[00:22] <kirkland> hallyn: "all but the XXX hunks" ... that's hard to read;  just send me a diff or a branch :-)
[00:29] <axisys> ivoks: do you have a same isolinux.cfg file that works... I cannot seem to make ubuntu installation read the preseed file ..
[00:29] <axisys> ivoks: posted my question here
[00:29] <axisys> ivoks: http://comments.gmane.org/gmane.linux.ubuntu.server/5034
[00:39] <nertil> is it posible to play games in server?
[00:40] <Pici> nertil: sure, the bsdgames package has a bunch of cli games, and nethack is always fun.
[00:41] <Pici> !info moon-buggy
[00:41] <pmatulis_> related to 'moon patrol'?
[00:42] <Pici> If I'm remembering correctly, its similar.
[00:42] <pmatulis_> i remember the arcade game
[01:35] <imthenachoman> hey guys. any recomendations on which firewall to use. I was thinking Shorewall?
[01:36] <twb> There is only one firewall: netfilter.
[01:37] <twb> Everything else is a (theoretically) "user friendly" layer on top of netfilter.
[01:37] <imthenachoman> okay. i see. then which "user friendly" layer do y'all recommend?
[01:37] <twb> If you don't want to learn, and you have simple needs, try ufw.  If you have complex needs AND you prefer to learn a wrapper instead of how to do it yourself, you can try shorewall.
[01:38] <twb> IMO shorewall only pays off if you have a whole machine room and a big bag of exceptional cases.
[01:39] <imthenachoman> right now i am learning in a virtualbox ubuntu server install. eventually, i am going to get a VPS hosted webserver. so i want to learn something i can use for those
[01:40] <twb> For my network, with seven networks, about thirty hosts and a single router, I just use iptables-restore directly (on the router); the ruleset is only 86 lines long.  It would be substantially longer in shorewall.
[01:41] <twb> Because netfilter has simple sequences of arbitrary rules, but shorewall's rules are rigidly structured into tables.
[01:42] <imthenachoman> well i am thinking if i am using a vps hosted service, then i might not need firewall
[01:42] <imthenachoman> so i dont know...not gonna learn it if don't need it
[01:57] <shauno> be worth checking with your host on that one. most do recommend a firewall on a vps
[02:06] <EvilPhoenix> you definitely need a firewall on a VPS
[02:06] <EvilPhoenix> most VPS hosts put an internet-facing IP to the VPS
[02:07] <EvilPhoenix> and thus its visible to the entire internet and thus vulnerable to attacks
[02:30] <hallyn> kirkland: why did you take the '-s' option out of debian/rules?  it won't matter for arm builds, but just wondering why
[02:31] <kirkland> hallyn: i don't think i did;  probably in slangasek's change?
[02:32] <fakhir> is there a recommended bandwidth graph application? mostly concerned with HTTP traffic.
[02:32] <hallyn> kirkland: ah. hm.
[02:51] <hallyn> kirkland: http://launchpadlibrarian.net/64430687/debdiff  should fix the arm build failure
[03:34] <yellowblue> Im true gangsta
[03:34] <yellowblue> !ops
[04:33] <kirkland> hallyn: uploading/sponsoring now
[04:33] <kirkland> hallyn: done
[05:19] <_1094kms> I just did an 'iptables -F' and ubuntu 10.04 doesn't respond anymore - that should't happen right?
[05:20] <twb> _1094kms: wrong.
[05:21] <twb> _1094kms: that is by design, and your lesson today is "do not use iptables(8); use iptables-restore(8) (or iptables-apply)".
[05:23] <_1094kms> I guess that is a lesson - I don't have remote access :-) Does it shut down the iface?
[05:24] <twb> No, it flushes all the rules.
[05:24] <_1094kms> so why is it completely dead now? - just did an nmap from a remote machine and it doesn't respond at all.
[05:24] <twb> In a "default deny" environment (which is the only workable way to run a firewall), that will result in all packets being dropped.
[05:25] <_1094kms> strange - I tried the same command on my dev server and no problems.
[05:25] <twb> That's obviously because your dev server has a weak or nonexistent firewall
[05:26] <_1094kms> Will I have to reapply some rules manually or will a reboot suffice?
[05:26] <twb> iptables -F only affects the running system
[05:26] <_1094kms> kk - I guess the rest will have to wait till tomorrow - thx twb for that lesson :-)
[05:26] <twb> Unless you have an on-shutdown script that dumps the ruleset into a persistent area, a reboot should revert your change
[05:27] <_1094kms> I don't actually know where those rules came from. There was nothing in the if.up scripts
[05:27] <lifeless> SpamapS: hey, how is cassandra packaging atm?
