[00:07] <jkitchen> question: if I set up a 2 partition root, give one partition to /boot, and the other as a pv to an lvm vg... if I want to resize that partition/pv/vg later, can I do that online? (the root block device is a LUN from my san, so I can resize the block device itself)
[00:08] <TJ-> jkitchen: Yes, you can use pvresize to expand/shrink the PV underlying a VG
[00:09] <jkitchen> right, but what about the partition table
[00:10] <jkitchen> this may actually be moot because it looks like ubuntu (at least 12.04) has no problems booting off of an lvm root with /boot inside the lv
[00:10] <jkitchen> but it does look like partman is still creating a partition tabel
[00:10] <jkitchen> rather than just giving the whole block device over to lvm
[00:12] <TJ-> jkitchen: pvresize is used in conjunction with the partitioning tools
[00:12] <TJ-> jkitchen: With LVM though you could, instead, just add another block device as a PV to the VG
[00:12] <TJ-> jkitchen: no need to resize
[00:13] <jkitchen> that still requires manipulating the partition table to add a new partition though, right?
[00:13] <jkitchen> or are you saying expose another LUN to the vm just to add space? (ew)
[00:14] <TJ-> jkitchen: not if you add a new block device... e.g. sdd2 >PV1 > VG  .... then later +sde > PV2 >vgextend VG /dev/sde
[00:15] <jkitchen> yea, I don't want to have to create another lun just to add space. the SAN can grow a lun and we already do that with data volumes and such
[00:15] <TJ-> jkitchen: alternatively, if you want to stay in one partition of one block device, make the PV the last paritition, so when you extend the block device, you just redo the sector count of the final partition, then do pvresize on that partition
[00:15] <jkitchen> right now our root volumes are just direct partition tables on the disk, I'd like to just hand the block device to lvm and have it go nuts
[00:16] <jkitchen> yea
[00:16] <jkitchen> since it seems I can actually boot directly from the lvm with no problem, I may try to see if I can get partman to just use lvm raw on the block device without creating a partition table
[00:17] <jkitchen> but something tells me that's not going to be easy :)
[00:17] <jkitchen> I'll poke at it a bit
[00:17] <jkitchen> right now it seems like it's creating the pv as the only partition, so resizing that one partition and then pvresizing should be fine, but I'd rather remove that one step if I can
[00:18] <TJ-> you could do that if you create the base install using deboostrap followed by some config and tasksel
[00:18] <jkitchen> I'm preseed installing
[00:19] <jkitchen> https://gist.github.com/kitchen/ba72af4a48b94fb0f463 that's the relevant section of my preseed file
[00:19] <jkitchen> I still need to answer a couple of questions in it, I'm adapting my existing template to lvm, but that does produce a bootable system
[00:20] <jkitchen> (it just did, actually)
[00:24] <jkitchen> and now I'm reinstalling with the questions answered so hopefully it goes full auto :D
[00:44] <DeltaHeavy> Can anybody help me figure out why my PHP file tries to download to the client, rather than execute? https://gist.github.com/anonymous/af31d161f1b96b83fcfb
[00:44] <DeltaHeavy> I'm using 12.04
[00:46] <sarnold> teward: ^^^
[00:47] <hitsujiTMO> DeltaHeavy: did you install libapache2-mod-php5?
[00:48] <DeltaHeavy> hitsujiTMO: Why would I need libapache2-mod-php5 if I'm using nginx?
[00:48] <hitsujiTMO> ahh, nginx not apache :P
[00:48] <DeltaHeavy> Yar
[00:49] <hitsujiTMO> DeltaHeavy: have you installed php-fpm?
[00:49] <DeltaHeavy> Yep
[00:49] <DeltaHeavy> However I get access denied when trying to read this file with sudo: /var/run/php5-fpm.sock
[00:51] <sarnold> I'm surprised about the access denied; I figured you'd just get nothing until something wrote into the pipe..
[00:51] <DeltaHeavy> Got it fixed. Needed to change the default MIME type
[00:51] <hitsujiTMO> DeltaHeavy: try_files $uri =404 is serving the file. remove that line
[00:51] <sarnold> are you sure it's a pipe?
