[02:08] <frank_> hi
[02:10] <frank_> I want to install ubuntu server 12.04 LTS on a VM. The homedir should be encrypted for every user, but the server should be able to boot without any interveniance (since it's a VM, vmware, ha, etc). What do I select in the installer?
[02:11] <frank_> Encrypt the first user and add the second one from cli via adduser? How do I get the homedir of the second user to be encrypted?
[02:11] <sarnold> frank_: probably "encrypted user home directories" -- that sets up ecryptfs to mount the user home directories with a stacked encrypted filesystem when users log in. it does complicate user crontabs, but it does boot without intervention
[02:11] <frank_> I guess the installer doesn't provide support to create n user accounts during installation where the homedir is encrypted
[02:12] <ScottK> No
[02:12] <tyhicks> frank_: I'm not sure if the installer provides that
[02:12] <tyhicks> frank_: but see the --encrypt-home option of adduser
[02:13] <frank_> tyhicks: nice, thanks
[02:13] <frank_> sarnold: what do i select during installation? Use entire vmdk and set up lvm?
[02:14] <sarnold> frank_: I believe that would allow you to grow the image / filesystem later if you wish.
[02:15] <frank_> sarnold: done
[02:18] <frank_> my provider promised me only level 3 filtering, configuring apt tages ages but eventually will finish. On TTy3 I see failed to fetch http://de.archive.ubuntu.com/...
[02:18] <frank_> i guess he's lying, isn't he?
[02:19] <frank_> i guess during install I haven't got a links to verify
[02:20] <sarnold> try also w3m, curl, wget, links, elinks, might be something..
[02:20] <sarnold> or if all else fails, maybe bash's /dev/tcp
[02:22] <frank_> last time I was behind a watchguard, which was pretty funny when I did am echo of seq(1,100) in a .txt and then wget'ted it. First got lines 51-100. Then 1 - 50
[03:30] <centaur5> I have a NFS share that mounts and even though I specified rw in fstab whenever I try to write to the mount point it says read only. Any ideas?
[04:15] <hallyn> stgraber: all right no worries, i see what's going wrong with nested lxc in ppa.  i *could* wokr around it in the mountcgroups script but am going to spend some time doing what i think is the right thing when creating containers
[04:16] <hallyn> (which is keep curcgroup per cgroup mount, and check tasks file to figure out what cgroup we are in relative to the mounted one, not to the host's / cgroup)
[04:57] <avis_> quite the large room, hello everyone
[04:58] <avis_> I'm having a slight dependency problem when installing libapache2-mod-python: I recieve the following error: http://pxl.uni.cx/irc
[04:59] <avis_> basically says I have have unmet dependencies
[04:59] <avis_> Depends: python (< 2.7) but 2.7.3-0ubuntu2.2 is to be installed
[05:00] <avis_> however this is my phython version output - Python 2.7.3 (default, Aug  9 2013, 22:08:36)
[05:00] <avis_> sudo aptitude provides too many regressive dependencies to warrant going that route
[05:00] <avis_> Any ideas?
[11:38] <samba35> i have strange problem ,on my home system i am not able to boot with harddisk of vmware esxi 5.1 when i try to boot with harddisk  booting process goes in loop but if i boot with dvd and select boot from hardisk then system boot properly
[11:41] <samba35> sorry
[12:03] <samba35> is it possible to add other then ubunut iso image to create startup disk /using startup  disk creator
[15:13] <progre55> Hi guys. Is logrotate's delaycompress option needed for apache2, if I have set postrotate to reload apache? It's just, I rotate the logs weekly, and the access-forward logs can be really huge uncompressed.
[15:51] <andygraybeal> how do i transfer a KVM with an that lives in an LVM container to another machine?
[15:51] <andygraybeal> LVM container/partition ... i don't know the words.
[16:34] <yeats> andygraybeal: the fact that it's on LVM shouldn't matter - you should be able to move it
[16:35] <yeats> at the userspace level, the details of the storage medium shouldn't matter
[16:42] <andygraybeal> yeats, okay
[16:42] <andygraybeal> i have no experience with this, so i'm not sure ;)
[16:44] <andygraybeal> so... the over view is.. to copy the data out of the LVM partition into another LVM partion and then load up the KVM definition into libvirt?  is that basically what i need to accomplish?
