[00:08] <wotwot> hi
[00:14] <KB1JWQ> !hi
[00:16] <slicslak> so, what do you guys say, is `nohup /bin/foo_daemon &` an appropriate replacement for a real init script?  :)
[00:18] <twb> That all depends
[00:18] <twb> You would probably do better to make an upstart job
[00:19] <twb> http://paste.debian.net/106373/ <-- example
[00:20] <slicslak> thx twb, i had been looking at start-stop-daemon earlier and just hadn't had the chance to write the init script yet.  i'll check your paste
[00:23] <wotwot> is there a currently working upgrade path from dapper to lucid ?
[00:31] <twb> slicslak: you have upstart; don't bother
[00:32] <twb> slicslak: if you REALLY want to use s-s-d and sysvinit, look into the metainit(8) package, which will turn something like that upstart.conf into a sysvinit script.
[02:07] <slicslak> twb, oh interesting
[02:07] <slicslak> ya, upstart looks be good enough
[02:07] <twb> Note that metainit is currently dead in the water, because init developers are insular, NIH-happy bozos
[02:08] <slicslak> lol
[05:28] <fluvvell> any tips on how to boot 8.04 server cd under virtualbox without a kernel panic?
[05:30] <jmarsden> fluvvell: I think it needs the PAE extensions visible to it.
[05:32] <fluvvell> jmarsden, yeah thought I'd just turned them on - still panic
[05:33] <fluvvell> pae/nx extended features enabled.
[05:33] <jmarsden> Hmm, I have a Hardy server virtualbox VM here that works... I rememebr there was something I had to add to make it work, but it was created a while ago...
[05:34] <fluvvell> yes, I have another that works, created a while ago too. Its just I need a new one for a transfer task. :(
[05:35] <fluvvell> is there a vga=   something setting to set in boot options ?
[05:35] <jmarsden> OK, let's see if I can create a new one here :)
[05:36] <jmarsden> I'm pretty sure I didn't have to do that...
[05:38] <jmarsden> fluvvell: Enable IO APIC :)
[05:38] <jmarsden> (I created one which crashed and read the traceback... which was about the APIC...)
[05:39] <fluvvell> Ah
[05:39] <fluvvell> wow, thats it.
[05:41] <fluvvell> Yes there was that in the traceback - of course! Thanks jmarsden!
[05:41] <jmarsden> fluvvell: You're welcome.
[05:46] <fluvvell> Well thats an 8.04 bare bones server install completed and logged in, 34Gb virtual hard disk space!
[05:46] <fluvvell> Was that 5Min???
[05:46] <EvilPhoenix> 6.34 mins.
[05:46] <EvilPhoenix> since your last post.
[05:47] <fluvvell> Astounding.
[06:04] <twb> LP570805
[06:04] <twb> LP #570805
[07:01] <chaunch> hey guys
[07:02] <chaunch> im trying to setup openvpn to tunnel all traffic to my VPS, but I need to setup a bridge interface to do it
[07:02] <chaunch> im having problems doing it though
[07:03] <soren> twb: You asked yesterday (or the day before) why su forks instead of just execing.
[07:03] <twb> soren: yes
[07:03] <soren> twb: It's because it needs to know when the process finishes so that it can close the pam session.
[07:04] <soren> twb: If you compile su with out pam support, it just execs.
[07:04] <twb> Hum, OK.
[07:04] <twb> But why is there a session when I'm not using "su -" ?
[07:05] <chaunch> https://help.ubuntu.com/10.10/serverguide/C/network-configuration.html#bridging
[07:05] <soren> twb: Million dollar question.
[07:05] <chaunch> am i suppsoed to replace the original /etc/network/interfaces file with the one there?
[07:09] <soren> twb: For security, I suppose. If I configure some sort of auditing in pam's session stack, I don't want people to be able to evade that by not passing - to su.
[07:11] <soren> twb: It's a half-baked argument, though. At best.
[07:11] <twb> Oh well; I'm fighting lucid-security's sync()-happy dpkg right now, which is way more important
[07:17] <soren> twb: Oh. It's the session handling that does stuff like mount the user's home directory and such. That sounds pretty useful :)
[07:17] <KB1JWQ> soren: Is that automount stuff?
[07:18] <soren> KB1JWQ: Depends on your PAM configuration. For me, for instance, it's ecryptfs.
[07:19] <KB1JWQ> Ahhh.
[07:19] <KB1JWQ> Interesting, I haven't played with encrypted home directories on servers yet.
[07:21] <soren> KB1JWQ: Well, this isn't really server specific.
[07:21] <KB1JWQ> soren: Right, but this channel somewhat is. :-)  I tend to want servers available to serve data from various homedirs regardless of who's connected at any given point.
[07:26] <twb> Is ecryptfs as shitty as pam_mount?
[07:26] <twb> In terms of mounting, I mean -- irrespective of crypto-y-ness
[07:37] <goddard> I need some help pxe isnt working
[07:38] <goddard> trying to netboot lucid server from windows 7 using tftp and this http://www.liberiangeek.net/2010/05/how-to-install-ubuntu-10-04-lucid-lynx-without-cd-usb-via-the-internet/
[07:38] <goddard> but it doesnt work
[07:38] <goddard> any ideas?
