: I tried that too. | 06:50 |
jmarsden | Maybe the Linksys is more tolerant of the too-short out of spec cable than the others? | 06:50 |
jmarsden | OK. | 06:50 |
qman__ | that type of thing is generally handled by the NIC itself, with hooks for the drivers to interfere if needed | 06:50 |
psi-jack | I switched out cables for brand new freshly made cables. heh | 06:50 |
jmarsden | Which were over 3 feet, right? | 06:51 |
psi-jack | Correct. | 06:51 |
qman__ | if the switch is behaving correctly, and the cables are working, the next logical step is hardware failure, but two NICs at the same time is unlikely | 06:51 |
jmarsden | Seems odd indeed. Can you test with a different switch at some point? | 06:51 |
psi-jack | Yeah. Both the on-board NIC's won't do 1000, but are capable of it. | 06:52 |
qman__ | what happens if you connect the NICs to each other? | 06:52 |
psi-jack | jmarsden: Only have the one gigabit switch. | 06:52 |
jmarsden | Borrow one from a friend, move PC to friends house, etc etc... ? | 06:52 |
psi-jack | jmarsden: Heh, I literally just bought the switch from a friend. :p | 06:52 |
qman__ | gigabit is automagic, so you can connect anything to anything | 06:53 |
qman__ | while connecting two NICs in one machine to each other isn't very practical, it should at least get you a gigabit link light | 06:54 |
psi-jack | qman__: That's a negative. | 06:54 |
psi-jack | Gigabit did not go. | 06:54 |
qman__ | do they link at 100, or not at all | 06:54 |
psi-jack | 100, yes. 1000, no. | 06:55 |
psi-jack | And both nic's I see are using the sk98lin driver for them. | 06:55 |
qman__ | so far that's the only thing in common | 06:56 |
qman__ | one way to verify would be, if they have lights for gigabit, to reboot the machine and see if they go gigabit before the OS loads | 06:57 |
qman__ | though that'd be very inconvenient and somewhat inconclusive | 06:57 |
jmarsden | Boot from a *BSD LiveCD and see if its drivers do any better?? | 06:58 |
psi-jack | qman__: I believe when I first booted back up my system it was 1000, initially. | 06:58 |
psi-jack | jmarsden: That's also a thought. I'll prolly try that. ;) | 06:58 |
psi-jack | Once my backups finish. | 06:59 |
qman__ | that wou;d confirm it as being the driver | 06:59 |
qman__ | funny how that works | 06:59 |
qman__ | the driver removes functionality | 06:59 |
psi-jack | Heh. | 07:00 |
psi-jack | Well, both these systems are the last two still actually running opensuse, but that's very soon to be changing to ubuntu-server and ubuntu-desktop within the next few hours, hence the backup. ;) | 07:00 |
psi-jack | Hmm, now, the 3c940 apparently can use the skge driver as well? the sk98lin was actually in the modules.d blacklist. heh | 07:02 |
drurew | !img3 | 07:19 |
ubottu | Sorry, I don't know anything about img3 | 07:19 |
jmarsden | drurew: See http://theiphonewiki.com/wiki/index.php?title=IMG3_File_Format | 07:20 |
drurew | heh, im looking at already... | 07:21 |
drurew | tx tho | 07:21 |
jmarsden | No problem :) | 07:22 |
drurew | does anyone know of any "succesful" attempts at iphonelinux ? | 07:30 |
* drurew waits for a bunch of trouts to fly his way | 07:30 | |
drurew | let me rephrase...*successfully documented (including all sources) attempts | 07:32 |
jmarsden | Unless Ubuntu server was ported to it, that's off topic here... maybe ask at irc://irc.osx86.hu/iphonelinux instead? | 07:34 |
drurew | yeah your probly right...just the only way to install it is with ubuntu 8.10....so | 07:38 |
drurew | k | 07:38 |
drurew | laterz | 07:38 |
jtxx0001 | is it sensible to configure dovecot/postfix to use passwords which are different from the system passwords, particularly for users with root access? | 08:57 |
andol | jtxx0001: Could be, yes. Especially if there is also a webmail, which those users use from wherever. | 09:04 |
jtxx0001 | andol: would the best way to go about this be to use dovecot's authentication with a passwd file? | 09:07 |
andol | jtxx0001: No idea what the best way is for your needs. The easiest is probably to simply to configure dovecot to use another auth source, and have postfix auth by using dovecot. | 09:09 |
andol | jtxx0001: Then of course, there is always PAM. | 09:10 |
