[00:28] <Psi-Jack> Has anyone worked much with AoE targets for HA networked storage mediums? I'm curious how fault tolerant it can be.
[00:33] <fbc_> I have a server with a 10gig system partition and would like to move and remap /var to another drive(bigger) as that is where all the data goes.
[00:33] <fbc_> could someone point me to a guide or how to?
[00:40] <fbc_> don't all jump in all at once now.
[01:48] <Andre_Gondim> I have a ubuntu server with cups installed, how do I access the interface in other pc in the same network
[01:53] <pmatulis> Andre_Gondim: i think cups has a web interface
[01:54] <Andre_Gondim> pmatulis, yeah, it has, but how may I do the accesss by other machine not localhost
[01:55] <pmatulis> Andre_Gondim: port 631 i believe
[01:55] <Andre_Gondim> I did, but don't show me nothing
[01:56] <pmatulis> Andre_Gondim: what exactly did you do?
[01:57] <Andre_Gondim> I installed ubuntu server, and cups, then in other machine try ip_from_server:631 and shows me a error page, like that page doesn't exists
[01:59] <pmatulis> Andre_Gondim: i guess the cups daemon isn't running
[02:00] <pmatulis> Andre_Gondim: you should get something like what i have on my system:
[02:01] <pmatulis> oops
[02:01] <pmatulis> cupsd   1938 root    7u  IPv4  25155      0t0  TCP 127.0.0.1:631 (LISTEN)
[02:01] <Andre_Gondim> root     19103  0.0  0.1   6872  2620 ?        Ss   21:45   0:00 /usr/sbin/cupsd -C /etc/cups/cupsd.conf
[02:02] <Andre_Gondim> pmatulis, like this?
[02:02] <pmatulis> Andre_Gondim: for me, it's listening on localhost
[02:03] <pmatulis> Andre_Gondim: change that in your config file
[02:03] <pmatulis> Andre_Gondim: what output do you get to:
[02:03] <pmatulis> sudo lsof -i4tcp:631 -nP
[02:04] <Andre_Gondim> cupsd   19103 root    6u  IPv4 165972      0t0  TCP 127.0.0.1:631 (LISTEN)
[02:04] <pmatulis> Andre_Gondim: there you go.  change the config file so it listens on your IP address
[02:05] <Andre_Gondim> pmatulis, the  config file is cupsd.conf?
[02:05] <pmatulis> Andre_Gondim: yeah
[02:06] <Andre_Gondim> I have inserted Listen my_other_machine_ip:631 and restarted the service cups, bug the result is the same
[02:06] <pmatulis> Andre_Gondim: your other machine?
[02:07] <pmatulis> Andre_Gondim: no, put the server's address
[02:07] <Andre_Gondim> oh, I got it
[02:08] <Andre_Gondim> pmatulis, many thanks, now I can setup my printer ;)
[02:08] <pmatulis> Andre_Gondim: good work
[02:10] <YDU_Remote> ...mornin y'all - anyone familiar with grub2 strangeness - i.e., on powerON (not reboot), grub just hangs with "out of disk" error...?
[02:12] <pmatulis> YDU_Remote: when you power on your machine does not boot but when you reboot it does?
[02:17] <YDU_Remote> pmatulis: Yeppers mate...exactly that. Strange, innit...
[02:18] <pmatulis> YDU_Remote: you might have a power issue (power supply or wall outlet)
[02:19] <YDU_Remote> pmatulis: All tested/checked/verified. This server installation also does the same thing at the lab, here, or anywhere else. Strange thing this - I've done the exact same installation/setup on several machines and this one is the only one that freaks like this. All the same hardware, same Ubu version, same everything.
[02:20] <pmatulis> YDU_Remote: but you're using the same p/s, obviously
[02:20] <pmatulis> YDU_Remote: or did you change it?
[02:21] <YDU_Remote> Used the PSU from a different box (the same, but just from a different box) -> same issue.
[02:23] <pmatulis> YDU_Remote: how many and what kind of drives?
[02:32] <YDU_Remote> pmatulis: Single drive, internal, 1.5tb => Samsung
[02:33] <pmatulis> YDU_Remote: swap in another one
[02:33] <YDU_Remote> pmatulis: Hard to do whilst this site is active - have to try that at COB
[02:33] <pmatulis> YDU_Remote: COB?
[02:34] <YDU_Remote> pmatulis: My bad - it's a WD (COB - Close of business)
[02:34] <pmatulis> YDU_Remote: alright
[02:36] <JasonMSP> Is it possible to adjust some setting so that I don't see the entire directory structure at the command prompt?  maybe just the current folder name?
[02:36] <YDU_Remote> These kinda niggly things really give me the shiites sometimes...re-installed grub2 so many times I feel like I'm dating it...
[02:38] <pmatulis> JasonMSP: yes, of course
[02:38] <pmatulis> JasonMSP: you want to "customize your bash prompt" (assuming you're using bash as shell)
[02:40] <JasonMSP> pmatulis: yes
[02:40] <pmatulis> JasonMSP: so google will tell you, basically you want to define the PS1 variable
[02:40] <JasonMSP> thanks
[02:41] <pmatulis> JasonMSP: your current prompt is given by 'echo $PS1'
[02:41] <EvilPhoenix> i'm trying to configure lighttpd to work
[02:41] <EvilPhoenix> i installed it with apt-get
[02:41] <EvilPhoenix> and i'm getting 403 forbiddens everywhere
[02:41] <EvilPhoenix> how can I fix that
[02:41] <pmatulis> EvilPhoenix: i guess you need to configure it
[02:41] <EvilPhoenix> no crap
[02:42] <EvilPhoenix> i came here looking for support with that
[02:44] <pmatulis> http://redmine.lighttpd.net/wiki/1/TutorialConfiguration
[02:44] <pmatulis> EvilPhoenix: that should get you started
[03:03] <tf2ftw> a
[03:05] <tf2ftw> chrismsnz,
[03:18] <|rt|> does anyone know of any reasons you can get really bad (4.3 MB/s) write speed and really good (187 MB/s) read with software raid in linux?
[03:18] <chrismsnz> raid 5?
[03:18] <|rt|> doesn't seem to matter what file system I use or kernel version
[03:18] <|rt|> yes 5 disk raid 5
[03:18] <chrismsnz> it's doing a lot of calculation per write
[03:19] <stlsaint> any of you folks ever use psad with server edition?
[03:19] <chrismsnz> i haven't seen it that bad, but it's bad
[03:19] <|rt|> nah...cpu's can do that really fast
[03:19] <|rt|> it's not cpu bound when this is happening
[03:19] <|rt|> like 97% idle cpu
[03:19] <|rt|> i've used 12 disk raid5 software raid without any issues in the past
[03:20] <chrismsnz> is one of the disks dragging it down?
[03:20] <chrismsnz> try to run hdparm on each individual device
[03:20] <chrismsnz> hdparm -tT
[03:20] <|rt|> yeah did that already...they are all about the same
[03:20] <|rt|> 50-60 MB/s
[03:20] <|rt|> but hdparm only test reads....and my slowness is writes
[03:21] <|rt|> I'll blow my raid volume away and run some bonnie++ tests on each drive
[03:22] <|rt|> but these same drives did a good job with OpenSolaris and ZFS
[03:22] <|rt|> if I can't get linux working right I'll try FreeBSD and just go back to ZFS
[03:23] <chrismsnz> zfs is nice :)
[03:23] <chrismsnz> if you're running a recent kernel the on-disk format for btrfs is now stable so might be worth a whack
[03:23] <|rt|> it is....I was hoping to play with BTRFS but the user space tools seem to be a bit lacking
[03:23] <chrismsnz> obviously not if it's production, though
[03:23] <|rt|> and it's documentation
[03:24] <chrismsnz> I prefer RAID 10 to raid 5
[03:24] <|rt|> chrismsnz: I started off playing with BTRFS on Maverick on this system just to see what it was like
[03:25] <|rt|> chrismsnz: yeah I agree only the raid 10 thing but these drives are only 250GB drives so for capacity I chose to stick to RAID5 for now
[03:25] <chrismsnz> in fact, you could probably use Linux's "raid 10" on your array - it outperformed raid6 in my tests
[03:25] <|rt|> but BTRFS doesn't do RAID5 yet
[03:26] <|rt|> yeah...I may do that just to test this....but I need to figure out what about my system Linux isn't likely
[03:26] <|rt|> liking
[03:26] <chrismsnz> well, if you want the bleeding edge - i heard that native linux ZFS is being released shortly as a compilable kernel module
[03:26] <|rt|> yeah I think i'll pass on that for now
[03:27] <|rt|> hopefully maverick +1 has BTRFS and the documentation and user space tools are improved
[03:27] <chrismsnz> natty narwhal D:
[03:28] <|rt|> integration with apt and snapshots would be nice too :)
[03:30] <chrismsnz> is your hardware exceedingly old or new?
[03:30] <|rt|> it's middle of the road
[03:31] <chrismsnz> a while ago linux was having trouble with an onboard sata controller I had and it affected speeds quite badly
[03:31] <|rt|> it's probably a year old or so but it wasn't cutting edge then
[03:31] <chrismsnz> shouldn't be a problem there :\
[03:31] <|rt|> yeah I've been looking for information about the SB700 chipset and performance issues but haven't found anything
[03:32] <|rt|> the system has an AMD Athlon 4850e cpu
[03:33] <chrismsnz> you're using software raid by mdadm right?
[03:33] <|rt|> yeah
[03:33]  * chrismsnz boggles
[03:33] <|rt|> i've tried the SATA controller in SATA mode and AHCI mode with the same results
[03:34] <|rt|> yeah i've never seen anything like this
[03:34] <|rt|> but i've tried a few older kernels on system rescue live cd's to make sure it wasn't a kernel issue
[03:35] <|rt|> what ever is going on doesn't seem to be tied to any kernel version
[03:35] <chrismsnz> you're using one parity disk?
[03:35] <|rt|> well raid5 doesn't have a dedicated parity disk...but it is n+1 parity
[03:35] <chrismsnz> err yeah
[03:36] <chrismsnz> i mean, i can see why it's slow to write... 5 disks = 4 reads and 2 writes per block
[03:37] <|rt|> I have plenty of pci slots open so one option may be to pick up a couple of 4 port sata controllers
[03:37] <chrismsnz> but like i said i haven't seen it that slow
[03:37] <chrismsnz> or 3 reads, hypothetically, plus parity calculation
[03:38] <|rt|> if the raid didn't rebuild at 50MB/s my first thought was partition mis-aligned but you normally don't see that much speed loss for that
[03:39] <|rt|> just did a bigger test to make sure I wasn't seeing some caching affect on the read spead
[03:40] <|rt|> wrote 7.4GB in 1767.36 seconds (4.2MB/s) and read 7.4GB in 49.046s (150MB/s) using dd
[03:41] <|rt|> time to blow away the raid and run bonnie++ on each drive to see what happens
[03:41] <|rt|> b/c if it's one drive the read speed could still be pretty good with raid5 as it could work around the slow drive
[03:41] <|rt|> not sure it's that intelligent though
[03:41] <chrismsnz> good luck :\
[03:42] <chrismsnz> also check dmesg and see if it's complaining about any of your hardware
[03:42] <|rt|> yeah hopefully something jumps out with the individual disks
[03:42] <|rt|> nothing in dmesg
[03:42] <chrismsnz> bummer
[03:49] <|rt|> chrismsnz: I think I found my problem
[03:49] <chrismsnz> ooh
[03:49] <|rt|> mkfs.ext4 on /dev/sda2 is going really slow....fast on all the other drives
[03:49] <chrismsnz> give it a prod with smartctl
[03:49] <chrismsnz> see if it's noticed a problem
[03:52] <|rt|> hmm sdd2 is slow as well...may have spoken too soon
[03:53] <|rt|> smart data is fine on all of the drives.....i'll have to swap what port the drives are plugged into and see if the problem follows ports, drives, or cables
[03:59] <Datz> hi, it seems as though cron isn't working on my system. Can someone help me out with a diagnosis?