[07:39] <goddard> no more address or address previously allocated by another server [02/02 23:31:02.324]
[07:39] <goddard> Rcvd DHCP Discover Msg for IP 0.0.0.0, Mac
[07:43] <goddard> it connects to dhcp i think just says no boot file name received
[07:54] <twb> Windows 6.1 ships with a TFTP server?
[07:55] <twb> goddard: the DHCPACK needs to pass the filename option, setting it to a string denoting the path to pxelinux.0 within the TFTP tree
[07:56] <goddard> ahh ha
[07:56] <goddard> got it
[09:21] <goddard> im trying to install ubuntu to compact flash and it isnt auto detecting what driver do i need any ideas?
[09:22] <twb> goddard: last time I did it, it just used the generic scsi driver -- the same as a SATA disk
[09:24] <DigitalFlux> Hi Guys
[09:24] <DigitalFlux> Anybody aware of a method to do a non-interactive execution of do-release-upgrade ?
[09:25] <twb> pass DEBIAN_FRONTEND=noninteractive to it?
[09:26] <goddard> twb that seems like it almost works but complains about some size problem
[09:26] <goddard> im using the scsi_debug driver
[09:26] <twb> DigitalFlux: note that if you're doing an actual upgrade from one release to another, you really really should read those messages.
[09:26] <DigitalFlux> twb: did that, didn't work
[09:27] <DigitalFlux> twb: I understand, i am just stuck with an old image template and need to automate it's upgrades
[09:27] <twb> Nod
[09:28] <twb> I don't support d-r-u because I don't like it, but if I tell you not to use it I'll get yelled at
[09:28] <twb> *I don't like or use it
[09:28] <goddard> ive never manually configured my drive before
[09:28] <goddard> I guess 8.4 gb is to small to auto partition
[09:29] <twb> goddard: I dunno.
[09:29] <twb> goddard: just allocate a single huge root filesystem, and no swap
[09:29] <twb> goddard: (I'm assuming this is a router or so, and you have more than enough ram.)
[09:29] <halvors> RoyK: I don't got that Samba server work last day, can you help me a bit more?
[09:30] <pit32> wont dance ?
[09:34] <goddard> twb i would although there isnt really an option for it i see raid lvm and iscsi
[09:34] <goddard> this isnt right
[09:34] <twb> goddard: choose "manual partitioning", silly
[09:36] <goddard> twb i go detect disks and it only shows me an 8.4mb thing and not my 12gb compact flash
[09:36] <goddard> and since i picked a driver it wont let me go back
[09:36] <goddard> twb this is pointless
[09:36] <twb> Oh 8.4 MEGAbytes
[09:37] <goddard> yeah but that isnt right
[09:37] <twb> I thought it was successfully detecting an 8GiB CF card
[09:37] <twb> Is the CF reader on the motherboard?
[09:37] <goddard> me to till i read mb haha
[09:38] <goddard> this is a little thin client
[09:38] <twb> If I were you I'd roll an image using live-build and/or LTSP
[09:38] <twb> Then just dd it on
[09:42] <goddard> twb well it was interesting more trouble than it is worth though
[09:43] <goddard> atleast i learned how to setup pxe haha
[10:43] <RoyK> halvors: sure, but I didn't quite understand _what_ that was that didn't work, and then you left
[11:34] <rtyuio> hello there
[11:34] <rtyuio> how to install complete java package ?
[11:36] <shadow42085> hey I am running a Ubuntu Server 10.04 with LAMP+Sendmail+Webmin I need to config sendmail woth webmin any ideas
[11:55] <airtonix> shadow42085: pretty sure webmin is not supported on ubuntu anymore
[11:57] <shadow42085> well what is recommened free Control Panel to use with LAMP+Sendmail or LAMP+Postfix
[12:00] <Daviey> shadow42085, The nearest to being recommend is ebox (now called Zentyal)... but it's still NOT recommended.
[12:00] <Daviey> This is an area of the server that isn't actively worked on as much as some would hope.
[12:02] <shadow42085> sheesh I am setting up a complete personal site except for domain with I use no-ip.com with DDNS
[12:02] <shadow42085> all I need is a control panel and a mail server
[12:03] <shadow42085> and the control panel sets it up I hate using command line and donlt want to use a GUI on the server
[12:06] <shadow42085> I could always use win server 03 but it's to old
[12:08] <Daviey> shadow42085, why not try the command line?
[12:08] <shadow42085> well if I try to config something in command I might mess up a config with
[12:09] <shadow42085> file with it**
[12:09] <Daviey> shadow42085, oh sure... but could you not do the same thing in a GUI?
[12:09] <Daviey> shadow42085, sudo cp /etc/some-config /etc/some-config.bak ... then you have a backup.