andol | jtxx0001: By the way, this is really not my expertiece. I know enough to figure out what I need to do myself, but not really to give other people (good) advice. | 09:11 |
jtxx0001 | andol: i'll look into PAM, but i think i've managed to get it working now. thanks for the help! | 09:12 |
acalvo | Hi | 09:28 |
acalvo | I'm trying to configure a DNS and DHCP server | 09:28 |
acalvo | DHCP is working | 09:28 |
acalvo | DNS seems to be working | 09:28 |
acalvo | but I cannot access anything from outside this lan | 09:28 |
acalvo | I can't even ping my gateway, but I can ping and resolve any other machines in the lan | 09:29 |
uvirtbot | New bug: #413985 in dhcp3 (main) "dhcp3-server fails to install on ubuntu 9.04" [Undecided,New] https://launchpad.net/bugs/413985 | 12:01 |
Rich-Newbie | Hi all, got a question about post fix. I have a simple email server setup, and can send and recive. I have setup masquerading_domain, with this I am able to send to outside email addresses, relaying through my isp. When the email arrives it showes name@server.mydomain.com, would I need to setup aliases to show it as name@mydomain.com. | 12:27 |
jdive | hello folks, i have an issue with a soft raid 5 array, ext4 .. userspace application get's cut from data read on the storage, this last for less than a sec, then goes again | 12:56 |
jdive | if someone could help on how to dig this out,i would appreciate it | 12:56 |
=== |sunny| is now known as dzzz | ||
uvirtbot` | New bug: #412501 in php5 (main) "php5 crashed with SIGSEGV in _Unwind_ForcedUnwind()" [Medium,New] https://launchpad.net/bugs/412501 | 13:25 |
=== Nightlurs is now known as Nightlurker | ||
uvirtbot` | New bug: #403599 in php5 (main) "Installing PHP5 performs a reload to apache instead of a restart" [Low,Confirmed] https://launchpad.net/bugs/403599 | 14:02 |
=== ^grim76^ is now known as Grim76 | ||
Maleko | hmm this chan is as good as dead | 14:22 |
ewook | no, idle is not dead. | 14:22 |
giovani | Maleko: it's a Saturday morning in the US ... clearly most people won't be around | 14:22 |
Maleko | oh.. | 14:23 |
ewook | and brunch/ early dinner in Sweden | 14:23 |
Maleko | weve just had dinner here and are now slowly approaching saturday's midnight in next couple of hours | 14:24 |
Maleko | btw any idea how i could start X through ssh? | 14:25 |
giovani | Maleko: x | 14:25 |
giovani | xorg isn't supported in #ubuntu-server | 14:25 |
ewook | X11-forwarding is a good suggestion. | 14:25 |
ewook | giovani: nah, but perhaps settings in sshd.conf server-side is? :) | 14:26 |
Maleko | i have some apps on my ubuntu server that need to run under wine but its being picky and refuse to launch without window manager. so i then installed both wine and window manager on the os | 14:26 |
Maleko | now i need to start the X.. tried startx but that doesnt work | 14:27 |
giovani | ewook: not when they relate to running xorg on the server | 14:27 |
ewook | giovani: I get your point. | 14:28 |
giovani | Maleko: what applications are you referring to? | 14:28 |
Maleko | gameserver daemon apps that run in windows cmd prompt mode | 14:29 |
giovani | Maleko: sound like applications suited for windows then | 14:30 |
Maleko | maybe but they should be able to run fine under wine with X support on the os | 14:35 |
giovani | Maleko: except that wine and x aren't appropriate on a server setup unless you really know what you're doing | 14:36 |
Boohbah | Maleko: maybe you need to get a windows server | 14:50 |
RoyK | Boohbah: shame on you and your foul language | 14:59 |
uvirtbot` | New bug: #414017 in samba (main) "Reloading /etc/samba/smb.conf smbd only" [Undecided,New] https://launchpad.net/bugs/414017 | 15:50 |
uvirtbot` | New bug: #414109 in samba (main) "package samba 2:3.3.2-1ubuntu3.1 failed to install/upgrade: il sottoprocesso post-installation script ha restituito un codice di errore 127" [Undecided,New] https://launchpad.net/bugs/414109 | 17:01 |
BlueFaceMonster | Hi - I have a question. Trying to set up my first email sever (emix/dovecot). I can connect to the IMAP no worries, and test emails genereted at the server get delievered to my client fine, but I'm not reciving email from external addresses. Any ideas? | 17:09 |
BlueFaceMonster | (and obviously by emix I mean exim 8o\ ) | 17:14 |
BlueFaceMonster | OK, update, so I can "telnet localhost 25" but can't telnet remotely - connection refused. How do I change this? | 17:25 |
