[03:16] <vorian> what is the easiest way to set up a mail server?
[03:17] <vorian> nvrmind
[03:17] <Delerium_> I'm no pro, but it depends on your infra (fix ip? MX record?  send and receive?
[03:18] <vorian> just something that will send mail when prompted my the right script
[03:18] <vorian> I think what i'm looking for is exim4
[03:19] <Delerium_> I **think** postfix might do the job too
[03:19] <Delerium_> depend on your scripting language?!
[03:20] <vorian> AH, I'll set it up and see
[03:21] <Delerium_> If you just want to send mail, just specify your ISP SMTP server and your mail should be routed correctly
[03:21] <vorian> kk
[03:21] <vorian> thanks Delerium_
[03:22] <Delerium_> mp
[03:22] <Delerium_> errr.. no
[04:49] <_Techie_> DHCP operating on a bridge interface seems to take considerably longer to assign addresses than when listening on a standard interface, is there any reason for this?
[05:11] <stgraber> _Techie_: It's not DHCP itself that takes longer. A bridged interface usually takes around 20s before it can start receiving/transmetting data over the bridge.
[05:12] <gobbe> yes
[05:12] <stgraber> _Techie_: if you go with a static IP instead, you'll notice the same delay until you can reach your bridged interface
[05:12] <gobbe> and bridge is meant to be used in static ip
[05:14] <stgraber> the bridge interface on the host shouldn't have a DHCP address or if you really need it, you should make sure you join a tap device (or similar) with a MAC address lower than all the other bridge members
[05:15] <stgraber> that's because a bridge will always take the lowest MAC address as it's own MAC, so if you don't control what the lowest address is, your bridge will change MAC address making your IP to change when using DHCP
[05:15] <stgraber> there's nothing wrong with using DHCP on a bridged interface though (like a VM, VZ container, ...)
[05:16] <gobbe> yes, it's working, but it's designed to work with static ip
[05:18] <stgraber> a bridge interface is a bridge interface, it's not any more designed to be used with a static or with a DHCP address. It's just simulating an ethernet switch.
[05:19] <stgraber> I agree it's good practice, as it's on a real physical switch, but there's nothing "designed" specifically for static or dynamic addresss.
[05:20] <gobbe> well, not designed in that way, but it's working better with static ip
[05:21] <stgraber> no. Your interface won't come online any faster with static than with DHCP. The only difference is that when using static everything will hang as it believes network is up whereas with dynamic, it'll just wait for the bridge to work and then get an IP and start everything.
[05:21] <_Techie_> my bridge has a static ip
[05:22] <stgraber> in that aspect, DHCP actually works better than static on bridged interface. Still, it's better to have static addresses on your network to avoid depending on a dhcp server, but that's just good practice.
[05:22] <_Techie_> once again, you lot have mis understood what i was saying
[05:23] <_Techie_> my DHCP server, listening on a bridge interface, is taking longer than expected to assign ip's
[05:23] <_Techie_> im not talking about right after boot or anything
[05:24] <stgraber> ah, I never noticed that specific issue (having all my DHCP servers in a bridged VM).
[05:25] <_Techie_> when i was using a single interface, windows wouldnt complain about not being able to mount share due to not being given an ip
[05:27] <stgraber> reading that last sentence makes me wonder if you're not speaking of bonding instead of bridging (single vs multiple interfaces)
[05:27] <gobbe> sorry, i was talking about bonding
[05:28] <stgraber> ok, now it makes sense :)
[05:28] <gobbe> too messed up, havent got yet my breakfast coffee :D
[05:28] <gobbe> sorry, but now i need to go, have to be at office in 45min
[05:29] <stgraber> what kind of bonding are you using ?
[05:29] <_Techie_> http://sprunge.us/TaIQ
[05:30] <stgraber> ouch, yeah, if you do it that way you'll have issues for sure ...
[05:30] <_Techie_> what the appropriate way of doing it?
[05:30] <stgraber> you probably want to use bond0 and the bonding module instead of using bridging
[05:30] <stgraber> http://wiki.debian.org/Bonding
[05:31] <stgraber> your current config basically means: take eth0 and eth1, put them in the same switch and add an IP to it
[05:31] <_Techie_> is there anywhere where i can see the differences between bonding and bridging?
[05:32] <stgraber> which means it creates a loop on your network by linking eth0 and eth1
[05:33] <_Techie_> and what does bonding do?
[05:34] <stgraber> bonding takes the two links and will load-balance the trafic between the two. If one goes down, it'll fallback to using just one.