[05:28] <twb> _1094kms: most likely someone has done "ufw enable".
[05:28] <twb> _1094kms: otherwise, something like shorewall or iptables-persistent might be installed.
[05:29] <_1094kms> I will check it out as soon as I get access back.
[05:36] <semanticpc> in bash script how do I redirect my output of a command inside the bash script to a file ?
[05:37] <kaushal> hi
[05:43] <SpamapS> lifeless: re cassandra packaging.. its about 60% from the archive.. 40% bundled jars..
[05:44] <SpamapS> lifeless: 0.7.1 came out yesterday... 0.7.2 should be out tomorrow to fix the huge bug introduced.
[05:44] <SpamapS> lifeless: I still have my doubts about their ability to support users long term.
[05:44] <SpamapS> lol.. status.aws.amazon.com is down
[05:45] <lifeless> SpamapS: do we have an alternative ready-to-roll for in-dc use?
[05:46] <SpamapS> lifeless: one that supports as many general use cases as Cassandra? no.
[05:47] <lifeless> whats the one that hadoop can run directly from
[05:47] <SpamapS> I think there are quite a few more focused scalable data stores out there... cassandra's power is its flexibility.
[05:47] <SpamapS> HDFS
[05:47] <SpamapS> yeah its pretty good
[05:47] <lifeless> thats right
[05:47] <SpamapS> or HBase
[05:47] <SpamapS> they're related
[05:47] <lifeless> HBase, thats the one
[05:47] <lifeless> is it in use inside canonical atm?
[05:47] <SpamapS> don't think so, but I could be wrong
[07:16] <havoc74> Good morning!
[07:18] <havoc74> I'm new to server admin and have a few questions. First I've got an IBM x235 with 2 36gb drives and 3 72gb drives. What kind of RAID setup would be best?
[07:25] <_ruben> a common setup would raid1 over the 2 26g disks and raid5 over the 3 72g disks .. tho it all depends on your needs/wishes/etc
[07:25] <_ruben> oh he'd left already :p
[07:56] <ZacLnxNewb> hi
[07:56] <ZacLnxNewb> I'm having trouble with DDclient
[07:58] <ZacLnxNewb> Can anyone suggest anything?
[07:58] <ZacLnxNewb> to use as a dynamic ip address updater?
[09:13] <huats> morning
[10:59] <lephisto_> ahm
[10:59] <lephisto_> https://bugs.launchpad.net/ubuntu/+source/lvm2/+bug/719333
[10:59] <lephisto_> can anyone please care?
[11:00] <lephisto_> .. rendering ubuntu lucid unusable for cluster usage
[11:00] <lephisto_> (except you wanna ruin the whole datacetnter)
[11:36] <lephisto_> dumdidumdidum
[11:36] <lephisto_> no one cares, really?
[11:43] <lephisto_> i'm uncertain what to do
[11:43] <lephisto_> ww
[11:50] <twb> Wait for someone to triage the ticket, I imagine
[11:54] <aliverius> $ sudo smartctl --attributes --log=selftest --quietmode=errorsonly /dev/sdb
[11:54] <aliverius> Num  Test_Description    Status                  Remaining  LifeTime(hours)  LBA_of_first_error
[11:54] <aliverius> # 1  Extended offline    Completed: read failure       40%         0         2780828360
[11:54] <aliverius> this means i have a faulty disk right?
[11:57] <patdk-lap> or a bad sector that hasn't been relocated
[12:10] <twb> patdk-lap: yes
[12:11] <twb> If a self-test fails, start shopping for a new disk
[12:11] <twb> The other error threshold stuff in smartctl is more paranoid, I tend to only look at self-test results
[12:59] <aliverius> patdk-lap: twb so it is serious or can i take other measures?
[12:59] <aliverius> lke letting the disk relocate the sector...
[12:59] <twb> aliverius: go buy a new disk
[12:59] <aliverius> i will rma that one
[13:00] <aliverius> for the second time
[13:00] <aliverius> i was going to rebuid the raid 1
[13:01] <aliverius> i havent done it before so at least let me educate myself before sending it back!
