[01:06] <lvmer> I could use some help with an lvm setup. I keep doing it wrong. I just did a fresh install of 12.04. So setup should be straight forward, but I could use some help.
[01:07] <xnox> lvmer: what's up. talk it out =)
[01:09] <lvmer> I'm trying to make a large stripped lv for a smb share
[01:09] <lvmer> no important / backup data will be on it
[01:10] <lvmer> xnox: I've got 3hd's  sda, sdb, sdc
[01:10] <lvmer> xnox: and right now sda seems to have a sub lv? for the OS [7gb] & swap [2gb]
[01:11] <lvmer> xnox: & 100gb free, how do I get the extra free space into a new vg = vg0 with sdb & sdc
[01:12] <lvmer> xnox: that's pretty much my question.
[01:12] <xnox> lvmer: checkout out output of $ sudo pvs; $ sudo vgs; $ sudo lvs
[01:12] <xnox> you probably want to make physical volumes out of your physical drives, then add those physical volumes into the volume group and then you can create big logical volumes
[01:13] <xnox> lvmer: read http://tldp.org/HOWTO/LVM-HOWTO/
[01:13] <lvmer> xnox: yah
[01:13] <xnox> it's very short and sweet =)
[01:13] <lvmer> lol
[01:13] <xnox> and tells you how to do all of this.
[01:13] <lvmer> yah I got all that... & it's easy on sdb & sdc, but I can't get sda to join
[01:14] <lvmer> the vg spans the whole sda, while it should only be for lv = root & lv = swap_1
[01:15] <lvmer> so I've got to like vg reduce, but it's weird xD
[01:17] <lvmer> xnox: want me to paste the cut out from pvs, vgs, lvs?
[01:17] <xnox> lvmer: right, it's usually to migrate one volume to a larger one. but if you really really want to resize it.
[01:18] <xnox> lvmer: resize the file-system first, then reduce the logical volume.
[01:18] <xnox> then migrate to stripped one......
[01:18] <xnox> but.
[01:18] <lvmer> xnox: I think the file system is already resized, I think I installed it that way with a 7GB partition.
[01:18] <xnox> if it's a new install, the easier way is to.
[01:19] <xnox> 1) connect all drives to a system & use command line from the installer to create a big vg across all of them & create your volumes
[01:19] <xnox> 2) do the install from d-i and select the pre-created logical volumes as targets
[01:20] <xnox> as a bonus you can use ubuntu 12.10 desktop cd to do the step 1, as it has all the lvm utilities =)
[01:20] <lvmer> yah I think I have 12.04 server
[01:20] <xnox> lvmer: sure, then you can create a big stripped partition. move your install from the 7GB logical volume to a big one. Then delete the small 7GB volume, enlarge the striped one.
[01:21] <lvmer> xnox: I see what you mean
[01:21] <lvmer> xnox: but I'd rather keep the system files separate from my eventual giant shared folder
[01:22] <lvmer> xnox: incase I have to purge data & can do it without messing up the shorewall configs etc.
[01:22] <xnox> lvmer: so what's the problem then?
[01:22] <lvmer> I'm scared to vgsplit xD
[01:23] <xnox> lvmer: why do you need vgsplit? the stippiness is the property of the logical volume, not the volume group (let me check that)
[01:24] <xnox> yeah. that's correct.
[01:24] <lvmer> http://paste.kde.org/607022/
[01:24] <xnox> and the way I did vgsplit, is to calculate it a few times & then add a 300 MB margin on top for safety.
[01:25] <lvmer> xnox: don't I need to get /dev/sda5 into it's own vg (ie: vg0)?
[01:26] <lvmer> xnox: instead of vg = US2
[01:26] <xnox> lvmer: $ sudo vgextend US2 /dev/sdb1
[01:26] <xnox> lvmer: $ sudo vgextend US2 /dev/sdc1
[01:27] <xnox> lvmer: $ sudo vgs
[01:27] <lvmer> xnox: but I think another vg would be better?
[01:27] <xnox> lvmer: no.
[01:27] <lvmer> xnox: can't I divide the hdd into 2x vg's?
[01:27] <lvmer> o
[01:27] <xnox> lvmer: a vg group is more or less the devices you expect to travel / be always attached together.
[01:28] <lvmer> xnox: OOOOOOOOOOOOOOOOOO
[01:28] <xnox> lvmer: typically you want it as large as possible.