[05:34] <_Techie_> okay, thats not what i need
[05:34] <_Techie_> both interfaces are not connected to the same set of devices
[05:34] <_Techie_> eth0 is a direct link to my computer, and eth1 is the rest of the network
[05:35] <_Techie_> so by that description, bonding is definitely what im after
[05:35] <_Techie_> not what im after*
[05:37] <stgraber> ok, so bridging is indeed what you want in that case. Basically your server is going to act as a switch with your computer plugged in in one port and the rest of your network in another
[05:37] <stgraber> and has itself an IP in that network
[05:37] <_Techie_> yep
[05:38] <_Techie_> as you can see, i had the two interfaces setup on different networks
[05:38] <_Techie_> which was fine, untill i wanted to use an autodiscovering app across the two
[05:39] <stgraber> I don't see why DHCP would be slow, other than some weird MAC address mangling done by the server (which shouldn't happen)
[05:39] <_Techie_> ill have a stab in the configs and come back
[05:40] <stgraber> I'm about to go to sleep here anyway, getting quite late here.
[05:40] <_Techie_> okay, have a good one stgraber
[05:40] <stgraber> thanks
[07:26] <chrislabeard> Hey guys can anyone help me format a secondary hard drive?
[07:26] <chrislabeard> I'm having a heck of a time
[07:35] <chrislabeard> when i do fdisk -l
[07:35] <chrislabeard> it gives me /sdb1 and /sdb2 partitions when I try to format the sdb2 it says it doesn't exist ?
[07:35] <chrislabeard> Any ideas
[07:38] <pting> chrislabeard, so ls -alh /dev/sdb2 returns nothing?
[07:38] <chrislabeard> ls: cannot access /dev/sdb2: No such file or directory
[07:38] <chrislabeard> I just wanna wipe the drive clean but it just seems like its impossible.
[07:39] <pting> chrislabeard, your box doesn't see it... do an ls \/dev/sdb*
[07:39] <pting> whoops, remove that \
[07:39] <pting> does it return aything?
[07:39] <chrislabeard> gives me /dev/sdb and /dev/sdb1
[07:39] <chrislabeard> So I guess the other partition is gone
[07:40] <chrislabeard> Why does it list it when I do fdisk -l?
[07:40] <pting> chrislabeard, so you should be able to wipe /dev/sdb
[07:40] <chrislabeard> K can you walk me through it.. So I will know the correct way to do it.
[07:40] <chrislabeard> I've tried so many different ways
[07:42] <pting> chrislabeard, so you wan to whipe sdb and format with ext4 or something?
[07:42] <chrislabeard> Yeah, I want to turn this hard drive into a TimeMachine drive so ext4 or 3 whichever
[07:43] <chrislabeard> I can wipe the whole thing
[07:44] <pting> chrislabeard, so you should be able to do sudo mkfs.ext4 /dev/sdb i believe
[07:44] <chrislabeard> pting: k let me try
[07:44] <pting> chrislabeard,  ... make sure that's the one you want to destroy though
[07:44] <chrislabeard> yeah my other drive is sda
[07:45] <chrislabeard> K its running
[07:45] <pting> chrislabeard, i think timemachine might require specific settings, i dunno if macs can read ext4
[07:46] <chrislabeard> if not I can just use that command and re format it
[07:46] <chrislabeard> http://sidikahawa.blogspot.com/2010/03/setting-up-time-machine-server-on-my.html
[07:46] <chrislabeard> that is the tutorial I was going to follow
[07:47] <pting> chrislabeard, http://serverfault.com/questions/190840/mount-a-ext4-partition-on-mac-os-x
[07:48] <chrislabeard> How do I see the mount point of the new drive?
[07:49] <pting> chrislabeard, you should read up on http://en.wikipedia.org/wiki/Fstab ... it's basically done through /etc/fstab
[07:50] <chrislabeard> Really I don't see the drive in there
[07:50] <pting> chrislabeard, but just to get started and to test, you basically make a directory, then type the following sudo mount -t ext4 /dev/sdb /mnt/mydirectory
[07:51] <chrislabeard> k cool I'll add that to the fstab
[07:51] <pting> chrislabeard, you dont' add that line into fstab; fstab has a specific format specified in that link i sent you
[07:52] <chrislabeard> oo
[07:52] <pting> chrislabeard, that's a command you use if you don't want to include it in fstab
[07:52] <chrislabeard> oo okay
[07:53] <chrislabeard> pting: So why ext4 and not ext3
[07:54] <pting> chrislabeard, it's just the newer format; but i don't think macs can natively read ext4, so i would pick another fs for your time machine if it's an issue
[07:54] <chrislabeard> K
[07:54] <pting> chrislabeard, i've never tried time machine, so i don't know anything beyond that
[07:55] <chrislabeard> well thanks for your help up to here
[07:55] <IrishWristwatch> I thought the latest time machine update broke time machine for ubuntu.
[07:55] <IrishWristwatch> Did they finally patch it?
[07:55] <chrislabeard> I don't know this was posted march 2010
[07:55] <chrislabeard> Is it not working now
[07:55] <pting> damn you apple
[07:56] <IrishWristwatch> Not sure, I thought I read somewhere that they slipped an undocumented change in the protocol and it broke time machine for ubuntu somehow.
[07:56] <IrishWristwatch> It could be patched by now.
[07:56] <chrislabeard> Yeah lol I hope so
[07:56] <IrishWristwatch> I really don't know, sorry.