[13:02] <patdk-wk_> aliverius, you can attempt to make the disk relocate it, by writeing to that sector
[13:03] <patdk-wk_> in my experiance though, it doesn't work well, and just rma it
[13:03] <aliverius> i will rma it
[13:03] <aliverius> afterall i forced it to work:
[13:04] <aliverius> when i first connected it it was doing some spinning noises and the system couldnot detect it at boot
[13:04] <aliverius> so i gave it a little kick
[13:04] <aliverius> and that fixed the problem:P
[13:06] <aliverius> if either the bad sector was created by me or it was there already, i will rma it
[13:07] <aliverius> i just did the smart test in case i could avoid it
[13:09] <patdk-wk> hmm, you can't create bad sectors :)
[13:09] <twb> patdk-wk: open the drive up in a dusty workshop
[13:09] <patdk-wk> twb, in normal usage :)
[13:09] <patdk-wk> sure, a hammer always works :)
[13:09] <twb> shake it, then
[13:09] <patdk-wk> twb, normal usage by people >5years old :)
[13:11] <twb> Hmm, why is my custom linux-image .deb 438MiB wide?
[13:12]  * patdk-wk always perferred his binarys long, vs wide
[13:12] <twb> Because dbg
[13:13] <twb> When you copy Debian's .config, you get -dbg turned on, but upstream's make deb-pkg doesn't know about it
[13:13] <twb> No wonder I needed like 12GiB
[13:18] <aliverius> can you guys help me rebuild the raid 1 array?
[13:26] <twb> Do I look psychic?
[13:27] <JamesPage> hggdh, zul: I just completed a minimal virtual install ISO test case - it all looks OK but the kernel installed is -generic-pae rather than -virtual
[13:27] <zul> for lucid or natty?
[13:29] <patdk-wk> twb, yes
[13:30] <twb> Funny, that's not what you thought.
[13:38] <JamesPage> zul: lucid
[13:38] <zul> JamesPage: weird ill try today
[14:08] <RoAkSoAx> morninh sll
[14:11] <compdoc> who is this sll person of which you speak?!
[14:16] <b0gatyr_> anyone with a trimonitor setup?
[14:19] <RoAkSoAx> kirkland: ping?
[14:24] <kirkland> RoAkSoAx: pong
[14:24] <RoAkSoAx> kirkland: Howdy!! Do you think we should ship the second stage action for powersave by FF? Anything else you'd like to see in PowerNap before that date?
[14:25] <kirkland> RoAkSoAx: hmm
[14:25] <kirkland> RoAkSoAx: i have not tested the 2nd stage yet
[14:25] <kirkland> RoAkSoAx: don't enable it by default
[14:26] <kirkland> RoAkSoAx: but if you're happy with it, then sure, ship it by FF
[14:26] <RoAkSoAx> kirkland: haven't yet implemented it :)
[14:26] <kirkland> RoAkSoAx: oh, well, don't rush it
[14:26] <kirkland> RoAkSoAx: PowerNap is looking *awesome* for Natty as it is
[14:26] <kirkland> RoAkSoAx: i thought of one thing this week... (forgotten it at the moment)
[14:27] <RoAkSoAx> kirkland: PowerNap in natty indeed rocks!!
[14:27] <RoAkSoAx> kirkland: when it comes back to your mind just let me know. Btw. where you able to figure out why things failed in your case?
[14:27] <kirkland> RoAkSoAx: it's working great now
[14:27] <kirkland> RoAkSoAx: oh, i remember
[14:27] <kirkland> RoAkSoAx: we do need a powerwake-now utility
[14:28] <kirkland> RoAkSoAx: that does a pm-powersave false
[14:28] <kirkland> RoAkSoAx: and fixes the state in powernapd
[14:28] <kirkland> RoAkSoAx: it should really just send a signal to powernapd
[14:28] <kirkland> RoAkSoAx: and let powernapd handle it
[14:28] <kirkland> RoAkSoAx: and then we need to call that in the powernap upstart script on pre-stop
[14:29] <kirkland> RoAkSoAx: so that if someone does "sudo stop powernap", it will undo the powersave state
[14:29] <kirkland> RoAkSoAx: otherwise, things get a little out of wack, if your system is powernapping, and then you kill powernapd from upstart
[14:30] <RoAkSoAx> kirkland: ok cool. Yeah I'll add the powerwake-now tool today
[14:31] <kirkland> RoAkSoAx: powerwake-now should be a *really* simple shell script, that just sends SIGUSR1 or SIGUSR2 to powernapd
[14:31] <kirkland> RoAkSoAx: powernapd needs a handler for SIGUSR2
[14:32] <RoAkSoAx> kirkland: yeah, will be exactly the same as powernap-now
[14:34] <lephisto_> re.
[14:55] <edge> I'm having trouble finding the documentation on adding things to the startup on 10.10. Does anybody have that handy? i want to start some services at boot.
[15:03] <hallyn> edge: you're looking for upstart documentation
[15:04] <hallyn> SpamapS: ^ where would you say a beginner should start looking?
[15:05] <edge> hallyn: having the name helps thanks. I didn't know what ubuntu called that system
[15:29] <RoAkSoAx> SpamapS: ping?