[01:28] <lvmer> xnox: lmao... xD wow
[01:28] <lvmer> xnox: this whole time I'm trying to split vg's
[01:28] <xnox> lvmer: the strippiness/mirroring/snapshoting is all property of logical volumes that sit within the volume group and really have no clue what is backing the vg.
[01:28] <lvmer> can I rename it?
[01:29] <xnox> lvmer: yes.
[01:29] <lvmer> sudo vgrename US2 vg0
[01:29] <xnox> lvmer: don't forget to run $ update-grub & update-initramfs and check the boot menu after you do =))))
[01:29] <lvmer> xnox: ooooo boi
[01:31] <lvmer> xnox: ah this makes more sense now as vg0
[01:31] <lvmer> updating grub
[01:32] <lvmer> xnox: so update-initramfs -u
[01:32] <lvmer> xnox: how should I check the boot menu?
[01:32] <lvmer> fstab?
[01:32] <xnox> lvmer: less /boot/grub/grub.cfg
[01:33] <xnox> lvmer: fstab is a good one as well =)))))
[01:33] <lvmer> lol
[01:33] <lvmer> was it a bad idea to do this?
[01:33] <lvmer> am I just going to have to re-install again lol. changing the vg of the boot lvm xD
[01:34]  * xnox had a typo in my vg at install. I did have a facepalm over "intarnelhdd" vg name
[01:34] <xnox> and renamed it =)))))
[01:34] <lvmer> xnox: and you were ok? ok xD
[01:34] <xnox> lvmer: no, you don't have to reinstall.
[01:35] <xnox> lvmer: yeah, it feels like black magic and pixie dust, but it's really stable device-mapper stuff =)))))
[01:35] <lvmer> ok
[01:35] <lvmer> should I do a test reboot? xD
[01:36] <lvmer> xnox: well hopefully it works
[01:37] <lvmer> xnox: if not... it's down to the basement
[01:37] <lvmer> xnox: k it worked xD
[01:38] <xnox> lvmer: born again ;-)
[01:39] <lvmer> xnox: yeeehawww. :)  So now I just need to create a massive lv
[01:40] <xnox> lvmer: well, create it ~50% capacity. Once you fill it up you can increase/decrease it, snapshot it, etc
[01:41] <lvmer> xnox: so $ lvcreate -i3 -L 300G -n lv1 vg0
[01:41] <lvmer> xnox: ?
[01:42] <xnox> lvmer: looks good =)
[01:43] <lvmer> what is a good stripe size? 4kb? 8kb?
[01:43] <lvmer> I think it's -I4
[01:45] <sarnold> lvmer: with no data to back me up, I'd assume _much_ larger sizes, something like 128kb
[01:47] <bananapie> Hi, I have three https servers running on the same IP. ports 443, 543 and 544. 443 has a private key and signed certificate. 543 and 544 both have the same private key, but each port has a different signed certificate. This is confusing apache, and apache is announce the certificate from 544 on 543. any ideas?
[01:49] <qman__> you specify which certificate file to use in your virtualhost configuration; there's nothing to confuse
[01:50] <qman__> ensure your virtualhosts specify which port they're on, and which certificate file to use
[01:50] <bananapie> Yes, that's all setup
[01:50] <bananapie> <VirtualHost *:544>
[01:50] <bananapie> I have SSLCertificateFile and SSLCertificateKeyFile in each of the 3 hosts.
[01:51] <bananapie> I think apache is confused because it has two different certificates for the same private key.
[01:51] <sarnold> bananapie: can you re-create that bug on a VM or something?
[01:51] <sarnold> two new keys, three new certs, and a bare-bones apache config?
[01:52] <xnox> lvmer: 4MB ?
[01:52] <bananapie> I'll try
[01:52] <qman__> yeah, there shouldn't be a problem with that setup, so it's either an apache bug or an error in the config
[01:52] <bananapie> ok
[01:53] <bananapie> I'll check if there are any apache2 updates
[01:53] <qman__> one possible workaround attempt in case of an apache bug would be to make a copy or hard link of the key by a different name, so there's three files according to the config but it's really the same key
[01:54] <lvmer> xnox: I keep getting not enough allocatable extents
[01:55] <lvmer> xnox: my command was: $ sudo lvcreate -i3 -I512K -L 350G -n lv1 vg0 /dev/sdb1 /dev/sdc1 /dev/sda5
[01:57] <xnox> lvmer: $ sudo vgs ?