[07:57] <IrishWristwatch> I guess... try it.  :)
[07:57] <chrislabeard> Yeah I'm gonna
[07:58] <chrislabeard> we will see
[07:58] <pting> so how do you perform an unattended install of a package when it asks whether a package's config should be replaced or not? something to do with debconf-get-selections ?
[08:00] <IrishWristwatch> does --assume-yes when running aptitude work?
[08:01] <pting> IrishWristwatch, i want to keep the current config, but i'll check aptitude's params out
[08:16] <Syria> Hi, why .htaccess is not affecting my site? is it because of the permessions?
[08:17] <Syria> When I try to change its permessions to 666 I get an error message.
[08:17] <Syria> -rw-r--r-- 1 root root     205 2011-01-03 08:02 .htaccess
[08:22] <joschi> Syria: check the value for AllowOverride for the directory
[08:24] <Syria> joschi How can I check its value please?
[08:24] <joschi> Syria: check your apache httpd configuration
[08:24] <joschi> Syria: /etc/apache2/
[08:29] <Syria> joschi is called apache.conf ?
[08:30] <joschi> Syria: basically every file in that directory. apache.conf is only one
[08:31] <Syria> joschi And which one  is respossible for the AllowOverride please?
[08:31] <joschi> Syria: check it yourself. you big enough ;)
[08:32] <Syria> joschi Okay thnx. :)
[08:34] <Syria> joschi I am viewing the files using nano on my ubuntu lucid server and most of them are blank! is this normal?
[08:55] <gobbe> what files?
[11:31] <uvirtbot`> New bug: #696753 in net-snmp (main) "Cannot disable connection logging" [Undecided,New] https://launchpad.net/bugs/696753
[12:27] <uvirtbot`> New bug: #696769 in clamav (main) "Please remove clamav source and binaries from lucid and maverick backports" [Low,Confirmed] https://launchpad.net/bugs/696769
[12:52] <uvirtbot`> New bug: #696774 in bacula (main) "package bacula-director-mysql 5.0.1-1ubuntu1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/696774
[13:57] <twister004> hi guys... is there a way in cron where we can delete old backups created by a cronjob?
[13:57] <twister004> im using ubuntu server 8.01LTS
[13:58] <Jeeves_> * * * * * rm -rf /
[13:58] <twister004>  8.04LTS
[13:58] <twister004> haha
[13:58] <Jeeves_> (that's a joke, obviously!)
[13:58] <Jeeves_> Before anyone starts using it and blaims me for their stupidity :)
[13:59] <Jeeves_> twister004: What are you talking about?
[13:59] <twister004> maybe ill have to do some kinda date comparison.. can anybody guide me?
[13:59] <Jeeves_> Which backups?
[13:59] <twister004> i am backing-up sql databases
[13:59] <_ruben> if it's a cronjob creating the backups, just extend that cronjob to delete old ones
[13:59] <twister004> i want to delete backups that are older than a week
[14:00] <twister004> _ruben... I'm new to cron.. can you guide me?
[14:00] <Jeeves_> twister004: You're misinterpreting the purpose of cron
[14:00] <Jeeves_> cron is just a scheduler
[14:00] <Jeeves_> it runs tasks you create at time you tell it to run them
[14:00] <twister004> yes... i know
[14:00] <Jeeves_> So, your question has nothing to do with cron
[14:01] <twister004> but i can use it to periodically check for old backups and delete them.. only.. i dont know how to do it
[14:01] <Jeeves_> No, you cant
[14:01] <Jeeves_> You can use it to run a script (you still need to create) to "periodically check for old backups and delete them.."
[14:01] <twister004> hmm.. any idea how i can proceed?
[14:01] <twister004> yes
[14:01] <twister004> sorry.. that's more like it
[14:01] <Jeeves_> So, shall we forget about cron for a moment?
[14:02] <Jeeves_> Shall we?
[14:03] <twister004> yes
[14:03] <twister004> i think i got it....what i can do is "find" files older than seven days and delete it ( /usr/bin/fin <location> -mtime +7 -delete)
[14:03] <Jeeves_> ok, so you've got this script that creates your backups.
[14:03] <twister004> Jeeves.. can you confirm if this'll work"?
[14:04] <Jeeves_> Yes, you can do that
[14:04] <twister004> oh gr8!!!
[14:04] <_ruben> assuming your /usr/bin/fin is something similar to the find command
[14:04] <twister004> :)
[14:04] <twister004> yes it is
[14:05] <Jeeves_> And you try it with -ls first, before you erase whole / :)
[14:06] <twister004> yes
[14:13] <gobbe> i would do script which finds and deletes them
[14:13] <gobbe> and run that from crontab
[14:54] <sander^work> How can I turn on the boot menu, so I can choose another kernel than default one?