[15:59] <RoAkSoAx> kirkland:
[15:59] <RoAkSoAx> kirkland: http://paste.ubuntu.com/567721/
[16:00] <Frenk> Hey, I have a question. One person with a certain IP cant connect to the server. I cleared denyhosts, and iptables -L |grep IP doesnt show it
[16:00] <kirkland> RoAkSoAx: don't "exec" that
[16:01] <kirkland> RoAkSoAx: i don't think you want to "exec" that
[16:01] <kirkland> RoAkSoAx: and drop the full path
[16:01] <RoAkSoAx> kirkland: oki
[16:01] <kirkland> RoAkSoAx: just use "powerwake-now" in pre-stop
[16:01] <kirkland> powerwake-now || true
[16:03] <RoAkSoAx> kirkland: alright. Anyways, when powerwake-now is issued, then powernapd will evaluate if it is in powersave mode. If it is, it will take recover action and restart everything. if not, will do nothing
[16:03] <Frenk> Where do I have to look?
[16:08] <SpamapS> hallyn: re startup docs, upstart.ubuntu.com
[16:15] <Frenk> Hey. I have a IP adstill ipress and the person cant access services on the server. Even ping does not work. For all others it does work well. There is no entry in deny-hosts or iptables. Any suggestions?
[16:41] <RoAkSoAx> kirkland: it seems that I'll have to put an sleep 2 or so in the pre-stop part. Otherwise, the daemon will stop before actually taking recover action
[17:06] <lephisto> re.
[17:06] <hggdh> smoser: do we have an ec2 kernel for Karmic proposed?
[17:09] <smoser> hggdh, no. you want me to spin a karmic with --proposed to get kernels uploaded ?
[17:14] <hggdh> smoser: yes, please
[17:15] <Frenk> Hello,  I need your advice! I can not connect to a server, even I can not ping the server. I can tracoroute it and it ends one ip-adress before my server. then the connection is timed out. I cleared hosts.deny and all ip-table rules. The laptop can access any other website and service. =(
[17:16] <smoser> hggdh, karmic-server karmic-server-uec-20110216 is being built
[17:17] <smoser> maybe 8 hours until it is completely published
[17:17] <hggdh> smoser: thank you
[17:17] <hggdh> smoser: btw, did you spin a hardy-proposed also?
[17:19] <bluethundr> I am attempting to add a route to ubuntu 10.10 server but so far unsuccessful
[17:19] <bluethundr> http://pastie.org/1571401
[17:19] <bluethundr> I would appreciate any help you may provide
[17:21] <smoser> hggdh, hardy is diferent.
[17:21] <smoser> i need to get a ppa build of the kernel from jjohansen
[17:21] <smoser> hardy is a mess for this
[17:22] <jjohansen> smoser: why a ppa over the build that already exists for hardy?
[17:22] <smoser> or, if i can get a non-ppa build of the ec2 kernel for hardy, i'll collect the kernel and ramdisk and put them in -sandbox
[17:22] <smoser> well, for one, at the moment, no modules
[17:22] <smoser> but for two, we need it in the ppa
[17:23] <smoser> i think it is probably best to do it with -sandbox at first until we've tested, adn then push it to our ppa when its good (at which point hardy builds will pull from it)
[17:23] <jjohansen> smoser: right but what is going wrong that we need a new ppa verses using the regular build
[17:23] <smoser> but, from jjohansen i need a linux-modules
[17:23] <smoser> so, nothing really i guess.
[17:31] <RoAkSoAx> kirkland: /win 4
[17:32] <kirkland> RoAkSoAx: ?
[17:32] <RoAkSoAx> kirkland: happens to me all the time
[17:33] <RoAkSoAx> kirkland: sry. anyways, just finished with the changes to powernap. You can test now and upload if everything looks good :)
[17:36] <kirkland> RoAkSoAx: cool;  will be later this evening
[17:36] <kirkland> RoAkSoAx: i'm virtual sprinting today on something else ;-)
[17:36] <RoAkSoAx> kirkland: no worries :) I have cluster stuff to work on anyways :)
[18:08] <ahasenack> hey, is there a ppa for euca2ools? I'm on lucid and have 1.2-0ubuntu10.1
[18:11] <ahasenack> the eucalyptus site has a debian repository, but not ubuntu, I might try that
[19:10] <zul> hggdh: im going to re-install cempedak
[19:11] <fullstop> Hi all.  I'm setting up NFS4 between a few hosts following this guide: https://help.ubuntu.com/community/SettingUpNFSHowTo
[19:12] <fullstop> I have it working.  I'd like to allow group read / write between hosts, but I'm having permission problems.