[16:57] <yeats> andygraybeal: I believe so, yes
[16:57] <andygraybeal> okay thank you yeahpla
[16:57] <andygraybeal> errrr yeats
[16:58] <andygraybeal> i've never done snapshotting of LVM yet
[16:58] <yeats> andygraybeal: the main point is, at the OS level, you can think of this as "directories" rather than worrying about which type of storage medium it is
[16:58] <yeats> andygraybeal: very simple
[16:58] <yeats> 'virsh snapshot-create <vmname>'
[16:59] <andygraybeal> hmm.. with virsh?  it deals with LVM ?
[16:59] <andygraybeal> i was thinking that i would need to do LVM snapshot of the partition
[17:00] <andygraybeal> or.. i'm listening, i'm full of confusion :)
[17:00] <yeats> oh LVM?
[17:00] <andygraybeal> yes, i got a libvirt box on lvm
[17:00] <andygraybeal> not a image file
[17:01] <yeats> is your end goal that you want a snapshot of your VM?
[17:01] <andygraybeal> well.. the goal is to move a virtual machine from one physical machine to another physical machine.
[17:01] <yeats> andygraybeal: oh - I see - "snapshot" has a specific meaning in KVM, so I thought you meant that ;-)
[17:02] <andygraybeal> yes, sorry, i meant LVM snapshot, i forgot it meant something in the KVM world for two seconds
[17:05] <progre55> Hi guys. Is logrotate's delaycompress option needed for apache2, if I have set postrotate to reload apache? It's just, I rotate the logs weekly, and the access-forward logs can be really huge uncompressed.
[18:28] <axisys> why does one precise server showing linux 3.2 and the other precise server showing linux 3.5
[18:28] <axisys> ?
[18:39] <sl8_slick> I'm having an issue with my ovh server, I just added an IP block and everything went fine, but it broke one of my ubuntu VMs. It refuses to connect to the internet or ping the gateway, and when I type ifconfig all the information is correct
[18:49] <RoyK> axisys: probably because those that say 3.5 were installed with 12.04.2, which uses 3.5 by default. earlier 12.04 releases uses/used 3.2. you can install 3.5 on the older ones if you need to, but 3.2 will work as well. default kernel was changed to 3.5 to allow for better support for newer hardware
[19:09] <axisys> RoyK: oh ok..
[19:09] <axisys> I will just upgrade the kernel since these two servers will be failover for each other.
[19:09] <axisys> RoyK: thanks
[19:10] <RoyK> axisys: np
[19:21] <ddsss> fresh ubuntu server 13.04 install. installed from usb key. doesn't boot.
[19:21] <ddsss> simply flies thorugh intial output and then monitor says no signal.
[19:22] <ddsss> I can't even see what erro is there - so fast it goes.
[19:48] <axisys> ddsss: you get anything with shift-alt-F1 or F2 or F3 .. F7 ?
[19:49] <ddsss> axisys, yes - bios works just fine.
[19:49] <ddsss> axisys, then - ubuntu installs just fine.
[19:49] <axisys> huh?
[19:50] <axisys> shift-alt-F1 gives you anything?
[19:50] <ddsss> axisys, ahh - no.
[19:50] <axisys> its one of the vtys
[19:50] <axisys> how about F2 .. to .. F7
[19:50] <ddsss> axisys, Its a headless server install. so i didn't install gui.
[19:51] <axisys> 15:21:46 < ddsss> simply flies thorugh intial output and then monitor says no  signal.
[19:51] <ddsss> axisys, exactly!
[19:51] <axisys> if it is headless why expecting something in the monitor?
[19:51] <ddsss> axisys, I mean - server without gui.
[19:52] <axisys> can you ssh to it?
[19:52] <ddsss> axisys, but yes - there is  a monitor attached via vga cable.