[04:10] <chrismsnz> hi
[04:10] <chrismsnz> Datz: what makes you say that?
[04:12] <Datz> chrismsnz: I have cron jobs scheduled and they don't run
[04:12] <chrismsnz> ok, where did you schedule them?
[04:12] <chrismsnz> in crontab? cron.daily? cron.d?
[04:13] <Datz> crontab
[04:13] <chrismsnz> unprivileged user? or root user?
[04:13] <Datz> user
[04:13] <Datz> groups datz
[04:13] <Datz> datz : datz adm dialout cdrom plugdev crontab sambashare lpadmin admin
[04:13] <Datz> I added myself to crontab group
[04:13] <Datz> in hopes to fix it
[04:14] <chrismsnz> do you have an /etc/cron.deny or /etc/cron.allow file?
[04:14] <chrismsnz> (it's ok if you don't but if you do it will use it)
[04:15] <Datz> no, for either
[04:15] <chrismsnz> ok
[04:15] <chrismsnz> you're running 10.04?
[04:16] <Datz> yes
[04:16] <chrismsnz> type "status cron"
[04:16] <Datz> ok
[04:16] <Datz> status: Unable to connect to system bus: Failed to connect to socket /var/run/dbus/system_bus_socket: No such file or directory
[04:16] <Datz> this is a VM btw
[04:17] <Datz> don't know if it would matter in this case
[04:17] <chrismsnz> no... is it a normal server install?
[04:17] <Datz> yes
[04:17] <Datz> Linux VMserv 2.6.32-24-server #39-Ubuntu SMP Wed Jul 28 06:21:40 UTC 2010 x86_64 GNU/Linux
[04:18] <chrismsnz> does "ps -ef | grep cron" show you anything? is it running?
[04:18] <Datz> yes
[04:18] <Datz> root       757     1  0 Aug21 ?        00:00:01 cron
[04:18] <Datz> datz     30006 27262  0 02:34 pts/0    00:00:00 grep --color=auto cron
[04:19] <chrismsnz> okay
[04:19] <chrismsnz> show me the crontab line you used to schedule the task
[04:20] <Datz> I just used this for testing: * * * * * date > /tmp/foo.log
[04:20] <Datz> nothing in /tmp/ yet
[04:22] <chrismsnz> ok one sec, just going to check a box
[04:22] <Andre_Gondim> Datz, try */3 * * * * date > /tmp/foo.log
[04:22] <chrismsnz> try restarting the cron daemon in the meantime "restart cron"
[04:22] <Andre_Gondim> every 3 minutos
[04:22] <Datz> Andre_Gondim: I'll add that one as well
[04:22] <Andre_Gondim> and reload the cron liek chrislabeard said
[04:23] <Datz> I'm getting the same error I got with status
[04:23] <chrismsnz> ok
[04:24] <chrismsnz> try restarting dbus
[04:24] <chrismsnz> "restart dbus"
[04:24] <chrismsnz> then "restart cron"
[04:24] <Datz> should I try to restart cron with /etc/init.d/cron restart?
[04:24] <Datz> ok
[04:24] <chrismsnz> Datz: i don't know any more - half the friggen system is on upstart and half isn't
[04:24] <Datz> it holds up on restart
[04:24] <Andre_Gondim> Datz, yes
[04:25] <Datz> kk
[04:25] <chrismsnz> i'm testing these commands on a lucid box as i'm advising you, though :)
[04:25] <chrismsnz> yeah the init.d file just restarts it via upstart anyway
[04:25] <Datz> ah
[04:25] <Datz> yea
[04:26] <Datz> well I never restarted the system after some updates, and it keeps asking for a restart.. might this create a problem?
[04:26] <Datz> humm, well now there is a foo.log :P
[04:27] <Datz> so I think restarting cron worked :)
[04:27] <Datz> maybe I should restart ....
[04:27] <chrismsnz> yeah...
[04:27] <Datz> heh hehe
[04:27] <Datz> thanks
[04:27] <chrismsnz> i've had dbus die after an update
[04:27] <Datz> ah
[04:27] <chrismsnz> whole system seems to run on it now hehe
[04:27] <Datz> well.. bb in a few thanks again
[04:27] <Datz> ah
[04:27] <chrismsnz> cool
[04:32] <Tim_R> I have a question about doing virtual ethernet onto my server how would I do that? I also got virtual hosting on couple pages how would I set all this up
[04:33] <chrismsnz> you mean having aliased IP addresses on the same interface?
[04:33] <Tim_R> yes
[04:34] <chrismsnz> ok
[04:35] <chrismsnz> Here's a good howto on adding alias's for an interface and how to configure them correctly to be permanent
[04:36] <chrismsnz> http://www.cyberciti.biz/tips/ubuntu-linux-creating-ethernet-alias-for-eth0-network-device.html
[04:36] <chrismsnz> However!
[04:36] <chrismsnz> If you just want it for virtual web hosting, you'll be better off relying on apache's virtualhost configuration
[04:37] <Tim_R> well see I have virtual hosting now on my server
[04:40] <Tim_R> alright thanks chris for your help
[04:40] <chrismsnz> :o
[04:42] <tf2ftw> anyone have any experiance with MAMP and enabling htaccess?
[04:44] <chrismsnz> i don't use mamp, but the apache config to enable it should be the same
[04:47] <tf2ftw> the problem im having is rewrite is not working
[04:47] <tf2ftw> let me get the file one sec
[04:50] <tf2ftw> its httpd.conf correct?
[04:50] <chrismsnz> the htaccess file is usually ".htaccess"
[04:50] <chrismsnz> httpd.conf configures the server
[04:50] <tf2ftw> i mean to enable mod_rewrite
[04:51] <chrismsnz> I'm not sure how mamp has it set up
[04:51] <chrismsnz> yeah httpd.conf
[04:51] <chrismsnz> (probably_
[04:52] <tf2ftw> looks like MAMP loads a mod_rewrite.so
[04:54] <tf2ftw> ffffuuuuu they have AllowOverride in more than one place in httpd.conf and you have to enable all of them
[04:54] <chrismsnz> oh, weak
[04:55] <chrismsnz> you should just be able to enable it for your "location"
[04:55] <chrismsnz> htdocs or whatever
[04:55] <tf2ftw> hmm yeah I enabled it with an apache restart still no joy... still digging
[04:57] <tf2ftw> ah so you can use a ".htaccess" file in another name using AcessFileName. That would be good for security?
[04:59] <tf2ftw> do i have to have the htaccess file in the root of my web server or can i just use it in the dir im working in (ex /htdocs/site1/) ?
[05:00] <tf2ftw> Its always worked for me just having it in teh working dir.
[05:00] <chrismsnz> should be in the dir your working in
[05:21] <tf2ftw> chrismsnz, why would httpd.conf be empty in /etc/apache2/httpd.conf ?
[05:22] <twb> tf2ftw: because most stuff happens in the .d's
[05:22] <tf2ftw> is that in sites-enabled?
[05:22] <chrismsnz> tf2ftw: twb is right, if your setting it up on ubuntu/debian
[05:23] <tf2ftw> yeah its ubuntu
[05:23] <twb> chrismsnz: if it was C5 it'd be /etc/httpd :-/
[05:23] <twb> Stupid apache people think theirs is the only httpd
[05:23] <chrismsnz> yeah he was talking about mamp before
[05:23] <tf2ftw> well looks like i might have found my htaccess problem
[05:23] <tf2ftw> the script itself
[05:24] <tf2ftw> does this look right? http://pastebin.com/F3UJ1a5k
[05:25] <chrismsnz> not a regex expert, but the second ^ doesn't make sense to me?
[05:26] <tf2ftw> let me look into that
[05:27] <chrismsnz> well it's supposed to indicate the beginning of a string, so it doesn't make sense to have it in the brackets there when you're matching any character for the file name
[05:27] <chrismsnz> again, not 100%
[05:28] <chrismsnz> a rewrite of mine looks like this:
[05:28] <chrismsnz> RewriteRule ^.*$ /index.php [NC,L]
[05:28] <tf2ftw> you are correct about the second ^
[05:29] <chrismsnz> so try changing yours to "RewriteRule ^.+\.html$ /$1.php [QSA,L]"
[05:30] <chrismsnz> err you need the brackets actually to capture
[05:30] <chrismsnz> RewriteRule ^(.+)\.html$ /$1.php [QSA,L]
[05:31] <tf2ftw> aha!
[05:31] <tf2ftw> take that "/" out before $1 and you are a winner
[05:31] <tf2ftw> Thanks
[05:32] <tf2ftw> The strange thing is i copied this htaccess file from another project that was working.. i might have copied the wrong one (obviously)
[05:32] <tf2ftw> thanks
[05:33] <chrismsnz> _b
[05:34] <tf2ftw> here is something weird, this htaccess works on my local MAMP but not my Ubuntu apache2..
[05:34] <tf2ftw> getting an Internal Server Error
[05:34] <chrismsnz> check the error log it normally sheds a little more light
[05:36] <tf2ftw> ".htaccess: Invalid command 'RewriteEngine', perhaps misspelled or defined by a module not included in the server configuration"
[05:37] <chrismsnz> heh
[05:37] <tf2ftw> check this out, how the folder is not showing up on the web server
[05:37] <chrismsnz> "sudo a2enmod rewrite"
[05:38] <chrismsnz> then restart apache
[05:38] <tf2ftw> you nailed it again. shit im going to have to start paying you
[05:38] <tf2ftw> o.O
[05:39] <tf2ftw> luckily i am at the end of my server journey here (i hope)
[05:40] <chrismsnz> hehe, already getting paid but considering going consulting next year
[05:40] <tf2ftw> good for you man. I took the jump recently.
[05:41] <tf2ftw> this server is great for keeping everything on one box. no more passing thumb drives or emails.
[05:42] <chrismsnz> yeah, testing the waters here in NZ to see if I can make a living as a python/django dev & sysadmin
[05:42] <tf2ftw> good luck.
[05:43] <tf2ftw> ok im checking out. Thanks again
[06:00] <ohzie> MDADM: Who are the experts? =D
[06:09] <chrismsnz> shout out, if someone's around they'll let ya know :)
[06:09] <chrismsnz> what's your prob?
[06:11] <ohzie> Trying to figure out if I can do raid 6 with more than two parity volumes, also spreading parity blocks redundantly across drives instead of having specific drives handle it. I'm very new to software raid, and not sure how to procede. =D
[06:11] <chrismsnz> cool
[06:11] <chrismsnz> well, you'll be happy to know that raid5 and raid6 have distributed parity, not dedicated disks
[06:11] <ohzie> I did look over the manual, and I'm under the impressive from a combination of the man file and the linux-raid wiki that I can define parity drives with -spare, but I wanted to talk to someone who has done it before I've done it.
[06:11] <ohzie> before I do it*
[06:12] <ohzie> chrismsnz, Sweet!
[06:12] <ohzie> Then I think I read it wrong and that -spare is probably not what I thought it was.
[06:14] <chrismsnz> a spare isn't a parity drive, it's a drive that is inactive until there's a problem with the array - at which point it will jump in and start reconstructing
[06:14] <ohzie> chrismsnz, That's automatic?
[06:15] <ohzie> Not even pillar does that crap
[06:15] <ohzie> Why the hell am I using Pillar
[06:15] <chrismsnz> what the hell is pillar D:
[06:15] <chrismsnz> ah, san/nas brand
[06:15] <ohzie> It's awesome
[06:15] <ohzie> but it doesn't do that
[06:16] <ohzie> I mean if we needed the iops I'd understand, but we really don't.
[06:16] <chrismsnz> i think it probably does, they might call it something different
[06:16] <chrismsnz> hot-spare is the other common name
[06:16] <ohzie> When a drive failed in one of our bricks, I had to come in on a sunday to plug the new one in. :3
[06:17] <chrismsnz> yeah, obviously it has to be configured as a spare at first - remember it's completely unused until there's a problem
[06:18] <chrismsnz> how many disks are we talking?