[12:10] <Daviey> if all else fails, purge the package and reinstall that package... :)
[12:10] <shadow42085> thats true but I want to be able to have access to all my server programs not other junk
[12:10] <TeTeT> I'd recommend etckeeper if you want to keep a trace of your chanegs
[12:11] <ttx> Daviey: looks like we won' tmeet i nBrussels
[12:11] <ttx> I need to go home, feeling worse
[12:11] <Daviey> ttx, Oh no!
[12:11] <ttx> Tell James
[12:11] <Daviey> ttx, Will do.. really sad to hear that...
[12:11] <ttx> I just shouldn't have come
[12:11] <ttx> That was a bad bad idea
[12:11] <Daviey> ttx, I didn't know you were poorly.
[12:12] <ttx> well, soren is staying, so you'll be able to ask him about how I was in details
[12:12] <Daviey> ttx, Hmm.. my main concern is who is going to buy the beer?
[12:12] <shadow42085> I have been looking at ehcp any ideas
[12:12] <Daviey> ttx, Ah great.. Well yes, don't overdo it and feel better for next $event.
[12:13] <ttx> ack
[12:14] <Sheepherd> hey guys i want to set up my own private server with ubuntu. does it matter what hardware i get? the plan is a "HP ProLiant N36L"
[12:14] <Sheepherd> are there any ubuntu-recommended servers?
[12:16] <Daviey> Sheepherd, you will probably be fine with anything... HP is pretty well supported... for more info on certified hardware: http://www.ubuntu.com/certification/
[12:18] <shadow42085> Sheepherd linux in general is fine with almost anything
[12:19] <shadow42085> what about webcp?
[12:20] <shadow42085> I have to use free ones other wise I would get either zpanel for cpanel
[12:20] <shadow42085> or**
[12:21] <shadow42085> I have also found virtualmin
[12:21] <halvors> I am trying to setup Samba
[12:22] <halvors> My goal is to login using unix users
[12:22] <halvors> And get access to users /home directory.
[12:22] <halvors> But i can't get it, so can anyone her look at my conf file?
[12:22] <shadow42085> ok u can sftp most of the time
[12:24] <halvors> Here is my /etc/samba/smb.conf file: http://pastebin.com/jF7j4UxE
[12:33] <halvors> Can anyone help me?
[13:08]  * \sh swears at upstart
[13:11] <\sh> anyone awake who is running lucid with some portchannel bonds?
[13:11] <Pici> yes. no.
[13:12] <\sh> Pici: yes, you are awake and no you don't have portchannel bonds? ;)
[13:12] <Pici> \sh: Correct ;)
[13:12] <\sh> Pici: hehe...that's bad for me...I need some testers
[13:27] <patdk-wk> portchannel bonds?
[13:27] <patdk-wk> I have many types of bonds, but never heard of portchannel
[13:29] <patdk-wk> mainly use active-backup, 802.3ad, etherchannel
[13:29] <gobbe> i think this means just basic etherchannel/port-channel
[13:29] <patdk-wk> so in linux, balance-xor :)
[13:39] <\sh> patdk-wk: etherchannel...(in cisco speak, on linux it's normally named portchannel, but anyways)
[13:39] <\sh> patdk-wk: the problem is described here: http://www.shermann.name/2011/02/ubuntu-1004-lts-portchannel-bonds.html
[13:40] <\sh> patdk-wk: on linux I'm running bond mode 4 (lacp + xmit_hash_policy layer3+4, to be cisco compatible)
[13:40] <patdk-wk> you do know that mode 4 != etherchannel
[13:41] <patdk-wk> mode 2 = etherchannel
[13:42] <Daviey> \sh, I assume you have "start on (local-filesystems and net-device-up IFACE=eth0)" in your upstart job?
[13:42]  * patdk-wk gets annoyed with etherchannel though, normally use lacp or nothing at all
[13:43] <\sh> patdk-wk: mode 4 == lacp
[13:43] <\sh> patdk-wk: btw...I upadted the post with a picture of a bond setup which worked out of the box on Jaunty
[13:44] <Daviey> \sh, If so, and that still doesn't work.. you might need to add some sleep magic... :(.... perhaps James Hunt knows..
[13:44] <\sh> Daviey: that's not possible actually, because when you have more then one interface you need to wait for all of them, then it's much better to tweak networking.conf to serve your needs, but there needs to be a default solution, because it worked in the past (without upstart), why not now (with upstart)
[13:45] <Daviey> \sh, I imagine it's an ordering issue :-)
[13:45] <Daviey> \sh, If you restart the upstart job; does it work?
[13:46] <\sh> Daviey: yes..it's just not working after boot
[13:46] <Daviey> \sh, You might need a loop, sleep, check stanza :(
[13:49] <patdk-wk> hmm, I haven't layered any bonds yet on lucid
[13:49] <patdk-wk> looks like that is the real issue
[13:49] <\sh> patdk-wk: jaunty => perfect / lucid => fail...and it's not the bonding driver which fails
[13:50] <Daviey> \sh, Something we do on Eucalyptus to stop something happening too fast is, http://pb.daviey.com/nkiN/ ... Not Ideal!
[13:50] <patdk-wk> what if you do a postup to bond bond0/1 to bond2?