[01:57] <lvmer> xnox: shows 400GiB free
[01:57] <lvmer> xnox: 400.34g
[01:58] <lvmer> xnox: vsize = 409.64g  vfree = 400.34g  #pv = 3  #lv = 2 #sn = 0
[01:59] <lvmer> xnox: perhaps: -I512K   is throwing it off.
[01:59] <xnox> lvmer: for some reason in cannot allocate enough space for that volume. try lower size.
[01:59] <lvmer> xnox:  Rounding size (98560 extents) up to stripe boundary size (98562 extents)
[01:59] <lvmer>   Insufficient suitable allocatable extents for logical volume lv1: 20043 more required
[01:59] <lvmer> does that seem weird?
[02:00] <xnox> lvmer: can you try $ sudo lvcreate i3 -I4 -L350G -n lv1 vg0
[02:00] <lvmer> xnox: each extents is 512kb?
[02:01] <xnox> and see if that uses less extents.
[02:01] <xnox> lvmer: hah =)))
[02:01] <xnox> lvmer: the maximum cap you can have is the maximum free space on a single physical volume
[02:01] <lvmer> xnox: NO!
[02:02] <xnox> lvmer: otherwise it will not be real stripes =)
[02:02] <lvmer> xnox: idc if it's not real stripes
[02:02] <lvmer> xnox: dang..... so it won't do it cause 1 HD is 100gb? while the others are 200gb?
[02:02] <qman__> the stripes have to go somewhere
[02:03] <xnox> lvmer: then got for two stripes, they will be spread across 3 drives and then you can have >200GB volume
[02:03] <lvmer> xnox: so I just change i3 to i2?
[02:03] <xnox> lvmer: maybe you don't want stripes at all.....
[02:03] <lvmer> xD
[02:03] <xnox> lvmer: yeah. and it's better to use 4MB stripe.
[02:03] <lvmer> xnox: it's looking like I don't want stripes at all
[02:04] <lvmer> xnox: how does 2 stripes work across 3hdds?
[02:04] <xnox> lvmer: just create a volume with name and max size and then the full free space will be "a single filesystem"
[02:05] <lvmer> xnox: I guess I'm going to unfortunately give up on stripes.... one day though.. one day I shall stripe lots of 2TB HD's. xD
[02:05] <xnox> lvmer: because it needs two places to store each extend in the vg, and the vg has three physical buckets with free space.
[02:06] <xnox> lvmer: hence there will be spread across them, since they are all "available".
[02:06] <lvmer> lv's can definitely be resized right?
[02:06] <lvmer> xnox: smaller / larger right? xD
[02:07] <lvmer> xnox: nvm stupid question
[02:07] <qman__> lvmer, it will keep two copies of each chunk of data, and put them on any two of the available disks
[02:07] <xnox> lvmer: yes. but with reducing you resize the filesystem inside it first, with enlarging you resize the fs after.
[02:07] <qman__> I don't know what it does to decide that but I assume it just picks wherever the most space is available
[02:08] <qman__> or otherwise does it as evenly as it can
[02:08] <lvmer> xnox: oh so is i2 = raid 1 ? cause I was looking for a raid 0 (similar) file system.
[02:09] <lvmer> xnox: I thought all the extends would be in 1 place
[02:12] <lvmer> xnox: well I got a 400gb lv1 now... so this is awesome. Thank you very much
[02:13] <xnox> lvmer: https://docs.google.com/a/surgut.co.uk/document/preview?id=1bZ4yQIVgGaUGSYu3qiUHnQt3ieBZoqunP_DcleHCr3I#heading=h.8acu9ne05kh9
[02:13] <xnox> lvmer: is a table comparing raid levels and performance gains
[02:13] <xnox> lvmer: raid0 theoretically gives you N write speed, but the read speed is still x1
[02:14] <xnox> lvmer: while raid1 gives you xN read speed, with ~x1 write speed, at the cost of 1/n space.
[02:14] <lvmer> xnox: yah I know about raid's xD
[02:15] <xnox> lvmer: and lvm implements RAID1, 0 & 1-0 raid levels ;-)
[02:15] <xnox> with 0 level by default.
[02:15] <lvmer> xnox: and stripping would be a RAID 0 example? right? the disks just apparently have to be the same size like a regular raid
[02:16] <lvmer> xnox: ohhhhh 0 is default?