Grim76__ | BlueFaceMonster: Sounds like a Firewall/Port Forwarding issue. Also could be an ISP block on port 25 if you are doing this on residential Internet access. | 17:33 |
BlueFaceMonster | Hi, Grim76__. Not residential, but firewall/forwarding sounds about right. Is there a default firewall for ubuntu-server? I thought it might be AppArmor but that's not installed. | 17:34 |
Grim76__ | BlueFaceMonster: I don't recall there being one by default on the last installation that I did. There might be a firewall at the perimeter of your network that needs to be configured to point traffic to the server. | 17:36 |
BlueFaceMonster | Noted, will look into it. Thank you!# | 17:36 |
Grim76__ | BlueFaceMonster: Really quick. Can you telnet to port 25 from another machine on the same network? | 17:37 |
BlueFaceMonster | Not sure how to check that, it a hostes VPS | 17:37 |
Grim76__ | BlueFaceMonster: OK. I know some VPS providers have a firewall that you can self configure. Also you might try an sudo iptables -L to see if there are any rules in place based on their build. | 17:39 |
BlueFaceMonster | To be honest, Grim76__, I've got no idea what I'm looking at. Could you check it for me if you have a chance? http://www.pastie.org/585055 | 17:42 |
jmarsden | BlueFaceMonster: Before digging into iptables rules, are you sure exim is listening on the external IP address as well as on localhost? Try netstat -ntl | grep :25 to find out | 17:44 |
BlueFaceMonster | tcp 0 0 127.0.0.1:25 0.0.0.0:* LISTEN | 17:44 |
BlueFaceMonster | It's not! | 17:45 |
jmarsden | So configure it to do so and life will be better :) | 17:45 |
BlueFaceMonster | you make it sound so easy! :) | 17:45 |
BlueFaceMonster | You know what's coming next, don't you... | 17:45 |
Grim76__ | Thanks jmarsden...Forgot to check for that. | 17:45 |
BlueFaceMonster | ...how do I do that then? | 17:46 |
jmarsden | You;ve never configured exim in your life and have no idea how... right? | 17:46 |
BlueFaceMonster | Genius | 17:46 |
jmarsden | Wait a sec... I use postfix more than exim these days... let me check... | 17:46 |
BlueFaceMonster | Thank you! | 17:46 |
* BlueFaceMonster is slowly getting used to being a newb | 17:47 | |
Grim76__ | BlueFaceMonster: We all have to learn sometime. | 17:47 |
BlueFaceMonster | I only recently discovered ubuntu IRC, and so far I've configured a webserver, made a Python/Pylons webpage from scratch and now this... all in a week. It's great! Hopefully get to "pay it forward" someday and help other | 17:49 |
jmarsden | BlueFaceMonster: Let's try sudo dpkg-reconfigure exim4-config | 17:51 |
Grim76__ | BlueFaceMonster: Look at your configuration and look for dc_local_interfaces according to what I am reading that is where you specify what interfaces to listen. | 17:51 |
jmarsden | Grim76__: Probably better to use the "friendly" config tools rather than manual file editing, don't you think? | 17:52 |
sub | It depends who you ask :-) | 17:53 |
sub | hehe | 17:53 |
Grim76__ | Good point. | 17:53 |
BlueFaceMonster | nah, I'm an expert at "sudo nano /etc/foo.conf" now | 17:54 |
Grim76__ | BlueFaceMonster: Follow jmarsden and ignore me on that one. I am used to editing the config files directly. | 17:54 |
jmarsden | sub: It depends who you are and your level of experience, IMO. I'd hack it, but I've been doing Unix system and network admin for 15+ years... | 17:54 |
BlueFaceMonster | I'm happy with that, just can find exim.conf! Friendly tools it is... | 17:54 |
BlueFaceMonster | * can't | 17:55 |
jmarsden | Hmmm. When you installed exim what did you do regarding configuring it? Didn't you get the debconf dialogs about it then? | 17:55 |
sub | jmarsden: Yes, very true, I agree. I've been doing *nix admin for about 5 total (though I have been using Linux for 10!) | 17:56 |
sub | BlueFaceMonster: The friendly config tool is can be launched via: sudo dpkg-reconfigure exim4-config | 17:57 |
sub | s/is// | 17:57 |
BlueFaceMonster | I did, jmarsden, but I just forgot to add the external IP address. I have now, it was very easy and I am composing a suitable test email to myself... fingers crossed... | 17:57 |
sub | Ah sorry, misunderstood :) | 17:57 |