[00:53] <sarnold> teward: ah, nevermind, he fixed his problem :) thanks
[01:24] <jkitchen> seems online partition resizing is not possible :(
[01:25] <DeltaHeavy>  jkitchen Why not?
[01:25] <DeltaHeavy> Should be
[01:25] <jkitchen> DeltaHeavy: parted keeps wanting to try to resize the filesystem for me, even though it warns me not to do that (I don't want it to)
[01:26] <jkitchen> all of the docs I've found searching for "ubuntu online partition resize" say it can't be done online
[01:26] <DeltaHeavy> jkitchen: Extend it or shrink it?
[01:26] <jkitchen> extend
[01:26] <jkitchen> I increased the lun size. parted sees this free space
[01:26] <DeltaHeavy> Can't you do it via SSH X11 tunneling, or fdisk via SSH?
[01:26] <DeltaHeavy> I swear I've done it before
[01:27] <jkitchen> I'd like to eventually automate this, so X11 ssh tunneling ... hah.
[01:27] <jkitchen> didn't think about using just fdisk
[01:27] <jkitchen> sec
[01:28] <sarnold> I know I've used LVM to resize a filesystem before, back in 2002 or 2003. I was stunned it worked. :)
[01:28] <jkitchen> oh for sure
[01:28] <jkitchen> I do resize2fs on my bare block devices all the time
[01:28] <jkitchen> this is my root block device, which is a single partition turned into a pv, turned into a vg with 2 lvs on it
[01:29] <jkitchen> I just want to grow the partition so I can pvresize
[01:29] <jkitchen> really, I'd prefer it not have partitions at all and just expose the pv directly
[01:29] <TJ-> jkitchen: first change the PT with, say, fdisk, then do "partprobe /dev/sdX" to tell the kernel (check "cat /proc/partitions" before and after partprobe to be sure it was done)
[01:30] <jkitchen> seems partman-auto-lvm creates a partition though
[01:30] <jkitchen> TJ-: yea, but parted is giving me hell.
[01:31] <DeltaHeavy> jkitchen: You NEED partitions. Even LVMs need them. EVERYTHING needs partitions.
[01:32] <TJ-> DeltaHeavy: no, you don't need them, they're an anachronism
[01:32] <jkitchen> DeltaHeavy: I have PLENTY of filesystems on bare block devices
[01:32] <jkitchen> mkfs.ext4 /dev/vdb
[01:32] <jkitchen> works fine
[01:32] <DeltaHeavy> Really? I thought you needed them. Then again all I know is NTFS, EXT*, ZFS, and BTRFS.
[01:33] <jkitchen> partitions just take a big block device and make them a bunch of smaller ones
[01:33] <sarnold> DeltaHeavy: create a new zvol, then mke2fs on it directly -- no partitions needed :)
[01:33] <DeltaHeavy> I know, I thought they were required for most (I thought all) filesytems
[01:33] <jkitchen> DeltaHeavy: the filesystem doesn't care, it just sees a block device and writes to it
[01:33] <DeltaHeavy> Damn...didn't know that
[01:34] <jkitchen> /dev/vda1 is just as much a block device as /dev/vda :)
[01:34] <jkitchen> anywho
[01:34] <sarnold> fwiw, BSD systems have 'slices' instead, a separate partitioning mechansim that is different from the MSDOS-style partitions or GPT-style partitions..
[01:34] <jkitchen> yea
[01:34] <jkitchen> and solaris uses its on thing too (pre-zfs-root days)
[01:34] <TJ-> jkitchen: "parted resize" looks like it tries to be clever. Why not instead "parted rm 2" then "parted mkpart primary x y" ?
[01:34] <sarnold> and guess what, nothing checks that slices are entirely within a partition :) hilarity ensues..
[01:34] <DeltaHeavy> I thought you needed the MBR pointing to partitions and such for the BIOS to really access that stuff or something of the sorts.
[01:35] <jkitchen> TJ-: parted rm sounds scary.
[01:35] <jkitchen> :P
[01:35] <jkitchen> I'll try that though, sec
[01:35] <TJ-> jkitchen: Why? a partition is only 16 bytes in the PT ... that's all rm touches.