[13:50] <\sh> patdk-wk: as said, after boot, bringing down all bond interfaces and bringing them up properly, the setup works
[13:51] <patdk-wk> ya
[13:52] <patdk-wk> I mean, kill the bond-slaves in bond2
[13:52] <patdk-wk> and have bond0 and bond1 add themselfs to bond2
[13:55] <\sh> patdk-wk: I tried, and the boot just stops
[13:55] <halvors> (kl. 13.21 +0100) halvors: I am trying to setup Samba
[13:55] <halvors> (kl. 13.22 +0100) halvors: My goal is to login using unix users
[13:55] <halvors> (kl. 13.22 +0100) halvors: And get access to users /home directory.
[13:55] <halvors> (kl. 13.22 +0100) halvors: But i can't get it, so can anyone her look at my conf file?
[13:56] <halvors> (kl. 13.24 +0100) halvors: Here is my /etc/samba/smb.conf file: http://pastebin.com/jF7j4UxE
[13:56] <gobbe> halvors: afaik samba needs separate passwords for users, you cannot use linux-users directly
[13:58] <patdk-wk> \sh I'll play with it sometime :) just don't have time till nextweek probably
[13:58] <halvors> gobbe: But i can't access homes
[13:59] <halvors> Does not samba ot libpam-smbpass
[13:59] <halvors> syncronize it?
[14:07] <RoAkSoAx> morning all
[14:07] <Daviey> RoAkSoAx, !
[14:07] <Daviey> How is it going?
[14:08] <RoAkSoAx> Daviey: pretty good, how about yourself
[14:09] <jpds> Daviey: You're not Cisco.
[14:13] <Daviey> jpds, am i not? :o
[14:13] <Daviey> RoAkSoAx, Good thanks.
[14:16] <panfist> my server has some nfs shares that are failing to mount on boot, is there any way to proceed booting anyway? it seems like it's just stuck
[14:19] <thomas_s> panfist: use _netdev in the mount options (after "default" in /etc/fstab). That will make sure they're ignored until the netfs service runs
[14:20] <thomas_s> s/default/defaults/ - sigh
[14:21] <panfist> thomas_s: i'm having trouble just getting a terminal prompt. it goes straight from the bios screen to the mount.nfs errors. i thought there was a key to press to go to the grub menu or something, but nothing is working
[14:22] <thomas_s> panfist: oic. you've got the grub timeout set to 0 probably.
[14:22] <thomas_s> But the failure to mount the fs' (unless it's the root you're nfs mounting?) should drop you to single user mode
[14:23] <thomas_s> You may need to bring out the ol' bootable cd
[14:27] <alvin> thomas_s, panfist: While the man page will tell you otherwise, _netdev will not help you. See bug 384347
[14:27] <panfist> sorry, i just crashed, i tried to send ctrl-alt-f3 to the server, it's running in a virtual machine, but the command was sent to the host machine instead and unfortunately trying to switch back to ctrl-alt-f7 causes my host's kernel to freeze
[14:28] <alvin> I have a few clients that don't mount their NFS shares automatically and I mount them after boot with mount -a -t nfs (or nfs4)
[14:48] <zul> some upstreams should be dead
[15:17] <panfist> is there any way to configure my server so that it starts listening for ssh connections before it tries mounting nfs shares?
[15:20] <\sh> patdk-wk: it's became more complicated ... I can see in syslog that the active bond slave changed from bond0 to bond1 but the mac address of bond1 is not taken by bond2 ;)...strange
[15:26] <zul> SpamapS: ping did you test the mysql stuff first?
[15:53] <Roasted> hola
[15:53] <Roasted> anybody have experience with setting up freeradius on linux?
[15:58] <zul> SpamapS: never mind i suck
[16:14] <sushi_> I accidentally removed one files from nginx. I reinstalled it but I don't see that file even after reinstallation. Any idea why?
[16:18] <SpamapS> zul: I will have to disagree with you there. You stink a bit, but you don't suck. :)
[16:18] <SpamapS> sushi_: is it in /etc ?
[16:19] <sushi_> SpamapS: I think I might need to apt-get purge.
[16:19] <sushi_> Yeah the file was in /etc/nginx/sites-available/default
[16:19] <SpamapS> sushi_: no
[16:19] <SpamapS> well you can --purge..
[16:19] <SpamapS> or dpkg -i file.deb --force-confmiss
[16:21] <sushi_> Yeah I can see the file after clean reinstall. THis django deployment is just killing me.
[16:21] <SpamapS> sushi_: removing a conffile is considered the same as editting it.. dpkg preserves it
[16:21] <SpamapS> sushi_: so --force-confmiss will restore it
[16:22] <azertyu> hello, why java not available on repository ?
[16:22] <sushi_> SpamapS: Thanks. TIL moment.
[16:22] <azertyu> i need to need to remote control 2 bladesystem
[16:22] <azertyu> i got a serious problem with java
[16:22] <SpamapS> !confmiss
[16:22] <SpamapS> that might be a good thing to have in the bot
[16:30] <azertyu> lo
[16:30] <azertyu> any one there ?