jmarsden | Good. Looks like all is well, I'll go eat my breakfast... have fun with exim :) | 17:58 |
ivantis | Does anyone here run ubuntu server with mysql installed? | 17:58 |
BlueFaceMonster | Tadaa! Worked like a charm! My thanKs to you all! | 17:59 |
jdstrand | BlueFaceMonster: re firewall in Ubuntu> the default firewall tool is 'ufw'. It is not enabled be default. see 'man ufw' and https://wiki.ubuntu.com/UbuntuFirewall for details | 18:00 |
BlueFaceMonster | will do, thanks. | 18:01 |
jmarsden | ivantis: Ask you real question and find out who can help answer it... | 18:01 |
Grim76__ | BlueFaceMonster: Out of curiosity where did you get your VPS? | 18:02 |
BlueFaceMonster | tagadab.com - can't fault them, service so far has been impeccible and cheap and chips, too | 18:02 |
Grim76__ | BlueFaceMonster: Ok thanks for the information. | 18:02 |
=== JanC_ is now known as JanC | ||
vecy | hey guys i have purchased a Dedicated Server - first time with multiple IP's i have a IP range - 72.232.190.90 to 72.232.190.94 - but when i try to ssh into it using lets say 72.232.190.91 - i never am connecting, do i need to activate these IPs somewhere before i can use them ? if so how ? | 19:35 |
vecy | i am connecting fine using *.*.*.90 | 19:36 |
vecy | but not connecting at all using *.*.*.91-94 | 19:36 |
jmarsden | vecy: Do you have interfaces (or interface aliases) that actually are set to those IP addresses? If not, your sever will not respond to traffic send to those IPs. | 19:37 |
vecy | i honestly dont know, i am logged in as root and i have not set anything besides installed LAMP (apache ..) | 19:38 |
jmarsden | Do you *need* to use the other IPs for something in particular? If so, what? | 19:38 |
vecy | when i type 'ifconfig' i see only the *.90 set | 19:38 |
jmarsden | Then you are only listening on .90 | 19:38 |
vecy | well i want to setup my Apache for 5 websites if ip: 72.232.190.91 go to /home/user1/www if 72.232.190.92 go to /home/user2/www ... and perhaps some other stuff just incase | 19:39 |
vecy | do you know the commands how to set those ips up ? | 19:39 |
Daviey | vecy: Do you know how to do virtualhosts? It's similar | 19:40 |
vecy | ohh true vhosts i totally forgot about them, yes i can use those and bind to domain name even better... :) | 19:41 |
Daviey | thar you go :) | 19:41 |
jmarsden | vecy: if you decide you do need the other IPs, man interfaces and also less /usr/share/doc/ifupdown/examples/network-interfaces.gz will probably get you on your way to configuring interface aliases for the other addresses. | 19:42 |
vecy | thx :) reading on it now.. seems simple lets try rebooting now for changes to take affect.. a reboot is required right ? | 19:45 |
jmarsden | No, just restart networking or probably just ifup the newly defined interface aliases | 19:46 |
jmarsden | But I'd only add them if you need to use them; otherwise, keep it simple. | 19:46 |
vecy | ah i rebooted before ive seen your msg and its been 5minutes | 19:49 |
vecy | server still not up hmm could it get stuck anywhere ? | 19:49 |
jmarsden | That's a long reboot... well, it could if you broke things and misconfigured the network interfaces... | 19:49 |
jmarsden | Maybe you need to see whether your provider offers you remote console access of some kind? | 19:50 |
vecy | well the /etc/.../interfaces file had if i recall eth0 auto... then eth0 ... static my ip .. subnet.. gateway | 19:51 |
vecy | and i just copy pasted that from eth0 static .. 3 times changed ip only | 19:51 |
jmarsden | That's unlikely to work, you need separate interface aliases for each IP, I *think*... like eth0:0, eth0:1 ... which were in the examples. | 19:51 |
vecy | iface eth0 inet static; address 192.168.1.5; netmask 255.255.255.0; gateway 192.168.1.254 | 19:52 |
jmarsden | Get yourself remote console access and see what is going on. | 19:52 |
vecy | it had this line | 19:52 |
vecy | i just copy pasted and changed ip | 19:52 |
vecy | thats just an example ok | 19:52 |
vecy | ill give them a call to see if they can give me remote console access | 19:52 |
sub | try the last IP you configured | 19:52 |
sub | because jmarsden is correct about the interface aliases | 19:53 |
sub | also the gateway only needs to be configured once | 19:54 |
vecy | im such a noob lol first try locked myself out ... waiting for support to reply | 19:58 |