[06:19] <ohzie> I'm thinking 20.
[06:19] <chrismsnz> linux supports raid5 and raid6, but depending on the circumstances I normally lean towards Linux raid 10 or raid 1 + 0
[06:19] <ohzie> It depends on what they want to pay for.
[06:19] <ohzie> Between 10 and 20
[06:19] <ohzie> with the auto-rebuilding, i might be able to sell them on more
[06:20] <ohzie> with raid 6 can you have more than one backup of a parity block?
[06:20] <ohzie> like, losing three drives instead of two?
[06:20] <ohzie> or is more redundancy required for that?
[06:20] <chrismsnz> here's a big entry on raid levels http://en.wikipedia.org/wiki/RAID
[06:21] <chrismsnz> covers everything including linux's non-spec raid 10
[06:21] <chrismsnz> ohzie: not sure about that
[06:22] <chrismsnz> you can specify whatever redundancy you like with linux special mdraid 10 http://en.wikipedia.org/wiki/Non-standard_RAID_levels#Linux_MD_RAID_10
[06:22] <chrismsnz> ohzie: are you building an appliance?
[06:22] <ohzie> chrismsnz, Yes. :D
[06:23] <ohzie> Basically they want more storage in one of our locations, but they don't want to spend 80 thousand dollars on starting another pillar
[06:23] <chrismsnz> ohzie: you should play with FreeBSD and ZFS
[06:23] <chrismsnz> Linux has btrfs in the pipeline which is looking good, but not fully baked yet
[06:25] <chrismsnz> http://en.wikipedia.org/wiki/Zfs ZFS is awesome at managing massive amounts of storage
[06:26] <chrismsnz> under linux I'd probably set up the underlying RAID comprising all of the disks and plop a flexible LVM scheme on it
[06:40] <cerberos> I've got a new ubuntu server 10.04 VPS and I can't get ssh to work with keys (I've been trying on and off for days). my key is /root/.ssh/authorized_keys/id_rsa.pub - here is /etc/ssh/sshd_config http://dpaste.com/236672/
[06:58] <cerberos> doh, authorized_keys is a file
[08:47] <kaushal> hi
[08:47] <kaushal> can i get help regarding xen on Ubuntu Server ?
[08:47] <kaushal> or is there a irc channel for xen on Ubuntu ?
[08:48] <jpds> kaushal: #ubuntu-virt ? As far as I know, KVM is the only supported option.
[08:48] <kaushal> ok
[08:49] <kaushal> jpds: I am looking out for vmlinuz-2.6.32-24.xen and initrd.img-2.6.32-24.xen
[08:49] <kaushal> Cannot join to channel #ubuntu-virt (You must be invited)
[09:00] <kaushal> jpds: sorry about asking basic question
[09:01] <kaushal> so xen type is KVM ?
[09:01] <kaushal> not sure i understand that
[09:02] <jpds> kaushal: KVM is a different virtualisation technology.
[09:02] <twb> KVM is hardware virtualization.  Xen is paravirtualization (though it will fall back to full virtualization).
[09:03] <kaushal> ok
[09:03] <kaushal> so whats the difference between them ?
[09:03] <twb> The difference is in how virtualization is done.
[09:09] <kaushal> twb: Thanks
[09:31] <imyousuf> Hi
[09:31] <imyousuf> I am trying to install sun-java6-jdk because Cloudera Distributed Hadoop has it as dependency, but I can't because of dependency failure
[09:32] <twb> In Lucid, Sun Java is in the "partners" repo.  Have you enabled it?
[09:32] <imyousuf> twb: let me check
[09:33] <twb> (Incidentally, if you're trying to make proprietary "enterprise" crap work on linux, you might as well just shoot yourself now.  It's NOT FUN.)
[09:34] <imyousuf> twb: CDH is Open Source :) they contribute back to Apache AFAIK
[09:35] <twb> Everybody *says* that; I believe them when Debian ftpmasters allows it through the NEW queue.
[09:35] <imyousuf> :-D
[09:36] <twb> FWIW, "wnpp-check hadoop" has no hits
[10:02] <\sh> moins
[10:03] <siretart> \sh: OK
[10:04] <\sh> siretart, cool...I just found the discussion about the freeze exception for fai
[10:04] <\sh> siretart, http://www.mail-archive.com/ubuntu-motu@lists.ubuntu.com/msg04598.html but I don't read anything about a granting ;)
[10:07] <siretart> \sh: there wasn't much disagreement either.
[10:07] <siretart> ah, you mean it would annoy someone if we broke FF for fai?
[10:08] <\sh> siretart, yes...but when we use the discussion from that time as a "granted FFE for FAI"...I'm happy to upload
[10:09] <\sh> siretart, and only today I got a message on fai-linux that the PPA packages are working
[10:10] <siretart> oh, what a shame
[10:11] <siretart> \sh: TBH, I'd say just upload. the risk of breaking other unrelated packages is neglectable AFAIUI.
[10:11] <siretart> in terms of: 0
[10:11] <\sh> siretart, it won't break anything...
[10:12] <\sh> well, I'll prepare maverick packages + add the pointer to the ml archive for documentation...:) at the same time I'll do an FFe report for zend-framework ;)
[10:12] <siretart> cool!
[10:25] <qiyong>  reload squid
[10:25] <qiyong>   reload: Unknown instance:
[10:32] <drurew> anyone a cacti genious here? ; Im having a wierd problem where my graphs restart collecting at @ 2am..
[10:33] <twb> Most people here are humans.
[10:33] <twb> Or at least anthropoids
[10:33] <twb> (Oops.)
[10:35] <drurew> hey twb...
[10:37] <twb> wot?
[10:38] <drurew> "hey" as in "drurew waves to twb"
[10:42] <soren> jdstrand: It seems that the libvirt packaging branch finally has been succesfully imported. Do you want to start using it for our packaging work?
[11:28] <Zikey> Hi, I'm running ubuntu 8.04LTS raid1 (MD+LVM), the first drive died, can I directly replace the first drive by a new one and boot up ?
[11:30] <Zikey> I'm affraid of what will happen when ubuntu will find a new empty drive as the first drive, will it automatically ID it and replace it as /dev/sda, or will it create a /dev/sdc ? (/dev/sdb being the healthy disk)
[11:30] <twb> Zikey: that depends how grub is configured on the second drive
[11:30] <Zikey> it's lilo
[11:30] <twb> Zikey: basically, you can't rely on it working
[11:30] <twb> Well, lilo I can't comment on
[11:30] <twb> lilo doesn't work with LVM reliably anyway
[11:31] <twb> (Since lilo remembers block offsets into the disk, and LVM can move extents around.)
[11:31] <Zikey> with 8.04 we didn't have a change to install grub that didn't work with lvm
[11:31] <twb> Zikey: that's because you tried to put /boot on LVM on RAID, which is 1) stupid; and 2) not supported by grub.
[11:31] <Zikey> k
[11:31] <twb> Note that this can happen accidentally due to a bug in d-i in 8.04
[11:32] <Zikey> unfortunately I didn't have the choice
[11:32] <\sh> siretart, bug #627337
[11:32] <twb> (Namely: if you allocate /boot to md0 in partman, then configure LVM, it will forget about md0's configuration, go through the WHOLE install, then put /boot in with LVroot and force lilo.)
[11:32] <Zikey> can i boot the 8.04LTS server by pressing some key to drop in command line without starting everything ?
[11:32] <NightDragon> hello all
[11:32] <NightDragon> a quick question if you dont mind
[11:32] <twb> Zikey: well, you can boot with "single" which will give you a root shell.
[11:33] <NightDragon> i'm getting an errror after setting up cups as a server,
[11:33] <Zikey> twb, how do you do that ?
[11:33] <twb> Zikey: it's usually the second option in the bootloader "rescue"
[11:33] <NightDragon> Request from "129.93.xxx.xxx" using invalid Host: field "xxxx.xxx.edu:631"
[11:33] <Zikey> what do I need to type when LILO boot ?
[11:33] <NightDragon> (xxx's are censored of course)
[11:33] <twb> Zikey: otherwise, manually add "single" to the boot parameters, however that is done
[11:33] <NightDragon> when i try to axs the website, i get a '400 bad request', the website being the cups web admin page
[11:33] <twb> Zikey: I haven't used lilo this millenium, so I don't remember
[11:34] <twb> NightDragon: "censoring" a public IP isn't exactly going to protect you
[11:34] <NightDragon> i know.
[11:34] <NightDragon> matter of habit.
[11:34] <twb> Bloody daft habit
[11:34] <NightDragon> better safe then sorry mate
[11:35] <NightDragon> oh look
[11:35] <NightDragon> i fixed my own problem :X
[11:35] <NightDragon> sorry guys
[11:35] <NightDragon> lol
[11:35] <twb> I expect you fat-fingered the URL in your browser
[11:37] <NightDragon> no, actually i added the magic line in my conf file
[11:37] <NightDragon> (some line that was apparently needed)
[11:38] <NightDragon> in any case, while i'm here... does anybody know of anything that can enable one to get some really good control over printing on a print server? something that extends the basic functionality of CUPS perhaps?
[11:38] <twb> Yeah, thingy
[11:38] <twb> Some guy in europe makes it...
[11:39] <twb> pykota, that's it
[11:39] <NightDragon> twb: narrows it down a bit. :-P
[11:39] <twb> It's *really* annoying to work with, but AFAIK nobody else has much along those lines
[11:39] <twb> It does stuff like rasterizing the page and calculating ink usage
[11:40] <NightDragon> actually, i think i've heard of this
[11:40] <twb> And per-user ink/paper printing allowance, of course.
[11:41] <NightDragon> yeah, kinda lame that they want to almost charge you for documentation
[11:43] <twb> I think I basically said "I don't want to deal with this, give it to an intern to deploy"
[11:43] <NightDragon> lol
[11:43] <NightDragon> actually, i have a question for you
[11:44] <NightDragon> i'm really sick and tired of our office (of which i'm the SA) having direct printing capabilities to all of the printers in the office
[11:44] <NightDragon> the results of which are some pretty stupid crap
[11:44] <NightDragon> ...such as the fact that lordy help us if an IP gets changed inadvertantly, etc
[11:45] <twb> So move the printers to a new subnet and have the router block it
[11:45] <twb> *block access to it
[11:45] <NightDragon> ...or i have to log on as adminstrator (a PITA in XP) to add a printer
[11:45] <NightDragon> ...or when some fucktard prints off a thousand page job to the printer i use (a color printer, which happens to be located right next to me)
[11:46] <NightDragon> so my question...  i want to convince my boss to move to a cups/samba print server. What are the talking points of this (advantages, disadvantages, etc)
[11:46] <NightDragon> , and what should i really know that i might not already? :)
[11:46] <twb> SPOF
[11:46] <NightDragon> yeah
[11:46] <NightDragon> it has to be failproof
[11:47] <twb> Versus the administrative cost of you having to go around fixing people's printer configs
[11:47] <NightDragon> if i work with a single print server, can i (1) add printers to the server, and have it reflect regionally across many users?
[11:47] <twb> Yes
[11:47] <twb> Er, depends what you mean by "regionally"
[11:47] <NightDragon> 2) can i assign users into groups, and give them access to different printers depending on their groups
[11:48] <NightDragon> (i cant say 'globally' if i'm going to segment by groups lol)
[11:48] <twb> I *think* so; I haven't done it myself
[11:48] <NightDragon> ooh :D
[11:48] <twb> Obviously it'd also depend on whether you were going via IPP or Samba or both
[11:48] <NightDragon> samba
[11:49] <NightDragon> that way, it will use NT domain authentication for the usernames, wont it?
[11:49] <twb> So you have a homogeneous Windows network?
[11:49] <NightDragon> NTLM?