[02:17] <xnox> lvmer: kind-of..... each physical volume is treated equivalent, so if you have /dev/sda1 /dev/sda2 /dev/sda3 and /dev/sdb
[02:17] <xnox> it will not realise that actually there will be little performance gain spreading across /dev/sda*
[02:18] <xnox> lvmer:  but checkout the --type option
[02:18] <xnox> lvmer: for segment type.
[02:21] <lvmer> how do you search a 'man' file?
[02:21] <lvmer> man lvcreate
[02:22] <xnox>  /
[02:22] <lvmer> xnox: trying to search for type
[02:22] <xnox> man man
[02:22] <xnox> also helps
[02:22] <xnox> or 'h'
[02:22] <xnox> for help while viewing a man page
[02:24] <lvmer> xnox: $ mkfs -t ext3 /dev/vg0/lv1     or -t ext4?
[02:26] <sarnold> lvmer: I sometimes do MANPAGER=cat man foo | grep   or that sort of thing; 'less' is a very nice pager, but some tasks are easier done with other tools, and MANPAGER is a great way to get there
[02:36] <lvmer> Ok I'm trying to auto mount this new lv in $ /etc/fstab   but I have some questions
[02:37] <lvmer> the old files are still /dev/mapper/US2-root   instead of vg0?
[02:37] <lvmer> should the new one be /dev/mapper/vg0-lv1   ?
[03:41] <bfortified> I installed owncloud if i forward port 80 from the wan will i be vulnerable to attack can I change the port to somthing uncommon like i do with ssh?
[03:45] <SpamapS> bfortified: some apps are stupid and redirect base on Host: without the port.. but owncloud is probably not that stupid.
[03:47] <bfortified> SpamapS: i dont think I understand what you meen I have tried a dif port on apache and that appears to have worked, I navigate to that port and get the congrats your server is running but i caint seem to acces own cloud
[03:48] <bfortified> figured it out <ip>:port/owncloud
[03:51] <qman__> changing what port you use won't render you invulnerable
[03:51] <qman__> it just means you're a little harder to find
[03:54] <qman__> and by a little, I mean very little
[03:54] <qman__> the average chinese bot won
[03:54] <qman__> 't find you, and that's about it
[03:54] <qman__> smarter bots and any human will
[04:00] <patdk-lap> hell, google/bing/... will all find you very fast
[04:00] <patdk-lap> cause the first time you use a browser with their toolbar installed, it reports back to them, and you will get scanned in a few min, if not faster
[04:00] <patdk-lap> my friend tried that :)
[04:01] <patdk-lap> put up a new website, on a non-standard port, and noticed google search had already indexed it and it was searchable within 2min, and he realized he had google toolbar installed
[09:14] <Daviey> Mornin' people! How is everyone is glorious day?
[09:15] <koolhead17> hello Daviey !! Its bright & sunny as usual with humidity
[09:15] <koolhead17> :p
[09:18] <Daviey> koolhead17: I uspect your bright & sunny is in a different league to mine :)
[09:18] <Daviey> suspect*
[09:18] <koolhead17> Daviey: yes its added with humidity and topical climate :D
[09:19] <Daviey> topical you say? :)
[09:20] <koolhead17> or equatorial where sun god literally vomits his heat with ultraviolet radiation :D
[09:21]  * smb thinks koolhead17 is missing Daviey pulling his leg...
[09:21] <koolhead17> *tropical
[09:21] <koolhead17> smb: :D
[09:22] <smb> :)
[09:22]  * Daviey must now be very careful he doesn't ever make a typo aguin.
[09:23] <smb> Daviey, We don't do typos, that is artistic freedom. :-P
[09:23] <koolhead17> smb: heh
[09:24] <Daviey> :)
[11:06] <Arne__> hello, I have a litte problem with my ubuntu server, I assigned a static IP to eth0, but ubuntu changes this ip sporadic to another .. someone who can tell me how I can change this and where?
[11:08] <xranby> Projektarybte: edit /etc/network/interfaces  http://www.fogproject.org/wiki/index.php/Ubuntu_12.04_static_ip_configuration
[11:09] <Projektarybte> xranby: this is exactly what I have in my config
[11:09] <xranby> is dhcpd still running?
[11:09] <xranby> dhclinet
[11:09] <xranby> dhclient
[11:10] <Projektarybte> oh, yes it´s running
[11:11] <Projektarybte> but I declared the network config as static
[12:45] <darthanubis> drush installs fine, but I can't use "You are running the provision script as the root user. Exiting" It does not like sudo
[13:09] <patdk-lap> darthanubis, hmm? isn't that ovious? don't run it as root, so you run it as root with sudo?