sub | You've gotta learn somehow :P | 20:00 |
vecy | is anyone here with LayeredTech by the way - just wondering where does one find the 'Automatic OS reloads' button which they advertise but not found within the Customer Portal | 20:01 |
vecy | so you guys are saying i should be doing | 20:22 |
vecy | iface eth0 inet static; address 192.168.1.5; netmask 255.255.255.0; gateway 192.168.1.254 | 20:22 |
vecy | for ip #1 | 20:22 |
vecy | iface eth1 inet static; address 192.168.1.5; netmask 255.255.255.0; gateway 192.168.1.254 | 20:22 |
vecy | for ip #2 | 20:22 |
vecy | ? | 20:22 |
vecy | and etc eth2, eth3.. | 20:22 |
jmarsden | No... use eth0, eth0:0 and eth0:1 since you only have one physical network interface, eth0. eth0:X are the aliases | 20:24 |
jmarsden | I'm out of here... on my way to attend a wedding... | 20:24 |
vecy | cool thx | 20:25 |
vecy | should | 20:44 |
vecy | auto eth0 eth0:1 eth0:2 eth0:3 | 20:44 |
vecy | cause any problems ? | 20:44 |
vecy | i added it and not again! it looks like i broke my server again, cant connect to it anymore | 20:45 |
giovani | vecy: you should never be editing core networking config files without an alternate means of accessing the server | 21:01 |
giovani | you WILL break something if you're not very careful and experienced with how the files work | 21:02 |
giovani | however, to answer your question, yes you can auto multiple interfaces on one line, that's fine | 21:02 |
giovani | you probably had a typo elsewhere in the config | 21:02 |
HellMind | When should I chroot? | 21:03 |
giovani | when you have a need for it ... | 21:03 |
giovani | when the process is self-contained within a specific directory structure, or can be easily adapted to | 21:03 |
HellMind | for security is useful? | 21:04 |
giovani | it can be | 21:04 |
HellMind | I want to run a ventrilo server | 21:04 |
HellMind | I dont know if doing that is necesary | 21:04 |
giovani | necessary? definitely not | 21:04 |
HellMind | but its a lot secure or its the same :( | 21:05 |
HellMind | doing that what Im preventing? | 21:06 |
giovani | there's no way for me to give you a simple answer | 21:06 |
giovani | the security advantages are, if the ventrilo software is compromised, they only have access to files that were placed in the chroot | 21:06 |
giovani | and they won't be able to access other files on your server, unless there's a bug or vulnerability in the kernel | 21:07 |
HellMind | but if the attacker gain root? | 21:07 |
HellMind | oh | 21:07 |
giovani | well it's unlikely that they'd gain root through a chrooted server | 21:07 |
infinity | They shouldn't, if Ventrilo isn't running as root, and if there are no root escalation vulnerabilities in your kernel. | 21:07 |
giovani | however, if someone gains root on a server ... it's game over | 21:07 |
HellMind | and there are many levels of chroot, because I saw you can chroot ftp | 21:07 |
giovani | there are not many "levels" of chroot | 21:08 |
giovani | chroot is a concept, it's either in use or not | 21:08 |
giovani | chrooting users to their home directory is a common step where users on a system aren't trusted | 21:08 |
giovani | (the ftp server is an example of where that might be used) | 21:08 |
HellMind | for every user I must creat an enviroment? | 21:09 |
giovani | if you wanted to chroot your users, yes | 21:09 |
giovani | but that's separate from chrooting a particular server binary | 21:09 |
HellMind | and what is the bad about it?, | 21:10 |
HellMind | wasting hd space | 21:11 |
giovani | chrooting is a pain to set up, it sometimes required a lot of work, and if you don't do it properly, you'll spend hours troubleshooting missing binaries, etc | 21:11 |
giovani | s/required/requires/ | 21:11 |
HellMind | I tried and I failed | 21:11 |
giovani | my point exactly | 21:11 |
HellMind | I wonder which files are required to create the environment | 21:12 |
giovani | ldd, and google usually help with that | 21:12 |
HellMind | I can remove all thats unnecesary | 21:12 |
HellMind | I got a guide | 21:15 |
HellMind | but not for ubuntu | 21:15 |
giovani | very little (if any) of this is ubuntu-specific | 21:15 |
giovani | so that's not a problem | 21:15 |