[11:49] <twb> I imagine s
[11:49] <twb> *so
[11:49] <NightDragon> yeah, we sit within a windows domain
[11:49]  * NightDragon is the stubborn bastard who will shoot himself before he picks a windows server over a good *nix server
[11:50] <NightDragon> come hell or high water lol... i hate working at a windows workstation as much as it is
[11:50] <NightDragon> well this is all moot, if you use samba you can authenticate over various things, including LDAP (which means AD), even if your not using NTLM directly
[11:51] <NightDragon> but question (4): Can you do cool things in cups like say... "users can print to this printer, but their jobs must have at least these many pages"
[11:51] <twb> Samba can talk to AD, but it can't *be* AD except in Samba 4.
[11:51] <twb> I don't know.
[11:51] <twb> Talk to #cups or so
[11:52] <NightDragon> and question 5)  Can you add imaging scanners into a print server? we have a really slick high-cap, high-speed scanner... and it would be nice to be able to network that so everyone can use it
[11:52] <twb> ditto
[11:52] <NightDragon> idd
[11:52] <NightDragon> well this is good stuff
[11:53] <NightDragon> ty
[11:53] <twb> No worries
[11:55] <NightDragon> of course, there is one added benefit to having a cups server... no waiting for 20 minutes while joe tool over there completes his book job
[11:55] <NightDragon> >:)
[12:10] <ttx> ScottK: re: cobertura-maven-plugin, I'll have a look -- however my maven foo is intentionally limited.
[13:21] <Zikey> do you know if mdadm support replacing a raid 1 disk with a higher capacity one ?
[13:21] <Zikey> basically can mdadm do raid1 with two differents disks size ?
[13:22] <Pici> I believe its limited to the smallest disk size.
[13:22] <Zikey> k
[13:22] <patdk-wk> yep
[13:22] <Zikey> I'm stuck, I boot up the livecd, got a terminal
[13:22] <Zikey> unfortunately mdadm is not running
[13:23] <Zikey> when I install/run it it doesn't detect my array
[13:39] <twb> Zikey: don't use ubuntu desktop CD for recovery, it sucks
[13:39] <twb> Find something that at least has mdadm and lvm pre-installed in the live CD
[13:40] <twb> And yes: RAID1 of inequal nodes will use the smallest node size.
[13:48] <NightDragon> ,
[13:54] <Zikey> When i run fsck on the filesystem, it says the fs is clean right away, is there any way to force it to verify it ?
[13:54] <patdk-wk> -f
[13:54] <Zikey> thx :)
[13:54] <patdk-wk> man fsck :)
[13:55] <Zikey> there is no -f in the man :(
[13:55] <Zikey> (ubuntu 8.04 LTS server)
[13:55] <patdk-wk> odd
[13:56] <patdk-wk> fsck --help :)
[13:56] <Zikey> doesn't work either :)
[13:57] <patdk-wk> it does on my 8.04 server
[13:57] <twb> In what way does it "not work"?
[13:58] <Zikey> not from the livecd it gives me an error (can not try again since it's checking now)
[13:58] <Zikey> I will tell you the error when done :)
[13:59] <Zikey> done
[14:00] <Zikey> no error, it just gives me: fsck 1.40.8 (13-Mar-2008)
[14:00] <Zikey> when i enter: fsck --help
[14:01] <patdk-wk> must be a livecd thing, I am doing this on a live system
[14:01] <twb> That's a really old CD
[14:01] <twb> fsck here is 2.17.2
[14:01] <patdk-wk> 1.40.8 on my 8.04 system
[14:02] <twb> Huh, ok
[14:05] <Zikey> anyone every tried to replace a raid1 disk of an array (mdadm+lvm2) ?
[14:05] <Zikey> I wonder if it's that easy to remove the first disk and put a new one
[14:09] <Zikey> do you know if 8.04 support hot disk swapping ? can I remove the first disk and replace it while the system is running ?
[14:13] <jdstrand> soren: if it actually works and has a sane workflow, sure
[14:14] <soren> jdstrand: Does bound branches+dch+debcommit count as sane for you?
[14:14] <patdk-wk> heh, hotswapping is a hardware issue, really
[14:15] <jdstrand> soren: sure, assuming it works :)
[14:15] <twb> patdk-wk: not just hardware
[14:15]  * jdstrand is slightly jaded
[14:15] <soren> jdstrand: We can always try and fail miserably.
[14:15] <zul> slightly? :)
[14:15] <soren> jdstrand: :)
[14:15] <patdk-wk> ya, but linux has supported hotswapping drives for a while now
[14:15] <jdstrand> :)
[14:15] <twb> Buses that don't support hot-swapping natively sometimes have non-standard vendor hacks to add it, and that requires a Linux driver for that vendor
[14:16] <twb> (Unless it's handled entirely in the hardware RAID card, in which case You Win!)
[14:17] <patdk-wk> I do it on non-supported busses too, just force a bus reset after I swap
[14:19] <Jeeves_> echo
[14:20] <Jeeves_> echo '- - -' > /sys/block/dev/blah/host/blah/scan
[14:20] <Jeeves_> orso :)
[14:24] <tynot> anyone know the syntax to open port 5900 for VNC server?
[14:24] <patdk-wk> run the vpn server?
[14:24] <patdk-wk> if you mean firewall, guess it depends on what you use to manage your firewall
[14:25] <tynot> nope. ubuntu's on a local network. just trying to vnc to it from a mac.
[14:26] <patdk-wk> heh? no firewall?
[14:26] <binBASH> living on the edge
[14:26] <patdk-wk> in my case, hope you trust your coworkers, or family :)
[14:26] <tynot> just for remote admin locally.
[14:32] <soren> jdstrand: Alright, I think we're all set. I've just pushed a couple of changes.
[14:37] <zul> kirkland: ping when you are around?
[14:39] <tynot> anyone? can't get vine vnc to run on a MacBook Pro in LAN.
[14:41] <jdstrand> soren: ok, so we are talking about lp:ubuntu/libvirt?
[14:41] <soren> jdstrand: We are.
[14:49] <soren> jdstrand: is that ok?
[14:50] <jdstrand> soren: as much as I understand udd, sure -- I just wanted to make sure we were on the same page. I'm grabbing the branch now
[14:50] <tynot> anyone care to share how to setup and run vncserver locally?
[14:50] <soren> jdstrand: It's pretty simple, really.
[14:50] <tynot> vnc4server's already installed and running?
[14:50] <soren> jdstrand: When you check it out, you  "bzr checkout lp:ubuntu/libvirt".
[14:51] <soren> jdstrand: If you've already branched it, you can "bzr bind lp:ubuntu/libvirt"
[14:51] <soren> jdstrand: That does the same thing.
[14:51]  * jdstrand nods
[14:51] <tynot> soren, thx. that's the syntax to run it from a client or server side?
[14:51] <soren> jdstrand: Oh, in the latter case, you also do a "bzr update" after binding, just in case. Otherwise it gets a bit confused.
[14:51] <soren> tynot: huh?
[14:51] <soren> tynot: Er... Talking to jdstrand about something completely different :)
[14:52] <tynot> I've both a MacBook Pro and a Windoze 7 PC trying to run Vine and UltraVNC, respectively.
[14:52] <soren> jdstrand: So, you make a change and use dch to add an entry to the changelog about it.
[14:53] <soren> jdstrand: like so: dch "whatever you want to be in debian/changelog"
[14:53] <tynot> sorry for the confusion, folks
[14:53] <soren> jdstrand: When you've done that, you run debcommit.
[14:53] <soren> jdstrand: This commits it to bzr with a commit log entry based on the debian changelog.
[14:53] <soren> jdstrand: Being a bound branch, this change is pushed to launchpad immediately.
[14:54] <soren> jdstrand: Once we want to roll a new upload, one of use will run "dch --release" which updates the timestamp in the changelog, and replaces "UNRELEASED" with "maverick" in the changelog.
[14:54] <soren> jdstrand: after that, you run "debcommit --release".
[14:55] <soren> jdstrand: This pushed the changes to the bzr branch and makes a tag.
[14:55] <soren> jdstrand: Once this succeeded, you build the source package and upload.
[14:56] <soren> jdstrand: It sounds tedious, but it's really a quite delightful workflow if more than one person is working on a package.
[14:56] <jdstrand> thanks for the explanation. I've only ever tried the udd thing with merges, and always had trouble with it
[14:56] <tynot> vnc, anyone?
[14:57] <jdstrand> soren: it is basically the same workflow I've used on other stuff with a bzr branch, except the debcommit parts
[14:57] <soren> tynot: What are you trying to do?
[14:57] <jdstrand> soren: so it should be quite comfortable
[14:57] <jdstrand> soren: how is the tag used in LP?
[14:58] <soren> jdstrand: I'm not sure it is, really.
[14:58] <soren> jdstrand: Well, you can see it in loggerhead.
[14:58] <tynot> soren, uh, not be in front of the linux console? just trying to admin the ubuntu server locally from a floor below me through vnc.
[14:59] <Zikey> If you change a disk (/dev/sdb) with another one, will it be mounted by default as /dev/sbc (because udev will keep /dev/sdb for the previous one) ?
[14:59] <Lichte> I just installed mysql-server and client and set the root password during setup, but I get this when I try 'mysql -u root -p' :  ERROR 1045 (28000): Access denied for user 'root'@'localhost' (using password: YES) ;   How do I get logged into mysql ???
[14:59] <soren> tynot: Why do you think you need vnc?
[14:59] <jdstrand> hallyn, kirkland: hey. fyi ^ please note we are going to try udd (Ubuntu Distributed Development) with libvirt. if you are unfamiliar with it, soren just gave a really good explanation of the workflow
[15:00] <soren> tynot: ssh is what you need, mate.
[15:01] <tynot> soren, clients are a Windoze 7 PC and a MacBook Pro. Either of them would be great to setup. I've UltraVNC on the PC and Vine on the MacBook. I don't know that I need VNC. I'm not married to it, although we're stuck in a PC world here without too much time.
[15:01] <soren> tynot: so stop wasting time on vnc and install ssh. It's all you need.
[15:02] <tynot> soren, ssh for Windoze?
[15:02] <zoopster> tynot: came in late, but just use putty on the win7 pc and native terminal ssh from OSX
[15:02] <zoopster> tynot: assuming you want to ssh into ubuntu server from those clients
[15:03] <soren> tynot: The idea of <insert whatever here> for Windows is a bit foreign to me, but I know for a fact that there are ssh clients for Windows.
[15:03] <tynot> I don't mind so long as we get a gui. Not adept enough to get around in a command line world anymore. I've already forgotten most of that stuff, unfort. VI was hell.
[15:04] <soren> tynot: You don't get a gui.
[15:04] <soren> tynot: If you want a gui, this is the wrong channel.
[15:04] <tynot> bye
[15:05] <soren> jdstrand: In fact, I'll roll a release straight away. The patch I added was just ACK'ed upstream, so I might as well.
[15:05] <jdstrand> cool beans
[15:05] <hggdh> Daviey: got a question re. euca2ools and Maverick
[15:06] <zul> hggdh: no
[15:06] <hggdh> Daviey: and good morning/afternoon
[15:06] <hggdh> zul: oh yes
[15:06] <hggdh> :-)
[15:06] <zul> hggdh: no really...no
[15:06] <hggdh> zul: absolutely yes. Just a small, tiny really, question
[15:07] <Daviey> hggdh: o/
[15:07] <soren> jdstrand: I suck.
[15:07] <hggdh> Daviey: bug 615442 -- would this be fixed on 1.3? Just asking, I intend to check on it
[15:08] <Lichte> I just installed mysql-server and client on 10.04 Server and set the root password during setup, but I get this when I try 'mysql -u root -p' :  ERROR 1045 (28000): Access denied for user 'root'@'localhost' (using password: YES) ;   How do I get logged into mysql ???
[15:08] <Daviey> hggdh: I don't think that bug is directly related to eucatools... i think it's eucalyptus
[15:09] <todd> Lichte: did you set a root password when you installed mysql?