[19:26] <smoser> hggdh, kernels/ramdisks for hardy are on their way up to ubuntu-testing buckets
[19:27] <smoser> i'll pastebin you something when i have it
[19:39] <hggdh> smoser: thank you
[19:39] <JamesPage> zul: did you get a change to look at the lucid minimal install?
[19:39] <zul> not yet but i will
[19:39] <hggdh> JamesPage, zul: looking at it (free time)
[19:40] <JamesPage> hggdh: thats great - thanks.
[19:40] <JamesPage> hggdh: I'm never sure which kernel should be installed for which type of install
[19:40] <smoser> hggdh, jjohansen just fyi, i updated https://wiki.ubuntu.com/UEC/Images/Publishing as to how to collect the hardy kernels and ramdisks
[19:41] <lenios_> does anyone know where (if available) to find how well a package has been tested?
[19:41] <smoser> hggdh, http://paste.ubuntu.com/567816/ <--- kernels and ramdisks are there
[19:41] <hggdh> JamesPage: it got hit by a time out (80 min, and nothing). I will run it by hand and see what gives
[19:42] <hggdh> smoser: thank you
[19:42] <RoAkSoAx> zul: I have cluster-agents almost ready. Will have it ready after lunch
[19:42] <smoser> for haryd, you'll just have to launch an instance with that kernel and that ramdisk
[19:42]  * RoAkSoAx lunch
[19:42] <zul> RoAkSoAx ac
[19:49] <sp00fz> can anyone tell me what does boot flags mean during encryption setup ?
[20:00] <JasonMSP> Im running a postfix, dovecot, mysql combination on my server.  Any good recommendations for an admin/gui tool?
[20:01] <Saturn2888> does anyone know which IRC channel I can use to get help with the ftp application? I'm trying to move a directory recursively from my machine to an FTP host. It's a long story, but I can only do it over FTP so I'm trying my best to use the put or send commands to do it, and those are only going to work with files, not directories
[20:04] <hallyn> dont' suppose you can make a tarball of it, xfer that, and extract it at the other end?
[20:07] <genii-around> Saturn2888: Have you tried something like mput -r *
[20:08] <hggdh> smoser: which AMI should I use?
[20:09] <smoser> http://uec-images.ubuntu.com/releases/hardy/release-20100827/
[20:11] <Saturn2888> mput? oh lemme try. genii-around from command line or ftp>? I was trying ncftpput just now and that didn't work either.
[20:12] <genii-around> Saturn2888: It's a command you can use from inside ftp
[20:12] <Saturn2888> genii-around: yeah, doing that now
[20:12] <Saturn2888> genii-around: local: -r: No such file or directory
[20:12] <Saturn2888> haha
[20:14] <Saturn2888> genii-around: it works w/ wildcards I think. I did /* and it started, but it's making the remote locations the same as local which is causing it to mess up
[20:20] <genii-around> Saturn2888: Well, depending on how many directories you have you might want to just do it manually. like use mkdir, cd into it, then use mput with full pathname and wildcard to put all from local dir named foobar into remote dir you just made called foobar , etc
[20:20] <genii-around> Apologies on lag, work got quite busy
[20:21] <Saturn2888> genii-around: I'm doing more tan 500 files :P
[20:21] <Saturn2888> than*
[20:23] <genii-around> Saturn2888: But roughly how many nested directories?
[20:25] <Saturn2888> genii-around: it's okay, I used Windows :P. A lot easier that way
[20:25] <Saturn2888> it was a lot. 5 deep
[20:25] <genii-around> Too bad they don't let you use scp
[20:26] <EvilPhoenix> eewwwww...Windows....
[20:26] <EvilPhoenix> >.<
[20:28] <Saturn2888> genii-around: I know! Haha. but actually, it was  ton of really small files and transferred really slowly over SFTP but was quick over FTP (trying to move over to a new server with SFTP, but have to use FTP for now)
[20:37] <Saturn2888> genii-around: thanks though :)
[20:50] <kirkland> hallyn: can you take a look at Bug 720210 ?
[20:51] <kirkland> hallyn: possible regression
[20:55] <hallyn> yeh I saw it go by
[20:56] <hallyn> did notice he's using -vga vmware, but haven't reproduced yet
[20:57] <hallyn> does 'bzr revert -r3006' revert *to* commit 3006, or revert commit 3006?