vecy | hey guys i am changing the motd.. are there any variable aliases i can use like Hello %username% welcome! << example ? | 21:16 |
vecy | i havent touched linux much so unsure | 21:16 |
giovani | vecy: /etc/motd is a static file | 21:16 |
giovani | /etc/issue on the other hand allows text substitution along the lines of what you'd like | 21:17 |
giovani | to my knowedge, you can't print the username though -- you could write a wrapper script to handle this | 21:18 |
vecy | ok one more thing lets say i have a program i want to start it everytime i reboot automatically | 21:21 |
vecy | how do i do such thing ? | 21:21 |
vecy | its web server.. but i need to run it manually all the time (xampp) | 21:21 |
giovani | what is xampp? | 21:22 |
vecy | its a combination of apache,mysql,ftp... all in one | 21:22 |
giovani | most server applications installed in ubuntu-server are automatically placed in /etc/rcX.d/ so that they start on bootup | 21:22 |
vecy | but i need to type ./lampp everytime to run it | 21:22 |
giovani | that doesn't sound like something built properly for ubuntu | 21:22 |
giovani | I'd recommend installing packages from ubuntu which will work well together, and properly integrate into the system | 21:23 |
HellMind | root@arctica:/opt/chrooted/ventrilo# su ventrilo | 21:23 |
HellMind | root@arctica:/opt/chrooted/ventrilo# | 21:23 |
HellMind | why I cant su :( | 21:23 |
giovani | HellMind: "su - ventrilo" | 21:23 |
HellMind | its the same | 21:24 |
HellMind | it returns # | 21:24 |
giovani | then you didn't set up the user properly | 21:24 |
HellMind | it says I must set it /bin/false | 21:24 |
HellMind | the shell | 21:24 |
giovani | if you set the shell to /bin/false ... then you can't log in as the user like you're trying to | 21:24 |
giovani | you can do one or the other, but not both | 21:25 |
HellMind | but I can run the server right | 21:25 |
giovani | yes ... | 21:25 |
giovani | but su is going to start the shell | 21:25 |
giovani | which you've disabled | 21:25 |
giovani | so clearly it won't work | 21:25 |
infinity | su -s /bin/sh - ventrilo | 21:26 |
infinity | That will log in as the ventrilo user. Not sure why you'd want to, but there you go. | 21:26 |
giovani | uh | 21:26 |
giovani | let's not provide workarounds like that | 21:26 |
giovani | that serves no purpose | 21:26 |
infinity | It's not a "workaround". | 21:26 |
infinity | If you need to test something as a user with a false shell, that's how you do it. | 21:27 |
giovani | it absolutely is -- either the user has a shell or not | 21:27 |
HellMind | giovani he is a pro | 21:27 |
HellMind | infinity you rock | 21:27 |
HellMind | i dont understand, if I can do that | 21:28 |
giovani | you shouldn't be doing that | 21:28 |
HellMind | and I see am not chrooted, | 21:28 |
giovani | that's the point | 21:28 |
HellMind | how can I see the chrooted environment | 21:28 |
HellMind | How can I see it working | 21:28 |
vecy | i asked this question on another linux channel without response.. lets try here | 21:28 |
vecy | hi guys is there a way to make aliases to commands .. lets say i have /opt/lampp/lampp (a webserver) can i make a global alias so i could just type startwebserver and it would automatically go to /opt/lampp/lampp | 21:28 |
giovani | vecy: yep, you're probably looking to set bash aliases | 21:29 |
giovani | this can be done in the .bashrc file in each user's home directory | 21:29 |
vecy | nice | 21:30 |
HellMind | How do I know my user is being chrooted | 21:30 |
infinity | Users aren't chrooted, processes are. | 21:31 |
HellMind | thats why I asked the lvl of chroot, | 21:31 |
HellMind | if I execute the sever using your tip, the process will be chrooted? | 21:32 |
infinity | Is /opt/chrooted/ventrilo a chroot containing what will be run? | 21:32 |
HellMind | yes | 21:32 |
infinity | If so, "chroot /opt/chrooted/ventrilo" would get you "in that environment". | 21:32 |
HellMind | but how should I start the process to chroot it then ? | 21:33 |
infinity | But, you probably want a nice init script or something that will start your vent server in the chroot. | 21:33 |
HellMind | "chroot /opt/chrooted/ventrilo" will spawn the shell chrooted | 21:33 |
HellMind | i got a .c script but it doesnt do chroot | 21:33 |
giovani | vecy: if you want to make aliases for every user on the system, best to place them in /etc/profile | 21:33 |