[15:09] <Daviey> hggdh: But can be fixed with what we have already.
[15:09] <todd> Try without the -p and see if it lets you in.
[15:10] <hggdh> Daviey: I had a look at the eucalyptus source, and there is ./clc/tools/src/euca_admin that seems to have the imports, but not quite as python would like them
[15:10] <Lichte> todd: ERROR 1045 (28000): Access denied for user 'root'@'localhost' (using password: NO)
[15:11] <soren> jdstrand: The quilt source packages are special. When I add a patch, I need to add it both to debian/patches /and/ apply it to the source tree.
[15:11] <hggdh> Daviey: one thing that I would like to have is euca-add-user (intead of going to the web interface)...
[15:11] <Daviey> hggdh: This is indeed the case... it's an issue that has been deferred whilst i waiting for upstream to land a patch i sent them
[15:11] <hggdh> Daviey: oh, OK. I am happy now
[15:11] <hallyn> jdstrand: alright, thx for the heads-up
[15:12] <Daviey> hggdh: We will have that working..
[15:16] <soren> jdstrand: Uploaded. Let's see what happens.
[15:16] <soren> jdstrand: Oh, bleh, we're frozen, aren't we?
[15:16] <jdstrand> soren: for beta yeah
[15:16] <soren> jdstrand: Oh well.
[15:17] <jdstrand> it'll get in after
[15:18] <soren> zul: do you happen to know the state of Xen dom0 in Maverick?
[15:19] <Lichte> damn, am I going to have to go back to 9.10 /
[15:19] <zul> soren: non existant
[15:19] <zul> soren: its something you will have to coble yourself
[15:20] <smoser> Daviey, euca-add-user is broken ?
[15:20] <Daviey> smoser: Depends..
[15:20] <Daviey> smoser: I think it is working on the CLC
[15:21] <smoser> ah.
[15:21] <soren> zul: Lovely.
[15:22] <soren> zul: Do you happen to know how Debian's Xen dom0 support is?
[15:22] <zul> soren: better than ours i think
[15:23] <daxroc> Afternoon all
[15:23] <ttx> hallyn: about bug 599342, do you think we can have a fix for the next hours (in time for Beta) or should I unmilestone it ?
[15:23] <daxroc> Is the cloud version of ubuntu commercial or is there a free version?
[15:24] <daxroc> or is it that you pay for the Management console ( landscape )?
[15:25] <hallyn> ttx: we have a fix, we're waiting for confirmation that it fixes the bug reporter's original problem
[15:25] <hallyn> i'll ping him one more time i guess
[15:25] <ttx> daxroc: we don't do commercial. Our cloud "version" is free and open source
[15:26] <hallyn> i suppose i'd argue that if noone else reports it, and he doesn't care for the fix, then we just leave as is?
[15:26] <ttx> hallyn: if it makes it before the euca upload/build/respin, then it'll be in beta, otherwise we'll fix it between beta and finalfreeze.
[15:27] <ttx> I guess we won't respin beta candidate just for this one.
[15:27] <hallyn> ttx: well the fix is there - should i jsut be proposing it for merge?
[15:29] <ttx> hallyn: it fixes the bug as far as you reproduced it ?
[15:30] <ttx> hallyn: looking at it, I think it can wait post-beta.
[15:30] <hallyn> ttx: i didn't reproduce the bug.  it requires mucking with my dns.
[15:30] <ttx> ok then we'll wait for OP comment, and fix it post-beta. Unmilestoning it.
[15:32] <ttx> Daviey: that makes your eucalyptus upload the only expected respin, as things stand.
[15:32] <ttx> so, the earlier, the better.
[15:32] <hallyn> ttx: for my education, what does 'a respin' entail, in particular on Daviey's part?
[15:32] <Lichte> I just installed mysql-server and client on 10.04 Server and set the root password during setup, but I get this when I try 'mysql -u root -p' :  ERROR 1045 (28000): Access denied for user 'root'@'localhost' (using password: YES) ;   How do I get logged into mysql ???
[15:33] <Daviey> ttx: Hmm.. Well - it could be defered to post-beta.. Whilst I really would like as many of the small fixes in beta, it's not gonna kill the show.
[15:33] <ttx> hallyn: to make it on the ISO, packages uploaded during the freeze need acceptation by release team, building, publishing. Then a new ISO must be built (ISO must be respun), then published to the tracker
[15:34] <ttx> the whole process can take a few hours
[15:35] <ttx> Daviey: how many of those would qualify as regressions ?
[15:35] <Daviey> ttx, possibly 0
[15:35] <ttx> I was mostly concerned by bug 617053
[15:35] <ttx> being an upgrade issue
[15:36] <ttx> fixreleased ?
[15:36] <Daviey> good point.
[15:36] <Daviey> ttx: Simple fix, but blocked on a more generic upgrade issue.
[15:37] <ttx> Daviey: are we expecting upgrades to work with current ? with next upload ? Or after that ?
[15:39] <Daviey> ttx next upload.
[15:39] <ttx> ok, then I'd like to get that new euca in.
[15:39] <Daviey> Whilst i'm not convinced many will be doing the upgrade initially, i'd really like that to be rock solid pre-beta
[16:12] <ecelis> z/win 3
[16:28]  * Daviey wonders if he is still "here"
[16:29] <Zikey> what is the best way to make a perfect clone (identifier+mbr+partition table+data) of a disk on ubuntu ?
[16:31] <Zikey> dd or ddrescue ?
[16:32] <patdk-wk> they do the same thing :)
[16:32] <patdk-wk> ddrescue is just for broken disks
[16:39] <tynot> softball ? syntax to promote user to admin, por favor?
[16:40] <ScottK> ttx: Thanks.
[16:41] <ttx> hggdh, smoser, kirkland, daviey: you have a few old TODOs on the https://blueprints.launchpad.net/ubuntu/+spec/server-maverick-dailytriage spec that you should mark DONE or POSTPONED
[16:41] <ttx> mathiaz: that was my fear with those "tracking-only" specs: that they are not up to date and they corrupt the rest of the charts and completion rates
[16:42] <ttx> adding work for the only benefit of getting the regular task in the same common list
[16:45] <zul> smoser: where is the ttylinux images again?
[16:45]  * Daviey screams at slow t'internet today.
[16:45] <Daviey> this is so painful.
[16:45] <smoser> zul, http://smoser.brickies.net/ubuntu/ttylinux-uec/
[16:46] <tynot> anyone? realize it's a rudimentary question, but I've forgotten the syntax to promote a user to admin. help, please?
[16:47] <mathiaz> ttx: hm - I wonder whether it helped people to actually do them
[16:47] <mathiaz> ttx: ie: does the fact that they show up on the task list made them being worked on
[16:47] <sherr> tynot: if you mean, add them to the "admin" group - see : man addgroup
[16:47] <tynot> never mind. got it.
[16:47] <mathiaz> ttx: if so I'd argue that the experiment was good
[16:47] <tynot> sherr, thx.
[16:48] <mathiaz> ttx: I'd argue that getting things done is more important than reporting being accurate
[16:48] <tynot> needed to give user on linux install admin priviledges is all.
[16:48] <ttx> mathiaz: the experiment continues :)
[16:48] <mathiaz> ttx: we shuld work on improving the reporting though
[16:48] <mathiaz> ttx: *should*
[16:48] <mathiaz> ttx: so that it's *also* useful on the reporting front
[16:49] <ttx> mathiaz: I'd argue that weekly tasks don't mix that well with tasks that don't have to occur on a specific day
[16:49] <ttx> (especially weekly tasks that you can potentially miss)
[16:49] <tynot> sherr, btw, what's the syntax here to address someone (in red) like you've just done to me. irc newbie.
[16:50] <mathiaz> ttx: true - tasks that should happen on a specific day should go on the calendra
[16:50] <mathiaz> ttx: IIRC jiboumans created an all event for weekly summaries in our calendar
[16:50] <mathiaz> ttx: that's another way to do it
[16:51] <ttx> mathiaz: it kinda conflicts with the event I already have on Monday to take care of it :)
[16:51] <mathiaz> ttx: lol - same here
[16:51] <ttx> but yes, calendar events are the way to track weekly-that-you-can-miss type of activity
[16:53] <qman__> anybody know of a good, recent samba performance tuning guide? everything I'm finding is for very old versions and assumes old network technology
[16:55] <qman__> my whole network is gigabit and the servers in question have RAID, so I'm dealing with significantly higher speeds than the "Windows 95 and NT" setups mentioned in the guides
[16:57] <_ruben> hmm .. wonder if i should look into the same, or just be happy with the like 300Mbps or so throughput i currently have
[16:58] <hallyn> kirkland: are there kvm backports for hardy?  (i don't see them in the ubuntu-virt ppa)
[16:59] <siretart> hm. I'm trying to create a new VM with virt-manager/libvirt both on maverick, but I get an "libvir: Security Labeling error : internal error error calling aa_change_profile()"
[16:59] <siretart> (typos copied)
[16:59] <siretart> is this a known issue?
[16:59] <siretart> btw, I see no apparmor warnings in /var/log/kern.log
[17:00] <qman__> and of course we have gigs of RAM where they had megs at the time, so throwing another 32k at a buffer here or there is wholly insignificant to the rest of the system, but might speed up transfers
[17:00] <sbeattie> siretart: by any chance, do you have auditd running?
[17:00] <smoser> ttx, do you know of any reason why i should not start testing ec2 20100830 ?
[17:00] <smoser> err.. 20100831
[17:01] <sbeattie> (it would cause apparmor messages to not appear in dmesg/kern.log out, but rather /var/log/audit/audit.log if you do)
[17:03] <siretart> sbeattie: no, there is no process running with that name. how to start it?
[17:04] <siretart> ah, installing the package auditd seems to help
[17:04] <sbeattie> siretart: no worries, it's not installed by default, wanted to make sure it was the reason you weren't seeing messages.
[17:04] <sbeattie> err, it wasn't the reason.
[17:05] <siretart> ah, now I'm seeing these two messages in the audit.log:
[17:05] <siretart> type=ANOM_PROMISCUOUS msg=audit(1283270700.144:20): dev=vnet0 prom=256 old_prom=0 auid=4294967295 uid=0 gid=0 ses=4294967295
[17:05] <siretart> type=ANOM_PROMISCUOUS msg=audit(1283270700.152:21): dev=vnet0 prom=0 old_prom=256 auid=4294967295 uid=118 gid=127 ses=4294967295
[17:05] <siretart> can someone translate this to english? ;-)
[17:07] <qman__> vnet0 went in and then out of promiscuous mode
[17:08] <siretart> and this is something that is not allowed? I want to create a new virtual machine, and eventually start it
[17:10] <jdstrand> siretart: those messages aren't denials. apparmor denials will still have the familiar 'DENIED' in the output
[17:11] <jdstrand> siretart: it is just letting you know that the process with uid=118 and gid=127 (presumably libvirt-qemu:kvm) put the vnet0 interface in promiscuous mode
[17:12] <jdstrand> siretart: do you have the domain xml for the machine you are trying to create?
[17:13] <siretart> jdstrand: I'm using maverick's virt-manager to create the machine
[17:13] <siretart> I guess the xml is generated on the fly
[17:13] <jdstrand> siretart: what is the name of the vm?
[17:13] <siretart> 'debtest'
[17:14] <jdstrand> siretart: can you paste the output of 'virsh dumpxml debtest'?
[17:14] <jdstrand> it should already be defined
[17:14] <jdstrand> (since the failure is in trying to start it)
[17:14] <siretart> no, it has not been defined yet. this is the step that fails
[17:14] <siretart> virsh list --all doesn't list any machine
[17:14] <jdstrand> siretart: are you sure? 'virsh list --all' should show it
[17:14] <jdstrand> hmm
[17:15] <jdstrand> maybe virt-manager tries to be smart and undefines on failure...
[17:15] <jdstrand> siretart: does the libvirt-qemu:kvm user have write permissions to the directory that you are trying to put the disk image?