[20:57] <hallyn> ah, i see
[21:00] <hallyn> kirkland: -vga vmware with a natty image fails similarly for me.  Waiting for a maveirck iamge to clone to test separately
[21:01] <kirkland> hallyn: $ kvm -m 1024 -smp 2 -hda natty-desktop.img -vga vmware
[21:01] <kirkland> hallyn: works like a champ
[21:02] <kirkland> hallyn: what's most impressive is that it can do 2360x1770 resolution ;-)
[21:02] <kirkland> hallyn: and qemu scaling actually let's that fly!
[21:02] <kirkland> hallyn: on my 1280x800 x201 :-)
[21:03] <hallyn> funky, for me it just failed
[21:04] <RoyK> just use virt-manager :)
[21:05] <kirkland> hallyn: oh
[21:05] <kirkland> $ dpkg -S /usr/share/qemu/vgabios-vmware.*
[21:05] <kirkland> dpkg: /usr/share/qemu/vgabios-vmware.bin not found.
[21:05] <kirkland> dpkg: /usr/share/qemu/vgabios-vmware.debug.bin not found.
[21:05] <kirkland> hallyn: i manually copied those into place yesterday
[21:05] <hallyn> haha
[21:06] <hallyn> phew, that makes more sense then
[21:06] <kirkland> hallyn: okay, we need qemu-kvm to put those symlinks there
[21:06] <kirkland> hallyn: drop me a merge request and i'll upload ASAP
[21:06] <kirkland> hallyn: or a debdiff
[21:06] <hallyn> hm?  no, if you use the real package i think they'll get installed
[21:06] <kirkland> hallyn: huh?
[21:06] <hallyn> i have those files
[21:06] <kirkland> hallyn: what do you have in /usr/share/qemu/*vmware* ?
[21:07] <hallyn> /usr/share/qemu/vgabios-vmware.bin  /usr/share/qemu/vgabios-vmware.debug.bin
[21:07] <kirkland> $ ll /usr/share/qemu/vgabios-vmware.*
[21:07] <kirkland> lrwxrwxrwx 1 root root 29 2011-02-15 12:09 /usr/share/qemu/vgabios-vmware.bin -> ../vgabios/vgabios.vmware.bin
[21:07] <kirkland> lrwxrwxrwx 1 root root 35 2011-02-15 12:09 /usr/share/qemu/vgabios-vmware.debug.bin -> ../vgabios/vgabios.vmware.debug.bin
[21:07] <hallyn> /usr/share/qemu/vgabios.vmware.bin  /usr/share/qemu/vgabios.vmware.debug.bin
[21:07] <kirkland> same/same?
[21:07] <hallyn> yes
[21:08] <hallyn> I vividly remember creating debian/links for just that reason
[21:13] <hallyn> kirkland: feh, vmware uses VBE
[21:13] <kirkland> hallyn: hmm
[21:13] <kirkland> hallyn: interesting
[21:13] <kirkland> hallyn: okay, well then perhaps the reason that mine's working is that I installed Ubuntu Server and then apt-get installed ubuntu-desktop
[21:14] <kirkland> hallyn: so perhaps my grub is running in a legacy/server/non-vbe mode?
[21:14] <hallyn> kirkland: i dunno, bc like i say i was able to run natty with -vga std before
[21:14] <hallyn> kirkland: doh!
[21:15] <hallyn> kirkland: it wasn't natty, it was lucid
[21:15] <hallyn> kirkland: so that seems to confirm that roland's is a dupe of our -std vga bug
[21:15] <kirkland> hallyn: mkay
[21:15] <kirkland> hallyn: can you mark that bug confirmed, then
[21:15] <kirkland> or dupe it
[21:15] <lullabud> i'm configuring a syslog server and i noticed the directive to enable --MARK-- messages... what are they useful for?
[21:16] <hallyn> kirkland: i commented that it's probably a dupe, but haven't yet marked it dupe
[21:18] <kirkland> k
[21:18] <b0gatyr_> hi guys, kinda unrelated but dunno where to ask I remember a shell tool that allows you to keep track of your to do's but forgot the name.. any clues?
[21:19] <kirkland> b0gatyr_: rtm?
[21:19] <kirkland> b0gatyr_: remember the milk?
[21:20] <hallyn> oh, there are shell tools for that?
[21:24] <lullabud> also, i'm digging around looking for the new approved way to configure remote syslog, but i can't find it.  all i see is that -r is deprecated, but there's nothing listed for the new method.
[21:27] <b0gatyr_> kirkland: not exactly, there was a shell tool to keep track of your to do's that you can even ssh into and check ur tasks but forgot .. however the site you suggest is not bad at all
[21:40] <airtonix> gcal
[21:41] <hallyn> my kingdom for a grub tree that's bisect-safe
[21:43] <RoyK> bisect safe?
[21:43] <RoyK> what is that in English?