[13:27] <dsop> SpamapS: ping
[14:54] <GhostFre_> How often does Ubuntu release security updates? Do you have like a patch thursday
[14:56] <TheLordOfTime> GhostFre_, security updates are controlled by the security team, if i remember things right, they're in #ubuntu-hardened.  They get released as they're processed, I think, though...
[14:59] <LauraA> Hi.  I just installed a new Ubuntu 12LTS Server (switching from Centos).  Everything's humming along except on boot, it just sits at the grub2 selection menu -- it doesn't auto-select the 1st entry.  If I hit 'Enter', everything continues on as normal.
[14:59] <LauraA>  This is changed behavior.  A week or so ago, it *did* autoboot OK.  Where do I look for the cause/fix of this?
[15:03] <smb> LauraA, One reason for not autobooting could be a failed previous boot, but then it should resume autobooting after one successful boot.
[15:24] <LauraA> smb: Hi.  I didn't know a prior failed boot 'gets in the way'.  I don't remember there being any failed boot, though.  Is there some way to verify/validate that that happened?  In the logs I guess .. but what to look for?
[15:27] <smb> LauraA, It would be some file that grub looks at (though I don't know from the top of my head which one). And that actually should be removed when the server boots to the prompt once, so the next boot would be normal again. Otherwise all the timeout values are normally set in /etc/default/grub and then configured when you run update-grub
[15:29] <smb> So the grub env is changed here in /etc/init/grub-common.conf
[15:29] <smb> grub-editenv /boot/grub/grubenv unset recordfail
[15:30] <smb> "grub-editenv /boot/grub/grubenv list" should show if recordfail is currently set
[15:31] <samba35> my ubuntu server (12.04.1 ) is on vmware  and ubuntu is a dmz  (192.168.3.100/24)in my network (using utm on vmware also ) i have wireless tv (192.168.200.117 )  how do connect wireless tv to ubuntu to use minidlna
[15:32] <LauraA> smb: Ok thanks.  I've got some sleuthing to do.  May be back later.  Cheers!
[15:40] <SpamapS> dsop: pong?
[16:02] <zul> hallyn: ping does this mean what i think it means http://libvirt.org/git/?p=libvirt.git;a=commit;h=2e03b08ead603c38c244aa9a1ecef6d73bb306be
[16:04] <hallyn> zul: <blink>  pretty sure it does
[16:04] <hallyn> zul: they have selinux, it has its own mknod denials...  why not use that?  huh
[16:04] <zul> hallyn: awesome
[16:05] <hallyn> zul: yeah, especially at the end of this cycle when we have stacking apparmor policies and user namesapces,
[16:05] <hallyn> libvirt-lxc is gonna be a mess to try and keep up with lxc
[16:05] <zul> *sigh*
[16:06] <zul> Daviey: ^^^
[16:07] <Daviey> ugh
[16:09] <roaksoax> jamespage: ok so i will get this up and running and test the ha stuff works
[16:09] <roaksoax> then will start modifying the hacluster charm
[16:10] <jamespage> roaksoax, OK - I'm really interested in seeing how the sync's happen between quantum and the hacluster sub
[16:10] <jamespage> for example when I 'add-unit quantum'
[16:10] <jamespage> quantum will get installed first, and then the hacluster stuff
[16:11] <jamespage> ...
[16:11] <jamespage> roaksoax, if you just want to test quantum you can run a much more minimal charm set
[16:12] <jamespage> keystone, mysql, rabbitmq-server and quantum should be enough
[16:13] <roaksoax> jamespage: yes yet you need to stuff works at least I have tried it with HAproxy
[16:13] <jamespage> roaksoax, I'm still seriously considering a refactor to push the quantum-server into the cloud-controller
[16:14] <roaksoax> errr the cluster stuff works with add unit
[16:14] <roaksoax> I agree but the server should probably be with the cloud controller
[16:15] <roaksoax> s/but/with
[16:17] <hallyn> note to self: make defconfig != make oldconfig !
[16:26] <jamespage> roaksoax, gonna talk with adam next week about that - it feels awkward
[16:28] <roaksoax> jamespage: the other option would be to have enough to separate charm then
[16:29] <roaksoax> so that it got me the flood in the cloud controller independently
[16:29] <roaksoax> err so it can be deployes
[16:29] <roaksoax> freakibg voice recognition
[16:32] <roaksoax> jamespage: there is going to be a problem if we remove unit and add unit again to reuse the mode though
[16:39] <lvmer> Hi, I have some netbios, wins, or name resolution issues. Can someone help me troubleshoot them?