[17:16] <siretart> jdstrand: yes, and the image has been created successfully. I've done 'sudo chmod 1777 /var/lib/libvirt/images'
[17:17] <siretart> it's owned by root:root
[17:17] <jdstrand> siretart: I'm confused by this statement: "the image has been created successfully", yet you don't have xml?
[17:17] <siretart> exactly
[17:17] <siretart> 304214    4 drwxrwxrwt   2 root     root         4096 Aug 30 21:24 /var/lib/libvirt/images/
[17:17] <siretart> 304331    0 -rw-------   1 root     root     8589934592 Aug 30 21:24 /var/lib/libvirt/images/debian.img
[17:18] <siretart> but no output of virsh list --all
[17:18] <jdstrand> siretart: you created /var/lib/libvirt/images/debian.img yesterday?
[17:18] <siretart> yes, I did. via virt-manager connecting to 'localhost'
[17:19] <jdstrand> siretart: what does /var/log/libvirt/qemu/debtest.log have to say?
[17:19] <siretart> the unusual part of this setup is that this is actually a kvm virtual machine, that I'm using for my maverick tests. I wanted to see what improvements mavericks brings in the libvirt camp, so I've installed libvirt-bin and virt-manager and tried to setup a qemu VM
[17:20] <kirkland> zul: yo!
[17:20] <siretart> that logfile contains the commandline to qemu, and "libvir: Security Labeling error : internal error error calling aa_change_profile()"
[17:21] <siretart> (the typo in 'libvir' is actually there)
[17:21] <jdstrand> siretart: I'm not 100% sure virt-manager can be used in this way...
[17:21] <jdstrand> siretart: can you paste the full command line?
[17:21] <siretart> LC_ALL=C PATH=/usr/local/sbin:/usr/local/bin:/usr/bin:/usr/sbin:/sbin:/bin QEMU_AUDIO_DRV=none /usr/bin/qemu -S -M pc-0.12 -no-kvm -m 496 -smp 1,sockets=1,cores=1,threads=1 -name debtest -uuid 1e6f1921-e986-84c3-9d4e-9c52c20fa7be -nodefaults -chardev socket,id=monitor,path=/var/lib/libvirt/qemu/debtest.monitor,server,nowait -mon chardev=monitor,mode=readline -rtc base=utc -no-reboot -boot d -drive file=/var/lib/libvi
[17:21] <siretart> rt/images/debian.img,if=none,id=drive-ide0-0-0,format=raw -device ide-drive,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0 -drive file=/var/lib/libvirt/images/debian-testing-i386-netinst.iso,if=none,media=cdrom,id=drive-ide0-1-0,readonly=on,format=raw -device ide-drive,bus=ide.1,unit=0,drive=drive-ide0-1-0,id=ide0-1-0 -device rtl8139,vlan=0,id=net0,mac=52:54:00:ac:04:8e,bus=pci.0,addr=0x3 -net tap,fd=27,vlan=0,nam
[17:21] <siretart> e=hostnet0 -chardev pty,id=serial0 -device isa-serial,chardev=serial0 -usb -device usb-tablet,id=input0 -vnc 127.0.0.1:0 -vga cirrus -device AC97,id=sound0,bus=pci.0,addr=0x4 -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x5
[17:21] <siretart> jdstrand: in what way? connecting to localhost and create a qemu VM?
[17:22] <jdstrand> siretart: importing an existing image
[17:22] <jdstrand> it might, I just don't know
[17:22] <siretart> well, the gui offered me that
[17:22] <jdstrand> ok
[17:22] <siretart> I can also delete and retry it, but I do remember that the error message was exact the same the first time I've tried that
[17:23] <jdstrand> siretart: can you paste the following: cat /etc/apparmor.d/libvirt/libvirt-1e6f1921-e986-84c3-9d4e-9c52c20fa7be*
[17:23] <siretart> sure: cat: /etc/apparmor.d/libvirt/libvirt-1e6f1921-e986-84c3-9d4e-9c52c20fa7be*: No such file or directory
[17:24] <siretart> root@siretart-desktop:/var/log/audit# find /etc/apparmor.d/libvirt/ -ls
[17:24] <siretart> 656106    4 drwxr-xr-x   2 root     root         4096 Aug 30 21:20 /etc/apparmor.d/libvirt/
[17:24] <siretart> 656107    4 -rw-r--r--   1 root     root          164 Aug 25 21:44 /etc/apparmor.d/libvirt/TEMPLATE
[17:27] <jdstrand> siretart: my feeling is this suggests no xml and virt-manager/libvirt is not erroring out in the right place and we aren't getting a useful error
[17:27] <jdstrand> siretart: can you file a bug with exact steps to reproduce. please file it against virt-manager for now
[17:27] <jdstrand> siretart: if it is somewhere else, we can move it
[17:28] <siretart> instructions are easy: install maverick, install virt-manager and libvirt-bin, try to install debian squeeze via netinst iso
[17:28] <jdstrand> siretart: you already said that you did something 'different' though
[17:28] <jdstrand> siretart: you used an existing image
[17:29] <siretart> I'm removing it and try with a new one
[17:29] <jdstrand> siretart: we need all those differences. virt-manager is confirmed to work with an iso install
[17:30] <siretart> yepp, I've now tried with creating a new image, exactly the same problem
[17:31] <jdstrand> siretart: please file a bug
[17:38] <kirkland> hallyn: there's a kvm-84 for hardy
[17:39] <kirkland> hallyn: we can't backport much beyond that as we dropped the dkms kernel module from the kvm package build
[17:39] <kirkland> hallyn: i know our kernel team are supposed to be publishing backported kernels soon
[17:40] <siretart> jdstrand: Bug #627514
[17:40] <kirkland> hallyn: with those, we could backport qemu-kvm, and depend on a particular kernel version, with a new enough kvm module
[17:40] <kirkland> hallyn: however, i would strongly advise against anyone wanting to run production KVM against 8.04
[17:40] <siretart> jdstrand: I've even included a screenshot of the error message
[17:40] <kirkland> hallyn: strongly suggesting that they either upgrade to, or deploy 10.04
[17:40] <kirkland> hallyn: since we have a newer LTS, with a production-ready hypervisor
[17:41] <kirkland> hallyn: 8.04's KVM was tech-preview, at best
[17:43] <kirkland> hallyn: and the backport package is visible at https://edge.launchpad.net/ubuntu/+source/kvm
[17:43] <kirkland> hallyn: that's in the official backports repo
[17:44] <jdstrand> siretart: thanks
[17:44] <siretart> jdstrand: anything else I should add to the bug while I have the machine still on?
[17:45] <jdstrand> siretart: not at this time, thanks
[17:59] <hallyn> kirkland: i've got a testbed set up, meanwhiel i made the suggestion
[17:59] <hallyn> kirkland: i'll be honest, the kvm bug backlog is my primary source of stress and dismay these days
[18:05] <mathiaz> ttx: heya
[18:05] <mathiaz> ttx: was wondering about your findings about hudson?
[18:07]  * SpamapS <heart> CI systems.
[18:08] <mathiaz> SpamapS: do you know of other CI systems?
[18:12] <SpamapS> mathiaz: CruiseControl is sort of the sad cousin of Hudson. ;)
[18:12] <kirkland> hallyn: ack
[18:12] <kirkland> hallyn: it's a PITA
[18:13] <hallyn> kirkland: also a PITA?  is launchpad pinning my cpu since last update
[18:13] <hallyn> $*&%(*$&%(*
[18:14] <kirkland> hallyn: launchpad?
[18:14] <kirkland> hallyn: as in, rendering a web page?
[18:15] <hallyn> yup
[18:15] <hallyn> just started this morning
[18:17] <hallyn> kirkland: would server team meeting today be a good time to bring up the question of updating lucid kvm to 0.12.5 or something?
[18:17] <hallyn> kirkland: (that would, for instance, solve bug 574665)
[18:17] <kim0> Hi folks .. I can't make today's meeting, since I have a committment. There's no real updates from my side since I'm mostly just polishing the cloud portal
[18:18] <kuttan_> hi is there any channels discussing snmp / oids thanls
[18:18] <kuttan_> thanks
[18:18] <kirkland> hallyn: you would have to upload 0.12.5 to lucid-backports
[18:18] <kirkland> hallyn: you can't do a major version bump in an LTS
[18:19] <kirkland> hallyn: but that's fine by me (uploading 0.12.5 to lucid-backports)
[18:19] <hallyn> should i ask if there are objections at team mtg, or just do it?
[18:20] <mathiaz> jjohansen: hi!
[18:20] <jjohansen> mathiaz: hi
[18:21] <mathiaz> jjohansen: is it normal that linux-virtual installs 124M of modules?
[18:21] <jjohansen> mathiaz: define normal?
[18:21] <mathiaz> jjohansen: one of the goal of creating the linux-virtual package was to have stripped down kernel
[18:22] <mathiaz> jjohansen: in lucid linux-virtual modules would take less than 40 MB
[18:22] <jjohansen> mathiaz: its a known issue, basically when -virtual was split off of server it lost its pared down modules config and I need to go in and trim again
[18:22] <mathiaz> jjohansen: ok great
[18:22] <mathiaz> jjohansen: is there a bug about that alread?
[18:22] <jjohansen> yeah, just a sec
[18:23] <jjohansen> mathiaz: Bug #621175
[18:24] <mathiaz> jjohansen: great - thanks
[18:52] <yann2> is it possible to use scp in combination with sudo? ie: i want to copy via ssh files from a remote server to my local server, the remote files belong to root:root and are readable only by root, and the root login is disabled via ssh
[18:52] <ttx> mathiaz: I just looked quickly into it. It's quite basic but pushes towards best practices. I'd compare it to Nagios
[18:53] <ttx> i.e. no black magic, but does what it should do, and well.
[18:53] <mathiaz> ttx: but in a different domain right?
[18:53] <yann2> is there anything shorter than copying the files to another location on the remote server, chmod them, copy them, move them again, and rechmod them?
[18:53] <ttx> mathiaz: oh yes.
[18:54] <mathiaz> ttx: where you looking at hudson as a product to package or to use in the team?
[18:54] <ttx> mathiaz: more as part of my "staying current" objective. I already looked at it from a packaging perspective
[18:55] <hallyn> yann2: so you want the target files owned by root:root?  Only way I can think of would be to do it as root from target machine, scp'ing with the source being remote :)
[18:55] <mathiaz> ttx: right - I'm more interested as a CI to use in the server team
[18:55] <mathiaz> ttx: I've just refactored my iso testing scripts
[18:55] <ttx> mathiaz: it's your classic Java thing: too many deps to package, compared to the ease of use of "just" deploying the prebuilt wart
[18:55] <ttx> war
[18:55] <mathiaz> ttx: and I may look at using hudson to track all the iso testing
[18:55] <mathiaz> ttx: now that I'm able to fully automate tests
[18:56] <yann2> hallyn, yeah, but then I couldn't scp them to the folder I want, as the normal user wouldnt have access to there :)
[18:56] <ttx> mathiaz: Hudson allows CI, but you can use to track any result
[18:56] <ttx> s/use/also use it/
[18:56] <yann2> also gets complicated with ssh agents, if I sudo -s to root on the remote host, I lose my agent connection
[18:56] <mathiaz> ttx: right - does it allow to attach random piece of  information to results?
[18:56] <hallyn> you do what?
[18:57] <mathiaz> ttx: like installation logs?
[18:57] <ttx> mathiaz: there is an interface that allows you to post a result, I think you could attach anything to it
[18:57] <hallyn> yann2: so i think the answer is "no, bc otherwise you lose any safety of disabling root remote logins in the first place :)
[18:58] <yann2> well not if the user has sudo anyway...
[19:12] <kirkland> hallyn: fyi, https://help.ubuntu.com/community/UbuntuBackports
[19:14]  * RoyK just wrote a perl thing to find duplicate files in a directory hierarchy - nice in case you have tons of mp3s or other large files on a chaotic filesystem
[20:07] <Krazyderek> looking for a good way to backup 12 google apps accounts, including sent mail
[20:08] <Krazyderek> is postfix the best solution? or do i want something a bit more custom to dealing with google accounts?