[16:33] <sushi_> plenty people here
[16:36] <azertyu> do you see my question ?
[16:46] <Tophat> ive got an opsview server setup to send email alerts, but now my sending SMTP acct is getting flooded with delivery failure messages -- ive uninstalled exim4 and still getting email failures, any way to stop the emails?
[16:54] <RoAkSoAx> zul: have a sec? I'm trying to package fence-agents, and I keep getting this error, in both the debian package, and compiling the source... any ideaS? http://pastebin.ubuntu.com/562065/
[16:55] <zul> looking
[16:55] <zul> RoAkSoAx: bad xml tag maybe?
[16:55] <RoAkSoAx> zul: that's the thing, the file is empty
[16:56] <RoAkSoAx> zul: i can't seem to figure it out, gonna have to try to compile it in maverick and see what happens
[16:56] <zul> fence2man.xsl is empty?
[16:57] <RoAkSoAx> zul: "python fence_alom" -> fence_alom is
[17:00] <RoAkSoAx> zul: this command ends up with empty output "python fence_alom ../../../fence/agents/lib/fence2man.xsl -o metadata"
[17:00] <RoAkSoAx> where fence_alom is created during build
[17:00] <RoAkSoAx> but is also empty
[17:01] <RoAkSoAx> maybe it is a bug in the makefile itself
[17:02] <zul> probably
[17:02] <zul> SpamapS: so the mysql source stuff
[17:08] <SpamapS> zul: yes?
[17:08] <zul> SpamapS: so i dont like it but is a necessary evil so i uploaded it
[17:12] <zul> SpamapS: so finish up the plugin and ill upload it as well
[17:12] <zul> and review it
[17:12] <SpamapS> zul: COOL.. ok n/m you don't stink. ;)
[17:13] <zul> SpamapS: heh
[17:27] <RoAkSoAx> zul: seems to be something wrong with python case all other that are not python based, they build ok
[17:29] <geekbri> if i only want to send outgoing mail from a server running ubuntu 10.04 what would be the best to install, exim ?
[18:07] <zul> SpamapS: can you look at https://bugs.launchpad.net/ubuntu/+source/nis/+bug/569757 its upstartish :)
[18:12] <SpamapS> zul: no sorry we fixed all the bugs in upstart that is user error ;)
[18:13] <zul> SpamapS: heh you just dont want to touch nis
[18:14] <SpamapS> DING DING DING
[18:14] <SpamapS> bingo
[18:14] <SpamapS> I used NIS successfully to deploy a roaming /home solution for a middle school computer lab once.
[18:14] <SpamapS> *never again*
[18:17] <zul> SpamapS: yes those were the days
[18:27] <RoAkSoAx> lol
[19:07] <airtonix> 4 yorkshiremen go!
[19:17] <gurkan_> hi all
[19:17] <gurkan_>  how to set the smtp server of my provider in my home's mailrc file i use mailutils 2.1
[19:17] <gurkan_> i saw the manual but not enough clear
[19:17] <gurkan_> someone have a mailrc sample
[19:30] <gnuyoga> gurkan_: what are u trying to do ?
[19:30] <gnuyoga> more details please
[19:32] <gurkan_> i try to set the smtp server of my provider in my home's mailrc file , it's for sending mail
[19:32] <gurkan_> i try to send mail myself
[19:33] <Roasted> anybody here familiar with freeradius? I keep going in loops with the authentication process.
[19:34] <gurkan_> i want to implement the line smtp://mysmtp.server.stuff in mailrc , i messed up the config
[19:43] <Sorinan> I jsut installed ubuntu server, but I'm getting an error when starting the system: modprobe: FATAL: Could not local /lib/modules/2.6.35-22-server/modules.dep
[19:44] <Sorinan> Altought the modules.dep exists, I already tried running depmod and regenerationg initramfs. Any ideas?
[19:46] <Sorinan> *Could not load /lib...
[20:17] <LaRockus> Has anyone here experienced an Ubuntu 9.4 Apache web server freezing up when the sendmail minute cron runs? I've looked all over but haven't been able to find any other instances of the same thing happening on the forums.
[20:33] <fluvvell> LaRockus, being a server, any reason it hasn't been upgraded to 10.04 LTS?
[20:37] <LaRockus> compatibility issues with our latest software release hasn't been through the QA process completely on a 10.4. I wish it had though, would love to get it upgraded.
[20:39] <LaRockus> I have 15 servers with the exact same configuration, only this one is giving me issue.
[20:40] <smoser> SpamapS, what woudl you do with: https://bugs.launchpad.net/ubuntu/+source/mysql-dfsg-5.1/+bug/711635
[20:40] <SpamapS> smoser: reading
[20:40] <thesheff17> LaRockus: anything in the logs...are the server patches up to date?
[20:41] <LaRockus> patches are up to date. Which logs specifically?
[20:42] <SpamapS> smoser: interesting.. I'd look at daemon.log .. is upstart trying to respawn mysqld?