[16:41] <lvmer> My samba share folders will not show up on the network, even without a firewall. & pinging from a different server yields closed ports.
[16:41] <lvmer> 139,445
[16:45] <jamespage> roaksoax, I think we should ignore the node reuse issue for the time being
[16:46]  * jamespage burys his head in the sand
[16:46] <jamespage> until the stop hook works or units are automatically terminated on remove-unit its never going to be pretty
[16:52] <roaksoax> jamespage: indeed. i think its worth pushing for any of those features
[17:07] <MontyMoose> Hello - quick bit of help if anyone's free. Running LTS 12.4. I have a rogue PHP script which runs from time to time and bogs the php-fpm process until I kill it. Problem is I have about 1000 php scripts running on the server and about 40,000 hits a day. I can find the PID from TOP, but can I find out which script is running on that PID? Which php process or which file etc?
[17:07] <MontyMoose> Ideally I want to know   PID 5475 = /var/www/thisfile.php
[17:08] <blkperl> MontyMoose: lsof
[17:08] <sarnold> MontyMoose: probably lsof or fuser can get you there
[17:08] <MontyMoose> thanks - I'll take a look
[17:08] <sarnold> MontyMoose: you do run the risk that the php interpreter you're using may have byte-code compiled it all once at first use, and might not be able to blame the right script, but it's a good starting point.
[17:09] <MontyMoose> pointing me in the right direction is a good start
[17:09] <sarnold> MontyMoose: if that doesn't work, ask about strace when you return. it's complicated and ugly :) but it's a good tool. hehe.
[17:09] <MontyMoose> I'm complicated and ugly - I'm sure we'll get on like a house on fire...
[17:11] <MontyMoose> hmm - lsof seems only to log that php-fpm is running, doesn't show me which script it's running.
[17:12] <MontyMoose> I think there might be a solution in PHP itself - there's a getpid() function which I could potentially log to mysql each time a script runs, but I'd rather not
[17:15] <MontyMoose> the other thing that's odd is that I tell PHP to timeout after 20 seconds, and this process will go on for ever until I kill -9 it
[17:15] <sarnold> MontyMoose: oh man :/ I was afraid of that.
[17:16] <MontyMoose> because?
[17:16] <sarnold> MontyMoose: you can attach strace to a process with strace -p pid
[17:16] <MontyMoose> rigiht
[17:16] <sarnold> MontyMoose: .. because debugging it by strace is ugly. :)
[17:16] <MontyMoose> hm
[17:16] <sarnold> MontyMoose: the horrible part comes from watching what it does and trying to figure out which of your scripts does those things. if you can catch sql being sent to mysql, you can grep for those, maybe narrow it down to a handful to instrument
[17:17] <MontyMoose> I see
[17:17] <SpamapS> MontyMoose: I've had this problem as well.
[17:18] <MontyMoose> of course as it's now working perfectly - I can't find a PID which stays long enough to trace!
[17:18] <SpamapS> MontyMoose: Logging the pid is the only way I solved it
[17:18] <MontyMoose> in php?
[17:18] <SpamapS> MontyMoose: yeah, the problem is that php opens the scripts, reads them all, parses them, and closes them, so lsof won't help you
[17:18] <MontyMoose> I thought that might be the case
[17:19] <SpamapS> MontyMoose: What you need to log is the URL
[17:19] <SpamapS> MontyMoose: and potentially the arguments
[17:19] <MontyMoose> yeah
[17:19] <MontyMoose> well I suppose once you've put a function together you can pretty much log anything you want/
[17:19] <MontyMoose> is it quicker to text or mysql?
[17:19] <SpamapS> MontyMoose: logging in mysql seems... overkill. Just log to a file on disk.
[17:20] <MontyMoose> ok
[17:20] <SpamapS> file_put_contents('/tmp/pid.log', $_GLOBALS['REQUEST_FILENAME'], FILE_APPEND);
[17:20] <MontyMoose> if I include other PHP files will they run in the side PID?