HellMind | it only does setgid( and setgroups( | 21:34 |
HellMind | So I should chroot lets say in the /etc/init.d script? | 21:37 |
infinity | Assuming the application doesn't chroot itself (and I'm pretty sure vent doesn't/can't), then doing it in the init script is the cleanest place, yeah. | 21:38 |
infinity | With a nice init script, you can also do some clever things like make sure the chroot is always fresh (copy over libc6, etc) before you start your application. | 21:39 |
infinity | I tend to do sketchy things like mount a tmpfs, copy the libraries that my application depends on, copy in the application itself, then start it. Then you get a nice, fresh chroot every time you restart. | 21:39 |
infinity | (Which means that package updates will stay in sync with the chroot) | 21:40 |
HellMind | if you do a sym link for that | 21:40 |
HellMind | you are unchrooting? | 21:40 |
infinity | You can't symlink out of a chroot.. | 21:41 |
HellMind | thats right :D | 21:41 |
HellMind | so all the symlink on the chroot will be unavailable | 21:42 |
HellMind | if you chroot? | 21:42 |
HellMind | it will show broken link | 21:42 |
HellMind | or something? | 21:42 |
infinity | Well, if they're absolute symlinks using full filesystem paths, yeah. | 21:42 |
infinity | Any symlinks that are internally consistent IN the chroot would be fine. | 21:42 |
HellMind | nice | 21:43 |
infinity | A symlink is literally just a text string. There's nothing fancy about it. | 21:43 |
infinity | So, if "foo" points to "/unf/whatever", if the latter exists in the chroot, the symlink works. | 21:43 |
infinity | If not, it doesn't. | 21:43 |
infinity | The file doesn't need to exist when you create the symlink, just when you want to resolve it. :P | 21:44 |
HellMind | what happen if you do a cycling linking | 21:46 |
HellMind | like a => b => c => a | 21:46 |
infinity | Don't? :) | 21:46 |
HellMind | dont know? | 21:46 |
infinity | Depends on your application. | 21:46 |
infinity | It won't hurt the system in any way. They're just files on the disk. | 21:47 |
infinity | But an application reading "a" could get itself caught in an infinite loop if it's written by people unprepared for such siliness. | 21:47 |
HellMind | but isnt about the app, is about the filesystem, the app doesnt resolve a symlink -_- | 21:49 |
infinity | The filesystem just hands back a pointer to the real inode. | 21:50 |
infinity | Which libc then hands off to the application as the new file handle. | 21:50 |
infinity | If the application then opens that, finds another symlink, and loops... And doesn't notice it's in a loop. | 21:51 |
infinity | *shrug* | 21:51 |
infinity | It's somewhat academic. The answer is "don't do that, then". | 21:51 |
sub | ln -s a b; ln -s b a; cat a gives me "cat: a: Too many levels of symbolic links | 22:19 |
sub | :-) | 22:19 |
r3rman_ | Hey gais - http://my.brandeis.edu/bboard/q-and-a-fetch-msg?msg_id=0000Dn << how does that translate into ubuntu land? | 23:15 |
r3rman_ | I want ot tweak my /proc/sys/fs/file-max, inode-max (didn't find this one!) and perhaps something on ulimit | 23:16 |
r3rman_ | I am getting Too many files open exception | 23:16 |
giovani | r3rman_: it has nothing to do with ubuntu | 23:17 |
giovani | r3rman_: those are linux kernel settings ... feel free to modify them in ubuntu, it's no different than elsewhere | 23:17 |
r3rman_ | giovani, I don't see the inode-max on my install tho' | 23:17 |
giovani | however, those posts are from years ago | 23:18 |
r3rman_ | yeah :s | 23:18 |
giovani | r3rman_: those are from 2.2 | 23:19 |
giovani | long ago, things have changed heavily in 2.6 | 23:19 |
giovani | adjust file-max instead | 23:19 |
r3rman_ | aaah, pkill -9 actually works, whereas killall -p 9 fails and stops when it matches one process you don't own | 23:20 |
r3rman_ | giovani, where is file-max? | 23:20 |
giovani | ... same place as you thought inode-max was | 23:20 |
r3rman_ | w00t | 23:20 |
giovani | googling clearly identifies all of this | 23:20 |
giovani | you should know that these settings don't stay after a reboot | 23:20 |
r3rman_ | yeah, so I'll emacs that file, set it high, like 65535, and I can put a init.d to reconf it though right? or just even ~.bashrc ? (or ~/.profile?) - does it take effect immediately though? | 23:21 |