[20:42] <smoser> well, the config probably tried to restart it
[20:42] <smoser> but its never going to start
[20:42] <thesheff17> LaRockus: usually freezes are kernel panics....have you been updating the kernels?
[20:42] <smoser> because it can't read /var/lib/mysql
[20:42] <smoser> :)
[20:42] <SpamapS> smoser: it shouldn't "never start" tho.. it should just error out
[20:42] <smoser> well, it is just hanging there.
[20:43] <SpamapS> Right, start mysql is waiting for something...
[20:43] <glycoknob> I have a question regarding ufw - I block outgoing connections but there are connections from port 80 to random higher ports (and random ips) that are blocked. apache should not doing this or?
[20:43] <SpamapS> smoser:     /lib/init/apparmor-profile-load usr.sbin.mysqld
[20:43] <SpamapS> smoser: is that running?
[20:44] <Roasted> anybody set up freeradius before? I'm having uathentication issues that I can't seem to track down
[20:45] <LaRockus> thesheff17, the install is less than a week old, unless there has been a huge kernel update in the last week should be good. the site actually stays up, and sendmail still sends when manually ran, its just extremely slow though, and does not run on through the cron
[20:45] <SpamapS> LaRockus: can you manually run the same command as the cron job runs?
[20:46] <smoser> SpamapS, http://paste.ubuntu.com/562208/
[20:46] <LaRockus> yeah, it takes about 10 minutes for each message to process.
[20:48] <thesheff17> LaRockus: maybe a hardware issue or a networking issues....how long does it take on the other servers? is it instant?
[20:49] <SpamapS> smoser: ahh!
[20:49] <SpamapS> smoser: its spinning forever in the post-start
[20:49] <SpamapS> smoser: or rather, it should only be spinning for 30 seconds
[20:50] <smoser> SpamapS, i wouldnt think its in post-start
[20:50] <SpamapS> smoser: status mysql
[20:50] <smoser> as there is no mysql process (in that pastebin)
[20:50] <SpamapS> smoser: right, because it exitted
[20:50] <smoser> mysql respawn/post-start, (post-start) process 24511
[20:50] <SpamapS> but post-start spins trying to ping the server
[20:50] <SpamapS> yep
[20:50] <SpamapS> should fail after 30 seconds
[20:51] <LaRockus> the transmission itself is almost instant, in and out through the mail queue in almost no time.
[20:52] <SpamapS> smoser: it should actually report the failure.. just that it will take 30 seconds :-P
[20:52] <thesheff17> LaRockus: sounds like a routing issue...I would check dns, network config, ping times, tracepath maybe even a different nic on the server.
[20:53]  * RoAkSoAx spent hours and hours trying to figure out what was wrong with the packaging/compilation of fence-agents to end up discovring that the cause of the failure was becuase of dash
[20:53] <LaRockus> with a routing issue would i experience any issues with the page displaying.
[20:54] <thesheff17> LaRockus: with apache?  not sure I understand the question ...displaying
[20:54] <smoser> SpamapS, yeah, its not failing after 30 seconds.
[20:54] <SpamapS> smoser: hmm.. logic error maybe
[20:56] <SpamapS> smoser: can you strace that sh process?
[20:56] <SpamapS> You actually might not be able to.. not sure if upstart ptrace's its post-start shells
[20:56] <SpamapS> I think not tho
[20:57] <LaRockus> I would expect to see issues with the web page displaying if there were physical issues with my nic, or with the DNS routing, or is my thinking wrong.
[20:58] <thesheff17> LaRockus: I admit I haven't used 9.04 much but when I did I saw it terribly slow on ssh for some reason even though apache worked fine....another thing you can do is just try 10.04 and see if you see the same problem and then you know it is software based.
[20:59] <smoser> SpamapS, upstart is re-execing that
[20:59] <LaRockus> I might have to try 10.04, I'm just reluctant with it being on a live customer. Thanks for the tips, im looking into those now.
[20:59] <smoser> i added a sh -x and redirected the post-start output to a file and i can watch it get called again after counting to 30
[20:59] <SpamapS> smoser: HMM
[21:00] <SpamapS> smoser: I wonder if that has to do with the respawning
[21:00] <smoser> respawn
[21:00] <smoser> yeah
[21:00] <smoser> it makes sense.
[21:00] <smoser> the job failed
[21:00] <smoser> upstarts trying to restart it
[21:00] <SpamapS> smoser: so maybe we need a respawn limit
[21:00] <thesheff17> LaRockus: there is just allot of factors from hardware, to networking config, to software...and only slowing eliminating them will help you find the problem.
[21:00] <SpamapS> I think the normal limit is like, 3 times in 5 seconds
[21:00] <SpamapS> But if we're delaying it by 30 seconds....
[21:01] <smoser> hm... can you exit from a pre-start in a way that says "don't bother"
[21:01] <SpamapS> actually we should probably change the goal back to stop
[21:01] <SpamapS> yes,   stop ; exit 0
[21:01] <smoser> why not stop exit 1
[21:01] <SpamapS> oh either is ok
[21:02] <SpamapS> 1 will report the problem to the user so yeah thats better
[21:02] <SpamapS> 0 is for if its been disabled
[21:05] <SpamapS> smoser: ultimately, we don't want to keep respawning over and over.