[17:21] <SpamapS> MontyMoose: yes
[17:21] <MontyMoose> *same
[17:21] <SpamapS> MontyMoose: include just inserts the code in the spot where the include is
[17:21] <MontyMoose> ok - so I need this in my global header really
[17:21] <SpamapS> MontyMoose: my example forgot to include getpid(), but you get the point :)
[17:21] <MontyMoose> yup
[17:21] <SpamapS> MontyMoose: you can enforce a global include in php.ini
[17:22] <MontyMoose> that's a good idea
[17:22] <MontyMoose> then every file has to run it
[17:22] <MontyMoose> I could wipe the file every night with a cron job, as I only need it if I notice a problem
[17:22] <SpamapS> MontyMoose: auto_prepend_file
[17:23] <SpamapS> MontyMoose: another option is actually to do an auto_prepend which creates a file per-pid, and then an auto_append that removes it
[17:24] <SpamapS> MontyMoose: this would be fairly awful on disk, so make sure its only in /tmp, but that would let you see it for the pid that goes haywire without filling up disk w/ log
[17:24] <MontyMoose> oh I see, so any processes which are running at that precise moment will have a PID open
[17:24] <SpamapS> MontyMoose: and when I say only in /tmp, I mean only in a tmpfs partition
[17:24] <MontyMoose> yeah
[17:24] <MontyMoose> could I store it to a session_variable?
[17:25] <MontyMoose> oh no - that would only work on the users machine
[17:25] <MontyMoose> daft idea = forget that :-)
[17:26] <MontyMoose> and we're not worried that file_put_contents will put a big strain on things?
[17:26] <MontyMoose> server gets about 50,000 hits a day
[17:30] <MontyMoose> thanks for all that - think I've got a plan now
[17:32] <SpamapS> hah, 50,000 / day == 0.58 requests per second
[17:32] <SpamapS> I think a rasbperry pi could handle that
[17:37] <sarnold> SpamapS: static, yes, running php and doing mysql, maybe it's pushing it a bit..
[17:46] <SpamapS> sarnold: not even
[17:47] <SpamapS> sarnold: you have 1.7 seconds to finish every request...
[17:47] <sarnold> SpamapS: somehow I read that as .58 seconds per request...
[17:48] <sarnold> SpamapS: that does make more sense, even with the limited ram on board :)
[17:57] <hallyn> utlemming: around?
[17:57] <utlemming> hallyn: yup
[17:58] <hallyn> utlemming: is there a bug in cloud-init in oneiric re handlnig of ssh public keys?
[17:58] <hallyn> (put another way - i think ther eis :)
[17:58] <utlemming> hallyn: not that I am aware of...what are you seeing?
[17:59] <hallyn> if i lxc-create -t ubuntu -n o1 -- -S ~/.ssh/id_rsa.pub -r oneiric, the resulting container does not have my ssh key in ~/ubuntu/.ssh
[17:59] <hallyn> utlemming: but with -r precise, it does
[17:59] <hallyn> uh, -t ubuntu-cloud, sorry
[18:00] <utlemming> hallyn: interesting...okay, that is likely a bug with the lxc-create template
[18:01] <hallyn> utlemming: the key info IS in /var/lib/cloud/seed/nocloud-net/meta-data in the guest
[18:01] <hallyn> which is why i figured it was a cloud-init bug, not template
[18:29] <subman> I don't seem to be able to get my sendmail to work.  I keep getting the following error in my returned email: http://pastebin.com/daEnLx4B
[18:31] <subman> Here is my sendmail.mc: http://pastebin.com/VrN1qDcv
[18:47] <roaksoax> jamespage: or better yet, still around?
[18:49] <ScottK> subman: Most people in Ubuntu use postfix.
[18:49] <roaksoax> hggdh: ping
[18:49] <roaksoax> hggdh: do you need sponsorship of cobbler to -proposed?
[18:50] <roaksoax> or has it been uploaded already?
[18:50] <hggdh> roaksoax: I certainly do. Can you please check if it is now kosher?
[18:50] <roaksoax> hggdh: is this were the SRU lieS? https://launchpad.net/~hggdh2/+archive/ppa
[18:51] <hggdh> roaksoax: no, it is in https://code.launchpad.net/~hggdh2/ubuntu/precise/cobbler/lp-967815/+merge/132172
[18:51] <hggdh> and bug 967815
[18:53] <roaksoax> hggdh: cool, I'll upload it to precise-proposed for the SRU to be procesed then
[18:53] <hggdh> roaksoax: IOU, thanks
[18:53] <hggdh> roaksoax: I am still to prepare a merge for Quantal
[18:54] <roaksoax> hggdh: has this been fixed in raring though?