[21:50] <hoechste_zeit> lullabud: if you have --MARK-- every 20 minutes, you know with this accuracy until when your server was alive.
[21:50] <kirkland> hallyn: you might ask in #bzr what are the tricks to doing a bisect
[21:50] <kirkland> hallyn: i bet the have some
[21:50] <kirkland> hallyn: or lifeless might indulge you here ...
[21:51] <smoser> zul, ping
[21:51] <zul> smoser: yessssssss?
[21:51] <hallyn> kirkland: yes, i asked about the bzr-bisect plugin.  no anwers :(
[21:51] <hallyn> answers
[21:51] <smoser> why is there a linux-xen at https://launchpad.net/~ubuntu-on-ec2/+archive/ppa/+packages
[21:52] <hallyn> kirkland: but i'll just do it manually.  just wish i wouldn't keep bricking systems with bad builds
[21:52] <zul> smoser: thats from long long ago delete them
[21:52] <smoser> i suspect that is not the same as the linux-xen package in hardy
[21:52] <smoser> really ?
[21:52] <zul> nope they arent
[21:52] <zul> those were just the kernel that i was using when i was doing the images
[21:53] <smoser> so where is the kernel supposed to come from for hardy images
[21:53] <zul> actually it might have come from there, did you guys update the kernel since?
[21:53] <kirkland> hallyn: snapshot, dude
[21:53] <smoser> i don't actually think so, no
[21:53] <kirkland> hallyn: when i'm doing something like that, i'll often cp the disk image to /tmp (which is tmpfs for me in memory)
[21:54] <kirkland> hallyn: keep a golden master
[21:54] <kirkland> hallyn: and create a victim
[21:54] <zul> smoser: interesting
[21:54] <RoyK> Someone told me link aggregation won't speed up a single connection - is this really so? I thought link aggregation was done on the link level, and if so, the data should flow over both, or all, links without regards to IPs or MAC addresses...
[21:54] <kirkland> hallyn: do my test, then rm the victim, and cp the gold master back to a new victim for each test
[21:55] <kirkland> smoser: are you going to review lynxman's merge proposal at https://code.launchpad.net/~lynxman/cloud-init/puppet-mcollective/+merge/49819 ?
[21:55] <zul> smoser: is your new hardy images booting?
[21:56] <smoser> idont know why i didn't see that, kirkland
[21:56] <smoser> zul, no. they go straight to terminated
[21:56] <RoAkSoAx> zul: what else do I need to get configured in dhcp for it to work with cobbler?
[21:57] <zul> smoser: erm...
[21:58] <zul> RoAkSoAx: we always used dnsmasq
[21:59] <zul> smoser: does the new image work with the old kernel?
[21:59] <RoAkSoAx> zul: lol! cause my config is preconfigured for dhcpd
[22:00] <zul> RoAkSoAx: not sure im trying to help put out a fire right now
[22:01] <smoser> zul, we're trying old image new kernel
[22:01] <hallyn> kirkland: of course i have a snapshot :)
[22:02] <hallyn> but it means bzr bisect wouldn't work since that state gets lost
[22:02] <zul> smoser: heh you know that last time i touched the hardy xen kernel stuff for ec2
[22:04] <hallyn> kirkland: but that's not even what i was talking about.  but if a particular commit results in a 'grub' that is just plain broken, that throws off the bisect algorithm.  (so again, just as well that i'm doing manual loose bisect, bc git-bisect does not deal with that well)
[22:05] <kirkland> hallyn: ah, gotcha
[22:11] <lullabud> re syslog forwarding, found some great documentation on the rsyslog site - http://www.rsyslog.com/doc/rsyslog_reliable_forwarding.html
[22:14] <JasonMSP> i've got postfix installed and it works fine but I want to reinstall from scratch again. apt-get install postfix doesn't push me to the install page because it knows it is already installed.  How can I override and tell it to do it again?
[22:23] <lullabud> dpkg-reconfigure postfix ?
[22:30] <lullabud> JasonMSP: ^^
[22:31] <jkg> quick stupid apache question: I have a "NameVirtualHost *" at the top of sites-available/default, which then contains a <VirtualHost *> container, with a ServerName and several ServerAlias lines. in sites-available/myotherdomain I have a <VirtualHost *> container with a ServerName and a ServerAlias. am I missing something stupid?
[22:31] <RoyK> apt-get install win7
[22:32] <RoyK> jkg: remove the default
[22:32] <jkg> on starting apache, I get a warning "[Wed Feb 16 22:28:38 2011] [warn] NameVirtualHost *:80 has no VirtualHosts", and visiting the 2nd site gives me a blank front page or 404s for internal pages, so it looks like the docroot for that Vhost is /srv/www -- it is configured as /srv/www/otherdomain/htdocs
[22:32] <jkg> the default is a live site.