[21:05] <SpamapS> smoser: so I think the fix is  respawn limit 3 95 .. which would mean it respawns 3 times in a row waiting 30 seconds between (and give it 5 seconds for churn)
[21:06] <SpamapS> smoser: will post in the bug..
[21:06] <smoser> i dont know.
[21:06] <smoser> you migh tconsistently miss the 3 restarts in 95 seconds
[21:06] <smoser> due to load
[21:06] <smoser> and still hang there
[21:07] <SpamapS> I don't see a way to know before starting mysqld, that mysqld is going to fail.. but we can probably confirm that its not starting correctly if it fails 3 times in a row.
[21:07] <SpamapS> the exec is instantaneous..
[21:07] <smoser> i think if the post-start doesn't ever ping after 30 seconds, i think 'stop; exit 1;' is proper
[21:07] <SpamapS> the mysqladmin is a unix socket only..
[21:07] <smoser> nothing is instantaneous
[21:07] <SpamapS> so the thing would have to be *destroyed* with load
[21:08] <smoser> its also running http://pastebin.com/ZgPrG9AQ
[21:08] <SpamapS> on a respawn, the exec will wait for almost nothing
[21:08] <SpamapS> the binary is already cached..
[21:09] <smoser> http://pastebin.com/PSCz4HpN is /usr/share/mysql/debian-start.inc.sh
[21:09] <smoser> forks left and right in that script.
[21:09]  * SpamapS just loves waiting 30 seconds for pastebin's ads
[21:10] <smoser> all sorts of things could make this take longer than 1.033 seconds to run a full post-start
[21:10] <smoser> anyway...
[21:10] <smoser> if the thing never pings inside 30 seconds, the idea is its dead.
[21:10] <smoser> respawning is pointless.
[21:10] <SpamapS> smoser: debian-start is after everything is ok
[21:10] <smoser> true.
[21:11] <SpamapS> 120 may be in order though.
[21:11] <smoser> so, i dont know. but why would you bother.
[21:11] <SpamapS> just .. because why not?
[21:11] <smoser> it is seriously not going to come up
[21:11] <smoser> the whole point of the pre-start is to do this delay and wait for you
[21:11] <SpamapS> I'm more worried that we're only waiting 30 seconds
[21:11] <smoser> what value is there in doing it 4 times.
[21:11] <SpamapS> Have had mysql servers wait *hours*
[21:12] <SpamapS> innodb recovery is a bitch
[21:12] <smoser> hm..
[21:12] <smoser> i dlon t knwo.
[21:13] <smoser> if post-start exits false, then upstart respawns ?
[21:13] <SpamapS> so you're thinking only try and respawn once?
[21:13] <SpamapS> no
[21:13] <smoser> does it ?
[21:13] <SpamapS> upstart respawns if the main process exits != 0
[21:13] <smoser> so the loop is caused by mysql actually dying
[21:13] <smoser> yeah.
[21:14] <SpamapS> if the process is up, but the post-start exits 1 .. I think its just going to fail the start event
[21:14] <SpamapS> but the state won't be changed
[21:14]  * SpamapS tests this thory
[21:14] <SpamapS> theory too
[21:16] <SpamapS> actually exit 1 in post-start does nothing
[21:17] <SpamapS> Feb  3 13:16:07 lucid init: test-pstart post-start process (1279) terminated with status 1
[21:17] <SpamapS> thats about all you get
[21:18] <SpamapS> So really all you're doing is delaying or cancelling the started event
[21:19] <SpamapS> (running stop in post-start will kill the main process and there won't be a started event)
[21:20] <azertyu> hi there
[21:20] <azertyu> is there anyone here
[21:20] <SpamapS> azertyu: yes!
[21:21] <azertyu> well i got a vps server
[21:21] <azertyu> i forgot to backup data
[21:21] <azertyu> now someone hack my server
[21:22] <azertyu> root password gone
[21:22] <markpof> uh oh
[21:22] <azertyu> my question is : is it possible to sync vps data to my home pc ?
[21:22] <SpamapS> azertyu: I'm sure you can get a massage and some tylenol...
[21:23] <SpamapS> azertyu: sure. just rsync the data.
[21:23] <azertyu> i got2T Þ of HDD at home
[21:24] <smoser> azertyu, you can certainly sync anything your user can get to, assuming you can get to the vps system.
[21:25] <azertyu> rsync is it a virtual raid ?
[21:26] <SpamapS> azertyu: no it just copies things very quickly
[21:26] <SpamapS> azertyu: if you want the data copied in real time that will be much harder
[21:28] <azertyu> what do you mean by quickly copies ?
[21:28] <azertyu> is that copies just the data
[21:28] <azertyu> ?
[21:29] <azertyu> first i want to know virtual raid is that existing ?