[14:57] <gobbe> press shift during boot
[14:57] <uvirtbot`> New bug: #694538 in bacula (main) "package bacula-common 5.0.1-1ubuntu1 failed to install/upgrade: el paquete bacula-common ya está instalado y configurado" [Undecided,New] https://launchpad.net/bugs/694538
[14:58] <RoyK> sander^work: or google for it and change /etc/default/grub
[15:01] <gobbe> yes, depends that if you want to boot only once, it's easier to use shift on boot
[15:01] <gobbe> if you want to change it for every boot editing ;)
[15:07] <uvirtbot`> New bug: #695857 in apache2 (main) "ssl protection not default for sensitive packages" [Undecided,New] https://launchpad.net/bugs/695857
[15:08] <sander^work> RoyK, I've googled it.. but cant find the option to change in /etc/default/grub.
[15:10] <RoyK> sander^work: this is mine http://pastebin.com/WAFiTCf5
[15:10] <RoyK> remember to update-grupb after changing that file
[15:11] <pickscrape_> HI, would this be a good place for a couple of questions about using raid/mdadm, or should I go elsewhere?
[15:11] <gobbe> what kind of question?
[15:11] <_ruben> it could be
[15:12] <RoyK> pickscrape_: don't ask to ask, just ask
[15:12] <RoyK> there's no such thing as stupid answers, only stupid questions
[15:12] <pickscrape_> Basicall my server has ubuntu server on it which is running on standard partitioning. I've installed a second on new drives in a raid1 (for boot) and raid5 (for everything else). It works perfectly fine.
[15:12] <_ruben> RoyK: shouldn't that be the other way around?
[15:13] <RoyK> :)
[15:13] <sander^work> RoyK, thanks! :-)
[15:13] <pickscrape_> Oh, and the "everything else" then has lvm on top
[15:13] <RoyK> pickscrape_: so, something like sda1 and sdb1 in a mirror?
[15:13] <pickscrape_> My question is, in order to make moving over to the new install easier, I'd like to be able to mount the raid stuff from my old install, so I can continue to use the old install while getting things ready on the new one. But I don't know if it would like that or not.
[15:13] <RoyK> different drives for raid5 or same ones?
[15:14] <RoyK> pickscrape_: use nfs
[15:14] <pickscrape_> RoyK: this is the same server. They won't both be running at the same time.
[15:14] <RoyK> pickscrape_: or do you mean boot into the old installation? sambe box?
[15:14] <RoyK> ok
[15:14] <RoyK> then it should be trivial, as long as the old installation support the filesystem used
[15:14] <pickscrape_> Boot into old install and mount the new local volumes from the new raid array without screwing up the new install :)
[15:15] <_ruben> if all drives can be hooked at the same time, it shouldn't be a prob
[15:15] <pickscrape_> Yes, they're all running now.
[15:15] <pickscrape_> And I'm pretty sure that 9.10 can handle ext4?
[15:15] <RoyK> I can't see how you should be ruining your new install by just mounting the filesystem(s)
[15:15] <RoyK> pickscrape_: cat /proc/filesystems or modinfo ext4
[15:16]  * RoyK doesn't remember if 9.10 does support it, but would be somewhat surprised if it doesn't
[15:16] <pickscrape_> Yeah, it's there.
[15:16] <RoyK> well, just try
[15:16] <pickscrape_> My worry was that if I tried to fiddle with the raid from the old install, it would play with settings stored on those drives and break it for the new install. I'm not sure how touchy such things are. :)
[15:17] <RoyK> iirc all of that is autosense these days
[15:17] <RoyK> anyway, it shouldn't be able to import something it looks on as broken
[15:17] <RoyK> s/import/mount/
[15:18]  * RoyK was thinking in ZFS lines
[15:18] <pickscrape_> There's no /dev/md* devices at the moment so I need to figure out how to get it to detect them
[15:19] <pmatulis> pickscrape_: you'll need to install mdadm and lvm2
[15:19] <pickscrape_> pmatulis: I already have those installed
[15:19] <pmatulis> pickscrape_: then you should be able to assemble the arrays
[15:20] <RoyK> does mdadm tell you anything worthful?
[15:20] <RoyK> perhaps the md mod isn't loaded
[15:21] <pickscrape_> lsmod shows no md module, but mdadm is started
[15:22] <RoyK> no raidxx modules either?
[15:24] <pickscrape_> RoyK: yes, plenty of those (covering what I need too)
[15:24] <pickscrape_> I think --assemble is what I need to look at
[15:24] <RoyK> pickscrape_: and the old install sees the partitions, right?
[15:24] <RoyK> //proc/partitions
[15:25] <pmatulis> pickscrape_: scan and then assemble
[15:25] <pmatulis> pickscrape_: did you use mdadm.conf?
[15:25] <pickscrape_> RoyK: yes, though I have to use parted to do it. fdisk and cfdisk complain about GPT
[15:26] <pickscrape_> pmatulis: the array was created for me using the ubuntu server 10.10 installer
[15:26] <pickscrape_> So I'm not sure what it used to store the configuration.
[15:27] <RoyK> hm... GPT might be the issue
[15:27] <RoyK> dunno if 9.10 can figure out that automatically
[15:28] <pickscrape_> RoyK: if I can access it manually then I'll be happy with this. Hopefully this old install won't be sticking around for too much longer. :)
[15:29] <pickscrape_> If not, then I'll just have to do it the other way round when I'm not needing the server to be working: boot the new one, mount the old and copy over that way.