[01:35] <sarnold> it does -sound- scary though :)
[01:36] <TJ-> I guess I'm getting blasê about it then, I do it so much
[01:36] <jkitchen> DeltaHeavy: bios, perhaps. but GRUB stage 1 fits in MBR, loads stage 2 which has smarts about partition table types
[01:36] <jkitchen> you can sorta think of lvm as "linux native partition table" but it's not.
[01:36] <jkitchen> it's so much more
[01:37] <jkitchen> MBR != partition table
[01:37] <TJ-> lvm came from IBM AIX and the system360 and Z series , I think
[01:37] <jkitchen> ahh ok
[01:38] <TJ-> although the lvm tools on AIX are a poor relation to the Linux tools
[01:40] <jkitchen> AIX.
[01:44] <TJ-> I go with GPT now, GRUB 2 will install into an EF02 type partition of about 1MB for legacy boot, as well as having the EFI boot-strap in the EFI system partition
[01:50] <jkitchen> oof. this is not great.
[01:50] <jkitchen> I'm gonna take another approach :)
[03:09] <MavKen> any downside of using 64bit ubuntu server on vps with 1 GB RAM? or should I go with 32?
[03:12] <sarnold> MavKen: even though the 64 bit variant will waste some memory with pointers that are 8 bytes instead of 4 bytes, you'll have far more registers available and an extended instruction set, that should help programs run more quickly
[03:12] <sarnold> MavKen: I'd install the 64 bit version
[03:12] <MavKen> ok... I have 14 clients, all with small wordpress sites
[03:13] <MavKen> I've always heard 64 does not provide any benefit if you have less than 4GB ram
[03:13] <MavKen> thanks for the info
[03:17] <MavKen> I am new to having root access... is cron job best way to have "apt-get update && sudo apt-get -y upgrade && sudo apt-get -y dist-upgrade && sudo apt-get -y autoremove" run every day at midnight?
[03:18] <MavKen> and is that even a good practice?
[03:20] <sarnold> MavKen: better is to use the unattended-upgrades package
[03:21] <MavKen> sarnold, ahh..nice
[04:03] <MavKen> is there a way to change the from address exim4 uses?  The tutorial I was following listed the last config screen, but my last screen was about splitting config into smaller files.  I want to change it from root@domain.com to webmaster@domain.com
[04:04] <MavKen> I tried exim -brw webmaster@domain.com but it hasn't actually changed the from address
[04:18] <sarnold> MavKen: probably better would be to change whatever it is that is sending the email to use the address you want
[04:19] <MavKen> ok
[04:23] <Lord_Set> Anyone here really familiar with MAAS running on 13.10?
[04:25] <Lord_Set> Or should I try a different channel for MAAS/Ubuntu Cloud problems?
[04:26] <bigjools> Lord_Set: #maas
[04:26] <Lord_Set> Thanks
[08:48] <freakynl> hi just installed ubuntu 13.10 on a server here and it doesn't automatically boot (get the grub menu - but no timer counting down and after 5 mins it still stands there - not really desirable for a server :))
[09:47] <freakynl> Hi,
[09:52] <freakynl> There's a small bug in targetcli. It's just a python file that needs to replaced. This has been affecting ubuntu since 11.10 or 12.04. It's still present in 13.10. Apparently nobody is interested in updating packages between releases. I'm not sure on how to progress... the bug has been open for over a year now
[10:20] <adam007> well i setup some script (php) in hosting and it works, but why it not works on ubuntu server ?
[10:20] <adam007> i just install apache, php and im following tutorial in net tuts
[10:28] <freakynl> adam007: it doesn't work... why? what's not working? what does the error log report, etc. Unless you can bestow some serious mental powers on people they can't guess your issue
[10:31] <cfhowlett> !details|adam007,
[14:57] <toyotapie> Hello. I wrote an init configuration file and dropped it in /etc/init/, how can I specify that the service can only start if openvpn is running AND the tunnel is up ?
[15:05] <hitsujiTMO> !upstart | toyotapie check out the cook book. you need openvpn to fire events for that
[16:31] <joshu> Hi I need some advice. I have a local ubuntu server and a remote ubuntu server. I want to copy the whole local disk to the remote local disk. I'm thinking of booting the local server with the livecd, using dd to create a drive image and than scp to remote and restoring the image. Alternatively I'm wondering if rsync can do the job of copying directly thus skipping the dd image part.