[21:32] <SpamapS> smoser: fyi the default is not to stop respawning unless it tries *10 times* in 5 seconds
[21:32] <SpamapS> smoser: doh
[21:32] <smoser> i suspect that someone just added exim4 to the list of packages supervised by ubuntu-server
[21:32] <smoser> http://ubuntu-server-new-bugs.notlong.com/
[21:33] <azertyu> well, what i want to kow is :  i want to synceverything on my home pc  from the vps in case if i change my vps  i need every to reconstruct
[21:33] <azertyu> is it possible ?
[21:33] <azertyu> rebuild
[21:33] <SpamapS> azertyu: you should look into etckeeper too..
[21:34] <smoser> azertyu, you could very likely do something like a 'ssh root@VPS dd /dev/sda > my.disk.img'
[21:34] <smoser> but that is going to take quite a while
[21:35] <SpamapS> hah.. and may not be usable on a different disk.
[21:35] <smoser> you're would be much better off just getting the filesystem contents, and posibly the partition table and MBR or something.
[21:36] <smoser> rsync -avz root@VPS:/ ./my-vps-backup
[21:36] <SpamapS> yeah,   ssh root@VPS tar -czf - > backup.tar.gz also works
[21:36] <SpamapS> or you could do what I do.. tar my boxes up and push the tarball into S3
[21:36] <smoser> the rsync would be interupptable would be the big reason to use that.
[21:36] <smoser> if you connection died it would resume well.
[21:36] <markpof> is the dd output conform to the .img format (if there even is one?)
[21:37] <smoser> there is no ".img" format really.
[21:37] <markpof> ah
[21:37] <smoser> but basically, it would work with kvm as a "raw" disk file
[21:37] <smoser> (qemu-img create -f raw my-disk.img)
[21:37] <markpof> raw
[21:37] <markpof> i see
[21:40] <azertyu> there is no result for etchkeeper
[21:40] <azertyu> !
[21:40] <SpamapS> azertyu: etckeeper
[21:40] <azertyu> yes
[21:41] <azertyu> ok
[21:41] <azertyu> find it
[21:41] <azertyu> cfengine is for what ?
[21:42] <azertyu> anyone know that package ?
[21:42] <SpamapS> cfengine is  more for managing multiple machines
[21:43] <azertyu> be explicit
[21:43] <SpamapS> cfengine more for managing multiple machines, $#@!
[21:44] <azertyu> what kind of tasks ?
[21:44] <azertyu> managing what kind of tasks ?
[21:47] <SpamapS> smoser: the plot thickens .. I think the default respawn policy may cause something weird. *ANY* respawn limit clause seems to fix this problem.
[21:47] <SpamapS> smoser: including 'respawn limit unlimited'
[21:48] <SpamapS> heh.. which is an undocumented feature.. fun
[21:49] <smoser> really.
[21:49] <smoser> thats interesting.
[21:50] <SpamapS> yeah I'm trying some crazy limits and it works as I'd expect.. failed to start job
[21:50] <smoser> so the pre-start isn't counted in the respawn i guess ?
[21:51] <SpamapS> post-start..
[21:51] <SpamapS> not sure
[21:53] <SpamapS> smoser: so there's something that gets inexplicably changed just by having a limit mentioned..
[21:53] <SpamapS> smoser: the code defines the defaults as a limit of 10, interval 5 seconds
[21:53] <SpamapS> smoser: but if I say   respawn limit 10 5 ... the job fails as I'd expect
[21:54] <SpamapS> smoser: likewise a limit of 1 100 .. or 100000 10000000
[21:54] <SpamapS> I think this one has to go upstream
[22:07] <mrothhh> were is the link to the hypervisor for zen bearmetal
[22:21] <SpamapS> smoser: commented. The workaround/fix for mysql seems simple.. I'm betting the fix in upstart is not.
[22:37] <SpamapS> zul: when can we kick nis out of main? ;)
[22:50] <MsBetty> What do I need to have a file and print server?
[22:51] <MsBetty> Apache and CUPS?
[22:51] <MsBetty> samba and cups?
[22:58] <qman__> MsBetty, that depends on your clients
[22:59] <MsBetty> qman__, two laptops on a home network
[22:59] <qman__> CUPS is needed if the printers are directly attached to the server regardless
[22:59] <qman__> if you have windows clients, you need samba
[22:59] <qman__> if you have linux clients, you have options
[23:00] <MsBetty> qman__, 1 laptop dual boots windows 7/#! and one is Linux Mint 10. An old desktop is getting reused for ubuntu server.
[23:01] <qman__> linux clients can connect directly to CUPS
[23:01] <qman__> but that's only for the printers
[23:01] <qman__> you'd still need a file server
[23:01] <qman__> linux clients can connect to samba file servers, so in a mixed network, samba is the most convenient
[23:02] <MsBetty> so I really just need to configure samba and cups properly. and make sure the clients have the appropriate packages?
[23:02] <qman__> that would work
[23:04] <MsBetty> qman__, thank you for the input. I'll go read the manuals and hopefully everything will go well.
[23:04] <qman__> no problem, good luck
[23:09] <airtonix> yo, i see that bug too when i connect to samba shares in nautilus
[23:10] <airtonix> !bug 712787