[15:30] <pmatulis> pickscrape_: can you pastebin output to 'sudo fdisk -l'?
[15:31] <_ruben> pickscrape_: doing it the other way around will probably turn out to be much easier in the end, due to the (probably) far less complex disk layout of the old install (easier to mount it in the new install)
[15:31] <pickscrape_> pmatulis: http://paste.ubuntu.com/549859/
[15:32] <pmatulis> pickscrape_: hm
[15:32] <pickscrape_> _ruben: yes, I agree. I wouldn't need any help at all with that. The downside is that it's less convenient, since I'd like to be able to use the server as well. Plus, I like to learn. :)
[15:32] <_ruben> ah ok
[16:16] <stack_>  I upgraded an 8.04 server to 10.04.  It has a software raid.  Every time it boots, it dumps to initramfs shell and I have to run "mdadm" to recognize the array and then exit to continue the boot.  How can I fix this?  One weird thing, my mdadm.conf file looks like the following: http://paste.ubuntu.com/549864/
[16:20] <RoyK> stack_: no idea - perhaps removing (or renaming) mdadm.conf will force linux to autodetect things?
[16:21] <stack_> I can wipe those duplicate lines in the config, right?
[16:21] <RoyK> IIRC removing the file will force linux to read the config from the drives
[16:21] <RoyK> which it should be stored
[16:21] <pmatulis> stack_: what kind of errors do you get?  set up a verbose boot if you don't see messages
[16:22] <pmatulis> stack_: it may be a UUID issue
[16:22] <RoyK> UUIDs shouldn't change, though, should they?
[16:22] <stack_> pmatulis: it just is waiting for root and then dumps to initramfs
[16:24] <RoyK> stack_: are the raid drives independant or do you run the root on the same drives?
[16:24] <stack_> the raid is the root
[16:24] <RoyK> ok
[16:25] <RoyK> sorry, never tried (yet) to upgrade from softmirrored root
[16:25] <pmatulis> stack_: like i said, boot verbosely (--verbose or --debug kernel boot option)
[16:26] <stack_> pmatulis: I'll have to try it and get back to you.  It's a production system.  We had a power outage, so I tried to figure it out then but didn't get far
[16:26]  * SpamapS streeettcchhess
[16:27] <stack_> thanks for your help, all... I'll dig further
[16:32]  * SpamapS is trying "terminator" for the first time.. seems to be just the thing for a big monitor
[16:33] <highvoltage> terminator ftw
[16:33] <pmatulis> SpamapS: i don't think it does anything that normal screen can't but the mouse support is convenient
[16:33] <RoyK> SpamapS: terminator?
[16:34] <SpamapS> pmatulis: even if screen does everything it does.. for some reason this made a lot more sense to me than screen's help pages ever did
[16:34] <SpamapS> Been using screen for years, never liked it much.
[16:34] <highvoltage> pmatulis: you can also split a window (or tab) into many horizontal and vertical splits, it's quite useful
[16:35] <pmatulis> highvoltage: but screen can do this too
[16:35] <highvoltage> and you can make one of the splits temporarily grow to the full window size. combination of screen/byoby/terminator is really great
[16:35] <SpamapS> and actually, I'm using byobu inside the terminator screens, which is probably silly but works fantastically
[16:35] <SpamapS> I just have two vertical splits now, but when I want to do something, I split horizontally, do it, then close the window
[16:35] <kirkland> SpamapS: what version of byobu are you running?
[16:36] <SpamapS> kirkland: in my hosted CentOS box.. "old" .. and then whatever maverick has.
[16:36] <kirkland> SpamapS: ah ... i was gonna show you something cool :-)
[16:36] <SpamapS> I only use it for ctrl-a N and ctrl-a C
[16:36] <kirkland> SpamapS: if you have a natty/byobu somewhere ...
[16:36] <SpamapS> kirkland: upgrade to natty is my 9:00am TODO ;)
[16:36] <sciguy16> Is it possible to give a normal user permission to reload apache?
[16:36] <SpamapS> which is in 24 minutes
[16:37] <RoyK> sciguy16: no
[16:37] <kirkland> SpamapS: heh, okay, poke me once you do
[16:37] <SpamapS> kirkland: poke queued
[16:37] <sciguy16> RoyK: ok, thanks
[16:37] <RoyK> sciguy16: you can allow it with sudo, though
[16:37] <RoyK> sciguy16: man sudo is a good start
[16:37] <SpamapS> ok this is weird..
[16:37] <SpamapS> ssh agent isn't working in terminator
[16:38] <highvoltage> pmatulis: indeed. I see them as two different completely different things though, they just happen to offer some of the same functionality
[16:38] <sciguy16> RoyK: thanks, I'll look into it
[16:39] <pmatulis> highvoltage: yes, i use terminator + screen
[16:41] <SpamapS> actually strike that.. ssh agent just isn't working. :(
[16:44]  * RoyK wonders when gnome apps came into ubuntu-server
[16:47] <compdoc> I install the gnome desktop on server. is that bad?