[20:08] <RoyK> fetchmail?
[20:11] <Krazyderek> @RoyK have you used it for a while?
[20:11] <RoyK> not with google
[20:13] <Krazyderek> hmmm i guess i'd like something that will download evethying at a set time, say 3am, to keep a backup of google, and in the event someone blows up google or something, that i could just install thunderbird on everyone's desktop and just point it at our linux box downstairs and keep on chuggin' along
[20:16] <RoyK> Krazyderek: fetchmail can feed the mail into an MTA like postfix
[20:19] <Krazyderek> RoyK: sorry MTA?
[20:19] <kirkland> smoser: ping
[20:19] <kirkland> smoser: have you gotten the ttylinux image running in UEC/Maverick yet?
[20:20] <smoser> i think so , yeah
[20:20] <kirkland> smoser: if not, do you have a running UEC where you can try it?
[20:20] <kirkland> smoser: can you check?
[20:20] <smoser> http://smoser.brickies.net/ubuntu/ttylinux-uec/
[20:20] <kirkland> smoser: i'm trying to run it in a vm
[20:20] <kirkland> smoser: ie, in qemu alone
[20:20] <smoser> it should be fine.
[20:22] <kirkland> smoser: hrm, it's not coming up ...
[20:22] <kirkland> smoser: well, i used a really old one
[20:22] <kirkland> smoser: do you have console working with these new images?
[20:22] <RoyK> Krazyderek: mail transport agent
[20:22] <smoser> what is console ?
[20:22] <RoyK> Krazyderek: something like postfix or sendmail or whatever
[20:22] <smoser> remember that maverick euca-console-output was recently broken.
[20:22] <smoser> i'm not sure if its fixed now or not
[20:23] <kirkland> smoser: euca-get-console
[20:23] <kirkland> smoser: so that i can see where it's stuck, if it's stuck
[20:23] <kirkland> smoser: what's the username/password in that image?
[20:23] <kirkland> smoser: root/linux ?
[20:23] <smoser> i dont know.
[20:24] <Krazyderek> RoyK: ah so would i setup both now? or just fetchmail? looks like i'd be using ETRN or ODMR modes and both are unfamiliar to me
[20:25] <RoyK> fetchmail can download your email easily using imap or pop3 and feed it to postfix, which can store it a dovecot store or something
[20:25] <smoser> i just verified: wget http://smoser.brickies.net/ubuntu/ttylinux-uec/ttylinux-uec-i686-11.2_2.6.35-16_3.tar.gz -O out.tar.gz && uec-publish-tarball out.tar.gz foobucket i386 && euca-run-instances --key mykey $AMI
[20:25] <smoser> and reached the instance fine.
[20:25] <smoser> consoel output is present.
[20:26] <RoyK> Krazyderek: postfix doesn't do stuff like imap or pop, it's a plain MTA, but a jolly good one. Dovecot can do the storage, as a POA (post office agent)
[20:27] <Krazyderek> RoyK: i have an ubuntu server book here for the postfix and devecot stuff, and i can probably feel my way through that for one user, it's just keeping everything in order for all 12 people that worries me
[20:28] <kirkland> smoser: you rock, thanks
[20:28] <RoyK> Krazyderek: dovecot can easily handle quite a few users
[20:28] <RoyK> thousands
[20:29] <Krazyderek> RoyK: but how does fetchmail?
[20:30] <RoyK> fetchmail is just run with a cron job
[20:30] <RoyK> it fetches mail, basically
[20:30] <RoyK> for 12 users, that won't be much of a problem
[20:31] <Krazyderek> hmm
[20:31] <Krazyderek> RoyK: but i do need it, right? postfix can't do this for me?
[20:31] <Krazyderek> RoyK: sorry first mail server
[20:32] <cloakable> postfix won't fetch mail for you, mail needs to be sent to it.
[20:32] <RoyK> afaik postfix can't do pop or imap, so it can't aquire email from another host
[20:32] <cloakable> postfix does smtp only, yea
[20:33] <RoyK> yes, as cloakable said, use something like fetchmail to download the stuff
[20:33] <pmatulis> Krazyderek: consider setting up a typical mail server and then have google accounts send a copy of all mail to it
[20:33] <RoyK> feed it into postfix, and have postfix send it to dovecot (or something)
[20:34] <RoyK> pmatulis: give you have a static IP, that's probably the best
[20:34] <cloakable> Unless your isp has a smarthost, of course :)
[20:35] <pmatulis> cloakable: smarthost is for outgoing mail, not incoming
[20:35] <Krazyderek> we'd have to pay at least $15 a month extra for a static
[20:35] <RoyK> Krazyderek: then try fetchmail
[20:36] <pmatulis> Krazyderek: i don't understand, you're worrying about google blowing up but you can't afford a static IP?  ;)
[20:36] <Krazyderek> hmmm fetchmail has I9 in their FAQ saying it doesn't play to well with gmail as of April 2008
[20:36] <cloakable> Krazyderek, pmatulis: Which would be a problem if your MX record was an ip address :P
[20:37] <RoyK> fetchmail -> postfix -> dovecot
[20:37]  * cloakable receives mail on a dynamic ip
[20:37] <pmatulis> Krazyderek: you can easily use the dynamic ip checker method - works very well
[20:38] <cloakable> Mmmm
[20:38] <JasonMSP> already did a google search.  Im unfamiliar with terminology so my search isn't great.  Im trying to understand SSL in order to configure VSFTPD for secure transfers (so passwords aren't sent in the clear).  I can easily setup the VSFTP.conf options, but I need help with what else to make happen for the connection.  Certificates, etc.. Can someone point me to a clear post that I can read up on?
[20:38] <Krazyderek> i do have a no-ip account for remote desktop and vpn addressing
[20:39] <kirkland> smoser: hrm, well, okay, i just published your latest tarball to my all-in-one-UEC-running-in-a-beefy-KVM
[20:39] <kirkland> smoser: published fine (thanks)
[20:40] <kirkland> smoser: running the instance, it goes to the "running" state no problem
[20:40] <kirkland> smoser: note that i'm using --addressing private
[20:40] <cloakable> kirkland: Use that as your MX record then *shrug*
[20:40] <kirkland> smoser: it stays in the running state (ie, it doesn't commit suicide)
[20:40] <smoser> do you get console output ?
[20:40] <kirkland> smoser: negative
[20:40] <smoser> then you need to fix that first.
[20:40] <kirkland> smoser: been running for ~350 seconds
[20:40] <kirkland> smoser: all i get is the timestamp back
[20:41] <kirkland> smoser: when i euca-get-console-output
[20:41] <smoser> your running into bug 619843
[20:41] <kirkland> smoser: well, instance id, then timestamp
[20:41] <cloakable> Erk
[20:41] <Krazyderek> pmatulis: i'm not sure if the "forward copy" option in google will work for sentmail, so you'd have to BCC mail to yourself and setup some kind of rule to put it in the sent items....errr..
[20:41] <kirkland> smoser: ah
[20:41]  * kirkland goes tackle that nasty bugger
[20:41] <pmatulis> Krazyderek: huh?
[20:42] <smoser> kirkland, that is probably supposed to be fixed
[20:42] <smoser> Daviey, its not fix released ?
[20:43] <Krazyderek> pmatulis: when you were saying setup a regular mail server and have google send a copy
[20:43] <pmatulis> Krazyderek: ok, it won't work?
[20:43] <Krazyderek> pmatulis: not for sent mail i don' think
[20:44] <pmatulis> Krazyderek: not sure why
[20:44] <Krazyderek> pmatulis: i thought the forwarding only appied to incoming mail
[20:45] <RoyK> http://karlsbakk.net/fun/sinking.mpg
[20:45] <Krazyderek> pmatulis: i'll try it now
[20:48] <kirkland> Daviey: what change did you make?  did you just add eucalyptus to the kvm group?
[20:49] <Krazyderek> pmatulis: ya incoming only
[20:49] <Krazyderek> i want to have a record of all sent mail backed up as well, so hopefully fetchmail is working better with google these days :S
[20:54] <Krazyderek> RoyK: i don't see a listing for postfix under the fetchmail website, do you have some recommended settings or a reliable tutorial?
[20:55] <kirkland> smoser: sweet!
[20:55] <smoser> what'd you find ?
[20:55] <kirkland> smoser: i've got console output, and i can see my failure :-)
[20:55] <kirkland> smoser: you da man
[20:55] <smoser> what is fialure ?
[20:55] <kirkland> smoser: well, i have a mismash of network hackery i need to sort out
[20:55] <kirkland> smoser: basically, i need to fix my networking setup in eucalyptus.local.conf
[20:56] <kirkland> smoser: but the bugger is most definitely running!
[21:00] <JasonMSP> okay lets ask that question again in a different way.  Can anyone point me to a primer on creating secure connections with ubuntu.  like the ultimate basics so that i understand how the process works and then can apply it to VSFTP?
[21:01] <JasonMSP> (im looking at TLS in wiki now)
[21:01] <qman__> secure and FTP are mutually exclusive items
[21:02] <qman__> you can sandbox it off in its own restricted corner, but you cannot secure FTP
[21:02] <JasonMSP> qman: vsftp from what I understand can be configured with TLS/SSL.  But again I really don't have any idea what im doing.
[21:03] <qman__> if it can, it's a new feature that I am unfamiliar with
[21:03] <qman__> I use the sftp built into openssh
[21:04] <qman__> as far as FTP daemons go, VSFTP is the best I have used, but it's still FTP
[21:04] <JasonMSP> Im working on the ability for clients to FTP into their sites.  Each has a user id with which they can FTP, SSH is disabled so they can't get in to the server, and they are chrooted to their web.com folder.
[21:05] <JasonMSP> Im still weighing if it is necessary to secure the connection, but i don't like the idea of passwords being sent in the clear.
[21:06] <JasonMSP> http://wiki.vpslink.com/Configuring_vsftpd_for_secure_connections_%28TLS/SSL/SFTP%29
[21:08] <qman__> centos, yuck
[21:08] <qman__> good information though
[21:08] <qman__> should be the same on ubuntu swapping yum for apt-get
[21:09] <qman__> and /etc/rc.d/init.d for the service command
[21:09] <JasonMSP> yes..  I tried this last week and didn't get it working.  It may be the clients im working with to get in which is why I need the bare basics to undersatnd the process.  I know how to create the certificates, but then the client connecting doesn't work with what I've tried so far.
[21:10] <JasonMSP> ive been using sudo service vsftpd restart
[21:10] <qman__> well, that's what's known as FTP/S, not SFTP
[21:10] <qman__> so make sure your clients are in the right mode
[21:10] <JasonMSP> ok then where can I get basics on the difference between all those.
[21:10] <qman__> with TLS it should still use port 21, too
[21:11] <qman__> SFTP is FTP over SSH
[21:11] <qman__> FTP/S is FTP with TLS encryption
[21:11] <JasonMSP> ok.  so then im looking for FTP/S...  Thanks!
[21:11] <qman__> a number of other protocols work like that too
[21:12] <qman__> TLS uses the standard, clear port and then starts encryption
[21:12] <qman__> where SSL uses a separate port, like how HTTPS uses 443
[21:12] <JasonMSP> does that mean password will still be sent in the clear?
[21:12] <qman__> no
[21:12] <qman__> the initial connection handshake is, then the STARTTLS command is sent
[21:12] <qman__> and the rest continues like SSL
[21:13] <JasonMSP> ah
[21:13] <JasonMSP> (i hear the angels signing in the background)
[21:13] <JasonMSP> *singing
[21:15] <qman__> now, I don't know if FTP/S transfers the files over the encrypted channel, or what other ports, if any, it uses, you'll have to look that up
[21:15] <qman__> but I do know that the authentication and control are handled over the secured connection
[21:16] <Krazyderek> RoyK: i sink i'm going to tackle this one tomorow, thanks for the tip i'll see how it goes on a small account first
[21:16] <Krazyderek> exit
[21:17] <JasonMSP> Does SFTP require users to have shell access?