[16:32] <yeats> joshu: so is the goal to mirror the same data, or what?
[16:33] <joshu> yeats yes correct the whole root volume
[16:37] <rbasak> I usually use rsync for this task. You'll need to partition and make filesystems at the other end, and then reinstall the bootloader.
[16:37] <rbasak> rsync -aH --numeric-ids, IIRC
[16:42] <joshu> is the data encrypted during transport with rsync
[16:56] <joshu> hmm gouging it seems as if dd is the way to go if I want exact copy of whole disk
[17:03] <semiosis> jamespag`: ping
[17:04] <jamespag`> semiosis, hello
[17:04] <semiosis> hi!
[17:04] <semiosis> just wanted to check in re: glusterfs MIR
[17:04] <semiosis> i saw your comment, but didnt know what to make of it
[17:04] <semiosis> people are asking me about gluster/qemu in ubuntu, idk what to tell them :)
[17:11] <ABC-XYZ> Yeah you can use rsync with ssh
[17:12] <ABC-XYZ> dd is usually pretty inefficient as it's a block level clone
[17:13] <semiosis> joshu: dd if=/dev/foo | ssh user@host "dd of=/dev/bar"
[17:14] <semiosis> joshu: or something like that :)
[17:14] <joshu> hey semiosis you stalker :P Yep I found that exact syntax on server fault. thanks
[17:14]  * semiosis everywhere
[17:15] <disposable> has anybody had any luck with the default version of LXC in 12.04? When I take these steps (apt-get install lxc; lxc-create -n test01 -t ubuntu; lxc-start -n test01 -d; lxc-console -n test01), all I end up with is a blank screen that says 'Type <Ctrl+a q> to exit the console'. It's like the container never really starts. What am I missing?
[17:19] <disposable> lxc-ls and lxc-info show it is running
[18:54] <jamespag`> semiosis, once the mir is complete, we can enable the feature in qemu
[18:54] <jamespag`> if we enable it now it just won't build until the mir completes which is not good
[18:55] <semiosis> is that happening for trusty?  just a matter of time?  or possible it will miss trusty?
[19:02] <jamespag`> semiosis, is being reviewed this dev cycle for trusty
[19:02] <jamespag`> that's no guarantee it will actually get to main
[19:29] <makara> i don't want to remember any passwords or passphrases beside the system login
[19:30] <joshu> I'm having trouble with permissions when doing this dd if=/dev/sda | gzip | ssh -i xxxxx.pem ec2-user@xxxxxxx 'gzip -d | dd of=/dev/xvdb'
[19:30] <joshu> I've tried adding sudo but then I get "sudo: sorry, you must have a tty to run sudo"
[19:30] <semiosis> jamespag`: ok, thanks for clearing up my confusion!
[19:30] <makara> i'm looking for an agent that can manage login keys for my servers
[19:31] <semiosis> jamespag`: if there's anything i can do to help it along, you know where to find me
[19:31] <makara> ssh-agent manages keys, but not passwords
[19:34] <makara> joshu, try `sudo -i`
[19:36] <joshu> didn't work :(
[19:38] <makara> joshu, output please
[19:39] <joshu> makara I solved it by using another ec2 instance. I think the AWS Linux AMI has some restrictions on sudo
[19:39] <joshu> on Ubuntu it worked fine
[19:41] <makara> joshu, did you create ec2-user?
[19:42] <joshu> yes that user is crewted automatically
[21:47] <bxz> is puppet completely broken in 13.10 or is it just me?
[22:09] <Jewsus> 'ello everybody
[22:10] <Jewsus> I have a question, does Ubuntu server's installation CD come with Realtek WiFi firmware?
[22:12] <sarnold> Jewsus: dpkg -L linux-firmware | grep -i rtl  shows a handful of RTL wifi firmwares; seeded-in-ubuntu linux-firmware  gives me the impression that it is probably on a server CD
[22:15] <Jewsus> kthx
[23:37] <miseria> "la verdadera felicidad de un ser humano, se logra cuando deja de ser esclavo, de la avaricia y la codicia" bienvenidos: http://castroruben.com *temo_a_un_ser_sin_rival*