giovani | you can either place the settings in /etc/sysctl.conf, or in /etc/sysctl.d/ -- read the README in that dir | 23:21 |
r3rman_ | thanks | 23:22 |
giovani | r3rman_: no no no | 23:22 |
giovani | r3rman_: sysctls don't belong in any of the places you mentioned | 23:22 |
infinity | Err. | 23:22 |
infinity | r3rman_: If you're getting "too many open files", it's probably not a kernel setting you're after, but a session limit. | 23:22 |
infinity | r3rman_: See /etc/security/limits.conf | 23:22 |
r3rman_ | aah, right, I bumped into that one before I fell asleep earlier, cheers infinity | 23:23 |
infinity | (See ulimit(1)) | 23:23 |
infinity | Default for open files is 1024. | 23:23 |
r3rman_ | infinity, you think 65535 is the max? | 23:24 |
accol | hey guys random question, if i get the 'host key verification failed' error when trying to connect to a server, what am i doing wrong (this is before i even get a chance to enter a password/username) | 23:24 |
infinity | r3rman_: I think that if you have a process using 65535 open handles, you have a very broken application. | 23:24 |
r3rman_ | infinity, or a very awesome one | 23:24 |
infinity | r3rman_: Raising ulimits is usuall a troubleshooting step, not a fix. :P | 23:25 |
r3rman_ | it isn't broke, it is just slamming two 8 cores like bitches, and ripping up lots of files, across two machines with 200Tb of storage. and doing crazy stuff. I've finally written something to organise my porn | 23:25 |
infinity | Not seeing how 8 threads translates to 65 thousand open handles at once, but sure. | 23:26 |
infinity | You'll be heavily I/O limited if you don't serialise that a bit. | 23:27 |
r3rman_ | infinity, yeah, you are right, I was thinking to dump it all into a berkley db, maybe I will | 23:27 |
r3rman_ | but, I have zillions of threads raeping the nets and when they complete they touch a file and impregnate it with awesome | 23:28 |
r3rman_ | that front end is well tuned, and gets the highest throughput right nao, but as I cranked it up past 11 it didn't like it | 23:29 |
r3rman_ | * soft nofile 65535 < infinity do I need to restart session now? | 23:29 |
infinity | r3rman_: Logging out and logging in should be enough to make it happy. | 23:30 |
infinity | (Or whatever starts a new session for your process... su, sudo, etc) | 23:30 |
infinity | r3rman_: You can verify you got it right with "ulimit -a" and check the values. | 23:30 |
r3rman_ | open files is still 1024 | 23:31 |
r3rman_ | aah need to log in / out | 23:31 |
r3rman_ | I could also run run a ulimit command to set it in memory now? | 23:31 |
infinity | r3rman_: Yeah. Limits are set by PAM, so you need a fresh session. | 23:31 |
r3rman_ | oh ok... I am accessing via vnc... perhaps I can logoff and on graphically... who knows, never tried | 23:31 |
infinity | r3rman_: ulimit will only let you change user limits up to the hard limit, since root controls those. | 23:32 |
infinity | r3rman_: But that would be lost on the next session start, hence why setting the default(s) for the user(s) makes a bit more sense. | 23:32 |
r3rman_ | ok. I've fusked my vnc now, I loaded gnome-session to see if I could logout that way, but that brough about the asdf abfh bug | 23:33 |
r3rman_ | lol | 23:33 |
r3rman_ | ulimit -n still gives 1024.... reshoe tiem | 23:37 |
uvirtbot` | New bug: #414232 in munin (universe) "munin-node tries to "autoconf" /usr/share/munin/plugins/postgres_space_" [Undecided,New] https://launchpad.net/bugs/414232 | 23:48 |
andol | By the way, if someone feels like confirming that one | 23:49 |
andol | ...feel free to do so :) | 23:49 |
psi-jack | Okay.. Since my on-board NIC's capable of doing Gigabit isn't working as I'd hoped they would... | 23:52 |
psi-jack | I'm looking for options of PCI-based NIC's that can do it and are well supported by Linux. | 23:52 |
psi-jack | Intel and or Netgear brands would be preferred. | 23:52 |
psi-jack | Anyone with suggestions of hardware and model numbers? | 23:58 |
giovani | any of the intel pro stuff is good | 23:58 |
giovani | not cheap though | 23:58 |
psi-jack | I don't mind spending about $100/card | 23:59 |
psi-jack | As long as they're worth it. | 23:59 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!