[21:18] <qman__> no
[21:18] <qman__> it uses sshd, but the users can be easily denied shell access
[21:18] <qman__> however, the users must have a valid shell defined
[21:20] <JasonMSP> ok so it would be just as easy for them to SFTP.  ive read that an option is /usr/sbin/nologin   ??
[21:21] <JasonMSP> (easy because the client software is easily come by)
[21:22] <qman__> lots of clients support SFTP, on many platforms
[21:22] <qman__> however, I don't think nologin works
[21:22] <qman__> I can test and verify
[21:22] <qman__>  /bin/false does not work
[21:23] <JasonMSP> i've got my ssh config file setup to allow only those in the group sshlogin
[21:23] <JasonMSP> so that should secure anyone else from shell access.
[21:25] <qman__> nologin does work
[21:26] <qman__> a bit surprising, guess I'm making some changes
[21:27] <JasonMSP> so as long as their shell folder is set with nologin, they don't have ssh access (because they are not in the group) they should only be able to access their folder.  What about CHROOTing them to their web folder, this should still be accomplished by VSFTPD
[21:28] <JasonMSP> (their folders via FTP that is)
[21:28] <qman__> VSFTPD has nothing to do with SFTP using sshd
[21:28] <qman__> they are completely separate
[21:29] <qman__> whichever one you are using, you need to configure chroot with
[21:29] <qman__> or both, if using both
[21:30] <JasonMSP> VSFTPD is not needed if you are doing SFTP.  ok.  Is this because the client software acts as if it is FTPing, but in reality it is being done through SSH?  As you can see im getting lost on the fundamentals which is what has been making this difficult.
[21:31] <qman__> SFTP is an internal function of sshd
[21:31] <qman__> it works like FTP but is built in
[21:31] <JasonMSP> ie its built in.
[21:32] <qman__> you can use other SFTP subsystems with sshd, but the internal one is the one I know and use with sftp-only users and chroots
[21:33] <JasonMSP> im on my third generation of working this solution. the first was straight VSFTP, then I was recommend SFTP and when I couldn't get either of those working I went back to VSFTPD.  I have a working in the clear setup with VSFTPD now.  I never was able to get a working SFTP solution (except for myself via key)  I've got passwords turned off in SSHD config.
[21:33] <qman__> vsftpd cannot be used as an SFTP subsystem, because it's not SFTP, it's FTP/S
[21:34] <JasonMSP> i understand.  they are separate (daemons?)
[21:34] <qman__> well, if you have password authentication disabled, SFTP users will need keys
[21:34] <qman__> a very secure configuration, but you may have trouble instructing users to use it
[21:35] <JasonMSP> EXACTLY!
[21:35] <JasonMSP> which is why i thought I would be able to use VSFTPD.
[21:35] <qman__> you can, but it will be FTP/S, not SFTP, and will require a different client mode
[21:35] <qman__> and I'm not sure on the specifics of how that one works
[21:36] <JasonMSP> client mode meaning protocol?
[21:36] <qman__> yes
[21:37] <qman__> also, FTP/S is not really standardized, some softwares implement it differently
[21:39] <JasonMSP> ok.  I just used WinSCP to connect in the clear with a password which means it is using the VSFTPD daemon
[21:41] <JasonMSP> that means that it is possible then to secure this connection with TLS if that link I sent earlier configures it correctly.
[21:42] <qman__> that page states that winSCP doesn't do FTP/S
[21:42] <qman__> I can neither confirm nor deny
[21:53] <WinstonSmith> try filezilla
[21:54] <WinstonSmith> does FTP/S i think
[22:01] <JasonMSP> you can select the protocol (FTP) and then it gives you a dropdopwn for encyption
[22:02] <JasonMSP> then you can choose none, SSL/TLS implicit, SSL explicit, or TLS explicit
[22:03] <JasonMSP> I used TLS explicit and it connected with the right certificate but they did not end up in the correct directory.  so im thinking its a config setting
[22:03] <Zikey> I have a file named "core" at the root (/), it's probably a coredump, is there any tool to know more about it ?
[22:04] <Patrickdk> only if you care to debug it
[22:05] <Zikey> I just would like to know which binary crashed
[22:05] <Zikey> it's probably written in the core file
[22:05] <Patrickdk> gdb I believe
[22:05] <Patrickdk> been a few years since I worked on one
[22:06] <Zikey> strings <core worked :)
[22:06] <Zikey> damn asterisk...
[22:09] <Zikey> last question, how do you force a fsck on next reboot ?
[22:18] <soren> Zikey: touch /forcefsck
[22:18] <soren> iirc.
[22:19] <soren> Yup.
[22:28] <Zikey> thx soren !
[22:33] <Kaelten> anyone have an example multipath.conf file they could point me at?
[23:02] <tucemiux> anyone knows how to configure a printer server manually?
[23:11] <DigitalDeviant> Would anyone here be helpful to help me setup a bind9 dns server on my ubuntu server? I have tried to follow multipe howto's and even webmin but i just can seem to get it working correctly. If someone could do a remote session with me and show/explain to me how it works I would be very greatfull
[23:13] <tomsdale> DigitalDeviant: sry - no experience here either but did you try howtoforge?
[23:13] <ScottK> DigitalDeviant: Look in the Ubuntu server guide on help.ubuntu.com.
[23:13] <ScottK> It has specifics on how to do it.
[23:14] <DigitalDeviant> yah, i have been looking at docs all day. They are confusing as hell to me. I have no problem setting dns up in windows, lol but since i am new to linux I keep hitting walls
[23:14] <ScottK> Did you try the Ubuntu server guide?
[23:14] <DigitalDeviant> yah, Ive spent a good 6 hours trying to get it to work
[23:14] <tomsdale> Does 10.04, apart from apparmor have a second firewall - I changed my ssh port to 31337 but the connection times out - on 22 it works. I already uninstalled apparmor
[23:15] <ScottK> tomsdale: There is one installed, but not active by default.  I'd recommend putting apparmor back as it won't interfere with that.
[23:17] <tomsdale> what could it be then that won't allow me to connect via ssh? netstat -tap shows it listening on 31337
[23:17] <DigitalDeviant> you behind a hardware firewall tomsdale ?
[23:18] <tomsdale> nop - all in the internal network
[23:19] <DigitalDeviant> try using a port scanner to the ip with the SSH server. see if the port is open from site to site first
[23:19] <tomsdale> already set loglevel to debug but I don't see any request being made in the auth.log. And apparmor is uninstalled as I mentioned.
[23:19] <tomsdale> good idea
[23:19] <DigitalDeviant> are you using fail2ban as well ?
[23:19] <tomsdale> yes, nmap says filtered, ELITE
[23:20] <DigitalDeviant> so you are using fail2ban
[23:20] <tomsdale> yes, but according to the log I'm not jailed
[23:20] <DigitalDeviant> have you added your ip to the allow list in fail2ban
[23:20] <tomsdale> eh - banned.
[23:21] <tomsdale> I don't think it's  my IP the problem - on port 22 all works fine. As soon as I change it to 31337 it stops working
[23:22] <DigitalDeviant> once you change the port, you have restarted the service right? i know stupid question, but sometimes people forget
[23:22] <tomsdale> I even did a good old fix all win problems reboot :-)
[23:22] <DigitalDeviant> lol
[23:22] <DigitalDeviant> ok
[23:24] <tomsdale> There somehow is another firewall actice I have the feeling.
[23:24] <DigitalDeviant> doesnt pure-ftp use port 31337. are you running that?
[23:25] <tomsdale> ah - there is bastille-firewall which is active.
[23:25] <DigitalDeviant> when i run into issues like that, i install webmin so i can look at everything all at once, it helps, then i purge it
[23:26] <tomsdale> that's it. it appears ubuntu server 10.04 uses by default apparmor and bastille-firewall. Two more secure than one ?
[23:27] <DigitalDeviant> i use ddwrt firewall and fail2ban. so im not to familair with software firewalls in ubuntu
[23:27] <sbeattie> um, what? bastille is not included by default.
[23:28] <tomsdale> hm - did a clean install from a howtoforge tut - checking whether it slipped in there somewhere.
[23:29] <DigitalDeviant> do i need multipe IP's to setup a dns server / lamp server ?
[23:30] <DigitalDeviant> i have two public ip's one pointing to dns port and the other to the lamp server. is the correct setup so far?
[23:31] <DigitalDeviant> or do i just use the one ip address for the dns and the lamp server
[23:31] <DigitalDeviant> im only hosting like 4 domain names
[23:32] <DigitalDeviant> neither will have email
[23:32] <fluvvell> I've got 10 domain names, dns and lamp all on the same ip. It really doesn't matter if you handle ports and everything correctly
[23:33] <DigitalDeviant> oh
[23:33] <DigitalDeviant> good then, that gives me ann extra ip :)
[23:33] <fluvvell> but I use fail2ban and shorewall (firewall)
[23:33] <DigitalDeviant> i use fail2ban and a hardware firewall with iptables
[23:33] <kirkland> smoser: still around?
[23:34] <kirkland> smoser: ideas?  http://pastebin.com/Vx24Xn7H
[23:36] <fluvvell> DigitalDeviant: are you wanting bind for external dns or internal network ?
[23:36] <kirkland> smoser: metadata failage
[23:36] <DigitalDeviant> external. My server is behind verizon fios and I need to setup fwd and reverse zones for them to setup the ptr records or something
[23:37] <SpamapS> DigitalDeviant: you want to have your IP resolve to your server's hostname?
[23:38] <SpamapS> DigitalDeviant: you need at least two servers to host your own DNS. Do you have two?
[23:38] <DigitalDeviant> yah, the other one is sitting here right now getting server loaded on it as we speak, but right now on the fios i only have one
[23:39] <DigitalDeviant> brb guys
[23:40] <kirkland> smoser: nevermind, i think i have it!
[23:45] <DigitalDeviant> right now, my domains are using the ns1.verizon.net and ns2.verizon.net per the tech support. So i need to setup the dns server to resolve the domains or someting..ive never done dns on ubuntu so im lost like a 9 year old in a candy store
[23:46] <DigitalDeviant> and the howto's are not helping me a bit
[23:46] <DigitalDeviant> im more of a watch and learn type person, ya know
[23:47] <DigitalDeviant> I installed webmin to help because a frind of mine said it would make it easier....pfft
[23:49] <DigitalDeviant> ive been at this all day with no progress and im about to throw my server out the window
[23:49] <tomsdale> DigitalDeviant: Don't know if it helps your cause but I'm atm installing ispconfig 3 in a 3 server setup. Haven't gotten as far as DNS yet but it seems you can create zones in a web interface.
[23:49] <DigitalDeviant> yah, you can do that with webmin as well
[23:49] <DigitalDeviant> its just too confusing for me
[23:50] <tomsdale> you're using bind?
[23:50] <DigitalDeviant> yah
[23:51] <tomsdale> on howtoforge some people us mydns - have heard it's smaller and easier but the hosters tend to prefer bind.
[23:52] <DigitalDeviant> yah, i was told to use bind
[23:53] <tomsdale> I mean a software with Version Nr. 9 - It gotta be good. Even windows only got to 7 so far :-)
[23:54] <tomsdale> sry - not really helping your problem :-(
[23:56] <tomsdale> http://oreilly.com/catalog/9780596100575  ??
[23:57] <DigitalDeviant> its all good, been going through this crap all day :(
[23:58] <ath88>  Hello, anyone whos nifty with sshtunnels got the time and temper to help me? I need to setup a tunnel from a netbook to my server, so i can access that netbook no matter what network it is logged onto. Is that possible?
[23:59] <DigitalDeviant> well the good news i guess is if im going to setup a second dns server, I might as well host mysql on it as well and get some load of the web server