[16:48] <RoyK> it's ok, but if you want a desktop machine, I'd recommend installing ubuntu-desktop
[16:48] <RoyK> the server stuff is there, or available, after all
[16:48] <RoyK> and installing the desktop machine will allow for better customization last I checked
[16:49] <SpamapS> RoyK: terminals are the window to the server. :)
[16:49] <compdoc> well, I need a server - its not used as a desktop. I just like the gui
[16:49] <RoyK> post-installing gnome led to a few headaches last I tried
[16:49] <RoyK> compdoc: it's the same thing, desktop and server, same distro
[16:49] <RoyK> it's just that if you want a fancy guy, starting off with -desktop will do things easier for you
[16:50] <RoyK> the only difference between ubuntu server and desktop are (1) default packages installed and (2) kernel chosen. for most use the desktop kernel will work just as fine as the server kernel
[16:50] <compdoc> I thought I read that Server has things set up differently
[16:51] <RoyK> just the server bits, which are minimal
[16:51] <compdoc> I think to work better under load
[16:51] <RoyK> that's HZ in kernel, yes
[16:51] <RoyK> also, desktop comes with a gui-based network setup, so it might not work out of the box with a static ip (it's set after login)
[16:52] <RoyK> but still, why do you want a desktop on your server?
[16:52] <compdoc> I have a large rsync transfer going for several hours now, and gnome wont open any programs like a term window
[16:52] <compdoc> so I have to wait
[16:52] <RoyK> just kill rsync, restart it in a 'screen'
[16:53] <RoyK> then you can do whatever you want without disrupting it
[16:56] <compdoc> a screen?
[16:56] <compdoc> a terminal window?
[17:19] <Datz> ooh, 67 days uptime, 68 updates :)
[17:19] <Datz> more than one update a day!
[17:20] <Datz> average anyway  .. good work :p
[17:21] <Datz>  11:22:08 up 67 days, 21:13,  2 users,  load average: 0.83, 0.94, 0.89
[17:22] <pmatulis> compdoc: 'man screen'
[17:24] <compdoc> thanks
[17:31] <thesheff17> when I use vmbuilder I'm getting no bootable device.  Did something change recently?  I see the file for the virtual machine.
[17:41] <thesheff17> nm I found a bug for it 659532.
[18:00] <yeags> hey all
[18:00] <RoyK>  
[18:01] <yeags> i'm looking at building an Ubuntu private cloud
[18:01] <yeags> Is it possible to have the nodes power up using WakeOn Lan?
[18:06] <baggar11> yeags: that's more of a NIC requirement, than ubuntu-server based
[18:10] <gobbe> wake-on-lan works quite bad in most of the cases
[18:10] <gobbe> i would prefer iLO or something like that
[18:14] <yeags> yeah iLO is much nicer, but i'm planning on a very cheap, basic home system
[18:15] <yeags> I just wanted a single big server as the controller which could run 2-3 servers, and then have 3-4 laptops I have here that I can power up and run services on ondemand
[18:15] <yeags> I was wondering if the software had any support for sending WOL packets to automatically bring up servers into the cluster to run more servers
[18:15] <yeags> Currently I just use Xen and manually fire up my servers
[18:15] <yeags> and manually move images from machine to machine
[18:16] <yeags> UEC looked like a nice option with a much nicer front end
[18:18] <yeags> So my plan is a 3TB main server with 8GB ram, quad core CPU
[18:19] <yeags> run 3-4 servers on that, and then have my spare laptops with 2GB RAM run development servers when I need them
[18:19] <yeags> I do a lot of technology validation and thought UEC would be a great platform for managing test systems
[18:21] <yeags> I could easilly use a linux WOL client to fire off the packets
[18:21] <yeags> is my plan a sound one?
[18:21] <gobbe> yes
[18:21] <gobbe> it sounds ok :)
[18:21] <yeags> :) smashing
[18:21] <yeags> I assume the main controller can also run servers without issue?
[18:22] <RoyK> yeags: what sort of storage solution?
[18:22] <yeags> well... I have options. I have a 6TB NAS (QNAP) and also the main server will have a 3TB array in it
[18:22] <RoyK> for large storage, I'd recommend zfs on some good platform
[18:22] <yeags> define large storage? :)
[18:23] <RoyK> 5+TB
[18:23] <RoyK> perhaps
[18:23] <yeags> well my data will reside on the QNAP and I will access via iSCSI
[18:23] <RoyK> thing is, large storage systems need to use checksumming
[18:23] <RoyK> since silent errors occur quite often with multiterabyte systems
[18:23] <yeags> I can't find any good architecture diagrams online
[18:24] <RoyK> and zfs has fairly good checksumming for this
[18:25] <RoyK> so with a box running multiple VMs, I'd say a separate box for storage would be decent
[18:28] <yeags> what is performance like of a gigabit ethernet network when the server image is on a remote NAS storage?