[22:33] <RoyK> jkg: then rename the default to a named one
[22:33] <lullabud> jkg: what's not working right?
[22:33] <RoyK> I've seen issues like that a few times
[22:34] <lullabud> jkg: oh, actually, shouldn't you configure your second site with <VirtualHost *> and then ServerName  ?
[22:34] <RoyK> no idea why, but in some circumstances, the default takes over
[22:34] <RoyK> just create a dummy for the default
[22:35] <jkg> RoyK: that behaves exactly the same :-/
[22:35] <jkg> oh I just dissite'd the default, renamed the file 'default' to 'apps' and a2ensite'd that.
[22:35] <RoyK> what is the servername? serveralas?
[22:35] <RoyK> jkg: pastebin the config
[22:35] <RoyK> !pastebin
[22:37] <jkg> is this enough? http://paste.ubuntu.com/567880/
[22:37] <RoyK> jkg: no </VirtualHost> there
[22:38] <lullabud> jkg: shouldn't you have a default site set up with no ServerName alias as a default catch-all?
[22:38] <jkg> there is a </VirtualHost> at the bottom of each file -- sorry, I kept it to just the top of each file to avoid boring you with the rest of the config
[22:38] <jkg> lullabud: I don't think so; I don't want it to respond to other names...
[22:38] <RoyK> jkg: just pastebin it all
[22:39] <RoyK> jkg: if you have a security issue, that should be fixed, not hidden
[22:40] <lullabud> jkg: if you don't want it to respond to all then why do you have namevirtualhost * ?
[22:40] <jkg> um. good question!
[22:41] <RoyK> lullabud: namevirtualhost * is not about responding
[22:41] <RoyK> it's about giving a name
[22:41] <RoyK> ServerName and so on is about responding
[22:42] <lullabud> that's inside of a <virtualhost> though, this was outside, defining the default.
[22:42] <jkg> http://paste.ubuntu.com/567881/ # full config
[22:42]  * RoyK sends lullabud into the RTFM part of the world
[22:42] <lullabud> http://httpd.apache.org/docs/2.0/mod/core.html#namevirtualhost
[22:42] <lullabud> that part of the world?
[22:42] <lullabud> yeah, i'm there.
[22:43] <RoyK> jkg: don't add them there - create new files in /etc/apache2/site-enabled
[22:43] <jkg> (obviously I've concatenated the 4 files together, so the Include sites-enabled/ is moot... unless that is the broken line, in which case it is intended to include everything I've put below)
[22:43] <RoyK> and symlink those to /ect/apache2/sites-available
[22:43] <jkg> RoyK: I have -- sorry, for simplicity I concatenated the 4 files
[22:44] <jkg> this is how Apache will read them anyway, right?
[22:44] <RoyK> jkg: and btw, ServerAlias takes several arguments, just list them, space separeted
[22:44] <RoyK> separeted, even
[22:46] <lullabud> RoyK: you're right, namevirtualhost is different than i though, specifying the IP#'s to listen on, not a domain list to respond to.
[22:47] <jkg> it definitely seems that when I hit it with a hostname not defined in what was the 'default' config (now explicitly called something else, but still with a ServerName) the DocRoot is /var/www as defined in apache2.conf, not as defined in my second VirtualHost container -- as if the second container (or sites-enabled file) isn't being read at all.
[22:48] <RoyK> lullabud: nameserverhost checks the host: header from the incoming HTTP request, not the IP
[22:48] <RoyK> as in
[22:48] <jkg> FWIW, apache2ctl configtest just gives that one warning and says Syntax OK
[22:49] <RoyK> GET / HTTP/1.1
[22:49] <RoyK> Host: karlsbakk.net
[22:51] <jkg> the exciting part of all this is, it worked great on 10.04, and 8.04, and someone persuaded me to migrate these sites tonight to our 10.10 box... and claimed to have tested it :-/
[22:57]  * jkg gives up, renames /etc/apache2, drops in the entire config from the 8.04 box and marvels as it all just works
[22:58] <jkg> god knows what was wrong, I can't see a difference :-/
[23:00] <jkg> thanks for the help investigating, anyway!
[23:02] <RoAkSoAx> zul: what should I submit for you to sponsor the new upstream release. The <new-release>.orig.tar.gz and <new-release>.debian.tar.gz ? (given that there's no .diff.gz)
[23:03] <RoAkSoAx> zul: or should I manually create a diff between <old version>.dsc <new-version>.dsc and post that too?