[19:52] <axisys> tty0 is vga and ttyS0 is console
[19:52] <axisys> where are you redirecting the display in grub?
[19:53] <ddsss> axisys, mmm. it was an automatica installl. how would I check?
[19:53] <ddsss> e in grub menu?
[19:53] <axisys> automatic install usually points to tty0.. but yes check in the grub
[19:54] <ddsss> i'll start it again. hang on:)
[19:54] <axisys> it should be something like console=tty0 and/or console=ttyS0,9600n8 or like
[19:54] <axisys> i would remove the quiet after --
[19:54] <axisys> and also remove the splash..
[19:55] <axisys> for server you really dont want to use those crap
[19:57] <ddsss> hmm. pressing e on ubuntu entry gives me some short script or a config or what not.
[19:57] <ddsss> setparams 'Ubuntu '
[19:57] <ddsss> recordfailt
[19:57] <ddsss> etc
[19:58] <ddsss> imhonestly not sure which one is video
[19:58] <ddsss> it does ave load_video though
[19:58] <ddsss> it does have load_video though
[20:02] <ddsss> ahhh! GPU power managament timed out - what does that mean?
[20:03] <ddsss> axisys, I've got the actual error  - I've had to record the whole thing on camera and then select the last part:)
[20:04] <ddsss> gma500 GPU: power managament timed out.
[20:09] <ddsss> axisys, ahh. it seems the new mobo I got for a nas has some completely unsupported graphics which require adding additional repos during install: gma500 on Launchpad.
[20:17] <ddsss> axisys, ha. I was just been able to ssh in. SO I guess it worked ll along  - just video driver wasn't present.
[20:56] <bananapie> Are hardware interrupts related to kernel frequency?
[21:13] <failmaster> i have a problem with switching passphrase luks auth on boot to key file based, extra details: http://askubuntu.com/questions/330660/what-is-the-correct-way-to-use-key-file-luks-authorization anyone?
[21:27] <RoyK> bananapie: no
[21:29] <bananapie> how can I tell if I have a bottleneck from my IRQ interrupts?
[21:31] <RoyK> can you pastebin /proc/interrupts?
[21:31] <RoyK> or perhaps how much hi is reported from top?
[21:31] <bananapie> http://pastebin.com/kWP0Vt9q
[21:31] <RoyK> also - what sort of server is this?
[21:32] <bananapie> HP Proliant I think.
[21:32] <bananapie> DL360 or something like that I bought back in 2010
[21:32] <RoyK> a router?
[21:33] <bananapie> Nope. It has 4 networks cards bonded into bond0 and bond1. It has a Sangoma A108DE card that receives PRIs from my telco and I convert the calls into SIP/RTP on the server.
[21:34]  * RoyK smells asterisk
[21:34] <bananapie> Nope :P I stopped using asterisk, I am running freeswitch.
[21:34] <bananapie> But close ;)
[21:35] <RoyK> goodie
[21:35] <bananapie> I can't tell if that is sarcastic
[21:35] <RoyK> I guess the NICs don't support much buffering or may be need to get their buffers increased
[21:35] <RoyK> no, it was not ;)
[21:35] <bananapie> What makes you believe that I have buffer problems?
[21:36] <RoyK> with higher buffers, you get lower interrupt rates
[21:36] <bananapie> ok
[21:36] <RoyK> check the module settings for the NIC driver(s)
[21:37] <RoyK> no idea if those MSI NICs support much buffering, though
[21:37] <RoyK> Intel cards are usually my best bet for such use
[21:37] <bananapie> I'll check. Broadcom Corporation NetXtreme II BCM5709 Gigabit Ethernet
[21:37] <RoyK> ouch
[21:37] <RoyK> those aren't too good
[21:37] <RoyK> but may be tunable for something better
[21:38] <bananapie> I like that the first page I found on google starts with "These cards are flawed".
[21:38] <RoyK> I mean - the NIC may be good, but Broadcom doesn't release much info about their hardware so much of the driver development is about reverse engineering
[21:38] <RoyK> hehe
[21:39] <bananapie> could the badness of these ethernet card impact the performance of my sangoma card?