[18:28] <yeags> surely I would be better actually storing the images locally on the server?
[18:29] <SpamapS> yeags: Not necessarily
[18:30] <SpamapS> yeags: if the local storage is very busy with a lot of concurrent requests, and your NAS has tons of cache and disks to work with.. it may do a better job
[18:30] <SpamapS> RoyK: whats the status of ZFS on Linux these days, any idea?
[18:31] <yeags> http://en.wikipedia.org/wiki/ZFS#Linux
[18:31] <yeags> not looking good
[18:31] <yeags> and Oracle would sure screw it up if they could
[18:31]  * yeags works for Oracle, unfortunately
[18:31] <RoyK> SpamapS: works well on fuse
[18:32] <RoyK> SpamapS: works better with openindiana :)
[18:33] <SpamapS> RoyK: Is Nexenta's ZFS using FUSE?
[18:33] <yeags> SpamapS: local storage will be much faster and responsive than the QNAP remote NAS
[18:33] <yeags> I'm planing on non OS file systems being on the NAS
[18:33] <yeags> like regular file shares
[18:33] <gobbe> SpamapS: no it's not, nexenta is solaris-based
[18:33] <RoyK> SpamapS: no, it uses a solaris kernel
[18:34] <SpamapS> gobbe: ahh.. right
[18:34] <SpamapS> I keep hearing that BTRFS has some serious problems with its metadata efficiency
[18:35] <RoyK> SpamapS: btrfs announced on their ml that raid[56] was on they way some 20 months ago
[18:35] <RoyK> seems there's a prerelease being announced soon
[18:35] <SpamapS> thats good to hear
[18:35] <RoyK> and btrfs isn't even close to zfs security-wise
[18:35] <SpamapS> baby steps
[18:36] <RoyK> SpamapS: really? if it took them two years to implement raid[56], how long to stabilise?
[18:36] <SpamapS> RoyK: who are "they" anyway?
[18:36] <SpamapS> wasn't Oracle a big component of BTRFS development?
[18:37] <RoyK> oracle gave the source away some four years back iirc
[18:37] <SpamapS> I'm more curious about who was doing the actual coding and designing
[18:37] <RoyK> see the source
[18:38]  * SpamapS must go afk... bbiab
[18:38] <RoyK> usually the bits are pretty well credited
[18:46] <sbeattie> SpamapS: chris mason is the primary developer for btrfs, he was working on it before he joined oracle. Not sure if he's roped other oracle devs (zach brown?) into developing for it.
[18:48] <yeags> Chris is a development director in Oracle, has 6 developers reporting to him
[18:52] <RoyK> 6 developers working on btrfs
[18:52] <RoyK> seems perhaps oracle doesn't want btrfs to succeed
[18:53] <RoyK> now they have zfs
[18:53] <yeags> would
[18:53] <yeags> wouldn't surprise me
[19:12] <pting> regarding unattended upgrades pf packages, is there something i can do to automatically select "keep existing config" when doing package upgrades... is there something i can set in debconf-set-selections?
[19:17] <snipeTR> hi
[19:17] <consumerism> i'm trying to diagnose a problem, not really ubuntu related but i figured this would be a good place to ask. occasionally, when i am navigating a website served by a ubuntu server in ec2, i get a timeout.  the problem usually persists for a few minutes and then goes away. apache access logs show no requests while the timeout happens, so the requests are never hitting the server (right?).
[19:18] <consumerism> nothing changes on the client end, the browser loads pages one minute, times out for a few consecutive minutes, then is able to load pages again. load on the server is negligible. the best i can come up with is some network problem outside my control between client and server, is there any other possibility? what could i do about this?
[19:18] <snipeTR> "Long-Therm" What is the difference?
[19:20] <pting> consumerism, firewall issues? possibly some rate limiting policies in apache? try spinning up another instance in ec2 and do an httperf
[19:21] <billybigrigger> where i can i poll imap/smtp info from?
[19:21] <billybigrigger> kind of like phpinfo for php, is there anywhere i can get info on my mail system?
[19:21] <snipeTR> hi guys? "Long-Therm" What is the difference?
[19:22] <Pici> snipeTR: Are you asking about LTS - Long Term Support?
[19:23] <snipeTR> yes what is this long term?
[19:23] <Pici> !lts
[19:23] <snipeTR> ohh ok tnx pici
[19:24] <snipeTR> mean only support time different? thru?
[19:24] <Pici> snipeTR: Essentially.
[19:24] <Pici> Also, LTS releases can be upgraded from one to the next without having to go through the intermediary releases.
[19:25] <Pici> So. 8.04 to 10.04 directly. Instead of having to go to 8.10 -> 9.04 -> 9.10 -> 10.04
[19:25] <snipeTR> tnx my friend. this info source ? URL?
[19:28] <snipeTR> Pici soo tnx byes
[20:15] <pting> has anyone used ubuntu-eucalyptus extensively? i'm thinking of building out a cluster for work
[20:16] <pting> ... we use ec2 so it seems like a perfect fit. i'm trying to figure out challenges i may encounter.
[20:17] <SpamapS> pting: #ubuntu-cloud is also a good place to chat about it
[20:17] <billybigrigger> using the ubuntu server guide...how come i can't send mail to gmail users?
[20:17] <SpamapS> billybigrigger: what is your evidence that you ca n't send mail to gmail users?
[20:17] <pting> SpamapS, thanks
[20:19] <billybigrigger> the fact that i can send to hotmail works..
[20:19] <billybigrigger> my test email to my gmail account didn't...this looks a little suspect though....
[20:19] <billybigrigger> Jan  3 15:12:14 timmy postfix/smtp[7303]: certificate verification failed for gmail-smtp-in.l.google.com[74.125.93.27]:25: untrusted issuer /C=US/O=Equifax/OU=Equifax Secure Certificate Authority
[20:19] <billybigrigger> i wonder if gmail doesn't like my self created CA
[20:19] <billybigrigger> i don't know where the Equifax Secure CA is coming from though
[20:21] <SpamapS> billybigrigger: that may be theirs
[20:22] <SpamapS> billybigrigger: do you have ca-certificates installed?
[20:22] <billybigrigger> i turned verbose logging of smtpd on, i'll pastebin my attempt to email my gmail account
[20:22] <billybigrigger> Jan  3 15:20:05 timmy postfix/local[7406]: 5C2AA61CC: to=<www-data@mail.thefrozencanuck.ca>, relay=local, delay=0.02, delays=0.01/0/0/0, dsn=5.2.0, status=bounced (maildir delivery failed: create maildir file /var/www/Maildir/tmp/1294086005.P7406.timmy: Permission denied)
[20:23] <billybigrigger> i keep getting those stupid errors in my mail.log too...any one have an idea where those are coming from?
[20:23] <SpamapS> billybigrigger: thats a local delivery failure
[20:23] <SpamapS> billybigrigger: www-data is getting email
[20:23] <SpamapS> billybigrigger: You can alias email to www-data to something else.
[20:23] <billybigrigger> wonder if its my webmail client...
[20:23] <billybigrigger> effin things up...
[20:23] <billybigrigger> anyway back to gmail...
[20:25] <billybigrigger> http://pastebin.com/1ficC5FT
[20:36] <pmatulis> billybigrigger: i see:
[20:36] <pmatulis> Jan  3 15:24:14 timmy postfix/smtp[7417]: C578361CC: to=<donzavitz@gmail.com>, relay=gmail-smtp-in.l.google.com[74.125.93.27]:25, delay=0.83, delays=0.36/0.01/0.24/0.22, dsn=2.0.0, status=sent (250 2.0.0 OK 1294086254 g3si36207791qcq.181)
[20:36] <billybigrigger> it never shows in my inbox though
[20:36] <billybigrigger> i've sent about 5 test emails today
[20:37] <billybigrigger> all within an hour
[20:37] <billybigrigger> haven't recieved 1 yet...hotmail test mail was instant
[20:38] <consumerism> pting: if it was firewall, it would not be intermittent...right?
[20:43] <pting> consumerism, doesn't sound like it, but maybe there's some sort of connection policy that throttles connections, i would first try to make repetative connections from within ec2 then try with a machine outside... or it could be that that particular instance is running on hardware that's having issues
[20:43] <pting> consumerism, what's the instance size?
[20:44] <pmatulis> pting: consider #ubuntu-cloud
[20:46] <pting> pmatulis, consumerism is having the ec2 connectivity instance issues, not me =)
[20:46] <pmatulis> pting: 'scuse me
[20:46] <pmatulis> consumerism: see above
[21:10] <erichammond> Amazon EC2 has rules about how many emails you can send per time period for new accounts.  You can get them lifted by applying for whitelisting.
[21:11] <SpamapS> erichammond: I seem to recall smoser asking some amazon people and they basically said that SMTP from EC2 was unreliable and unpoliced by them.
[21:11] <SpamapS> other than responding to spam reports by cancelling accounts I mean
[21:11] <SpamapS> but maybe I misunderstood
[21:12] <erichammond> SpamapS: There are real time block lists that ban email from EC2 IP ranges from time to time which I think is the primary "unreliable" factor.
[21:12] <erichammond> See also: https://aws-portal.amazon.com/gp/aws/html-forms-controller/contactus/ec2-email-limit-rdns-request
[21:14] <erichammond> I recommend sending email from EC2 servers through a non-EC2 email relay with a clean IP address.
[21:16] <erichammond> You can either use a third party (gets expensive for high volume) or lease a simple server on any non-"cloud" provider (i.e., one that doesn't let new folks sign up with a credit card and rent computers by the hour).
[21:55] <DesignsEdge> Afternoon all!  I hope that I can get some help here ---
[21:57] <IrishWristwatch> Maybe you can, but we don't know what the problem is.
[21:58] <DesignsEdge> well an end user just called with a problem - will brb -- sorry