[21:40] <RoyK> hardly, unless it's PCI
[21:40] <RoyK> you should have enough CPU cores to handle it
[21:40] <bananapie> the 4 ethernet cards are built into the motherboard and the sangoma is PCI express
[21:41] <bananapie> ok.
[21:41] <RoyK> then it should work well
[21:41] <RoyK> on which interrupt is the sangoma?
[21:41] <bananapie> 30
[21:42] <RoyK> rather a few interrupts from that one as well :P
[21:43] <bananapie> 916 million interrupts in 5 days, is that a lot?
[21:43] <RoyK> bananapie: this says a bit about what I think about asterisk :P http://karlsbakk.net/fun/asterisk-installation.wav
[21:44] <bananapie> That was brilliant!
[21:44] <RoyK> or http://karlsbakk.net/fun/asterisk_architecture.jpg
[21:47] <bananapie> Would it be a bad idea to allow 4 CPUs to the sangoma card?
[21:49] <bananapie> I really need to learn more about Interrupts. Do you have any documents that you especially like that gives good info about interrupts in Linux?
[21:50] <bananapie> what's odd about the card is the smp affinity was 00200020 instead of ffffffffff like I found in /proc/irq/default_snmp_affinity
[21:51] <Patrickdk> how about we stop all this *guessing*
[21:51] <Patrickdk> what is the *problem*?
[21:51] <bananapie> I randomly get talk-off on my PRI and some faxes fail. I thought it was a sangoma/freeswitch problem, but yesterday
[21:52] <RoyK> Patrickdk++
[21:52] <RoyK> bananapie: talk-off?
[21:52] <bananapie> I installed a Zabbix machine on the network to monitor my servers, all the servers generate  beautiful graphs.
[21:52]  * RoyK doesn't quite understand
[21:52] <bananapie> But for some reason, Zabbix can't get more than about 5% of the data it requests from the machine with the Sangoma in it.
[21:52] <RoyK> bananapie: t.38 is always hard to work with, though
[21:52] <bananapie> But I ping the machine and it responds fine.
[21:53] <bananapie> So, I don't know what is going on, so I suspect there may be an evil spirit living in my PCIE bus.
[21:53] <Patrickdk> seriously unlikely
[21:53] <Patrickdk> a bad driver, sure
[21:53] <RoyK> bananapie: have you tried to talk to coppice about the spandsp issues?
[21:53] <bananapie> The Evil Spirit of Sys-admin-know-nothing-about-his-IRQs
[21:54] <bananapie> Yes. I spoke with coppice, we talked about echo cancellation and stuff.
[21:54] <Patrickdk> bananapie, how many pri's?
[21:54] <bananapie> 8
[21:54] <bananapie> but, I disconnected 2 to see what would happen
[21:54] <bananapie> so the only 6 have been connected for the past few weeks
[21:54] <bananapie> Because of the funky behaviour from Zabbix that is isolated to this machine, I think something else is going on on the server.
[21:55] <Patrickdk> well, that should max you out around 17mbit
[21:55] <bananapie> But only 2 of the PRIs see any serious use, it is rare that the other PRIs use more than 3-4 channels a tthe same time.
[21:56] <Patrickdk> how did you monitor this, only 5% data?
[21:56] <bananapie> the graphs on Zabbix have a few dots instead of full lines like all the other servers that have the exact same hardware and Ubuntu version ( excluding the sangoma card )
[21:58] <Patrickdk> it is a pcie sangoma card?
[21:58] <bananapie> Yes.
[21:58] <Patrickdk> not the pci version?
[21:59] <bananapie> PCIE
[21:59] <bananapie> A108DE
[21:59] <bananapie> d = hardware DTMF and E = PCI Express.
[22:08] <bananapie> less -S makes /proc/interrupts readable on my screen :D
[22:50] <bananapie> hello?
[22:51] <failmaster> i have a problem with switching passphrase luks auth on boot to key file based, extra details: http://askubuntu.com/questions/330660/what-is-the-correct-way-to-use-key-file-luks-authorization anyone?