[18:54] <hggdh> ugh
[18:54] <hggdh> roaksoax: IDK, let me check
[18:54] <roaksoax> hggdh: alright, we need to fix it in raring... :)
[18:54] <hggdh> if the cobbler version is the same... then we need it there also
[18:54] <hggdh> on it
[18:55] <roaksoax> hggdh: yeah, ok, I'll strip the patch from your branch, test it and upload it
[18:55] <hggdh> roaksoax: thank you
[18:55]  * hggdh cancels the Raring cobbler branch creation
[18:57] <hggdh> yes, we need it on raring
[19:01] <lvmer> I've got a samba problem. I can't see 1 of my servers on the network. It should be identical to my 2nd one. I think there is a name resolution or netbios or wins conflict?
[19:01] <roaksoax> hggdh: does this look sane to you then? http://paste.ubuntu.com/1360923/
[19:03] <jamespage> roaksoax, yep
[19:04] <roaksoax> jamespage: ok so i was thinking this:
[19:04] <roaksoax> jamespage: the hacluster charm, sets in the config what resource to configure in terms of HA. it tells the name and type/resource to use
[19:04] <roaksoax> jamespage: currently, this is limited to only 1 resource per associated charm
[19:05] <sarnold> lvmer: does smbclient -L -I<IP> show you information from both  servers alright?
[19:05] <roaksoax> jamespage: so, to make things better, i was thinking on removing the definition of the resource to HA in the config for the charm (say HAProxy), and simply do that in the relation
[19:05] <sarnold> lvmer: (forgive me, it's been a decade since I've had to debug smbd, my commands are liable to be wrong, but I hope the ideas are there...)
[19:06] <roaksoax> jamespage: so, each charm to integrate HA, can simply send a list of resource to HA, with parameters, that will be set (an unchangeable) for the relation, rather than declaring them in the config
[19:06] <roaksoax> jamespage: this way, if a charm needs various resources in HA, then it can simply send the list with its options, that the hacluster will interpret and execute
[19:06] <roaksoax> acoordingly
[19:06] <roaksoax> jamespage: now, this could also allow me to define resources_active_passive and resources_active_active
[19:07] <roaksoax> so there would be a list for each type of resource
[19:07] <roaksoax> jamespage: does that make better sense?
[19:09] <lvmer> sarnold: not enough '\' characters in service... I ran \\192.168.0.18\share\    it leaves me with a  blank line
[19:09] <hggdh> roaksoax: it does sound correct
[19:10] <sarnold> lvmer: how about \\\\192.168.0.18\\share\\ ?
[19:10] <sarnold> .. double all the \ to get them through the shell..
[19:10] <JoeVLcek> Can someone please help me with a cloud-init issue regarding the default user?
[19:10] <JoeVLcek> cloud.cfg now has: users: -default
[19:11] <JoeVLcek> how does one specify the user name? for example how do I specify ec2-user ?
[19:11] <JoeVLcek> is the new user functionality documented some place that someone can point me to?
[19:12] <JoeVLcek> smoser: ping
[19:12] <lvmer> sarnold: it asks for root password xD  that should be disabled. So I can't get past it xD
[19:13] <sarnold> lvmer: there's command line options to tell it which user accout to try to use..
[19:13] <lvmer> sarnold: doh
[19:14] <lvmer> sarnold: ok it ran
[19:15] <lvmer> sarnold: command seems to execute, but no list back on either 17 or 18.  ip 17 still shows up in windows explore on LAN computers though.
[19:17] <roaksoax> hggdh: alright, I uploaded to raring, quantal-proposed, precise-propoesd
[19:17] <roaksoax> hggdh: so it is just matter for someone on the SRU team to review, and make it available for verification
[19:17] <hggdh> roaksoax: one day I will be as fast as you...
[19:20] <sarnold> lvmer: but both think they're alive? did eithe rknow about the other?
[19:21] <lvmer> they shouldn't know about eachother
[19:21] <lvmer> sarnold: they shouldn't know about eachother
[19:22] <sarnold> lvmer: no? I thought the whole point of magic nmbd things was that everyone knew about each other so they could elect leaders and the like?
[19:22] <lvmer> sarnold: I think it's a name resolution or net bios problem, because both samba configs are idential & all shorewall ports are accept
[19:22] <lvmer> sarnold: idk if I have magic nmdb
[19:23] <sarnold> lvmer: ps auxw | grep nm, I bet it's there... :)
[19:24] <hggdh> roaksoax: so I can delete the merge request I did for Precise, correct?
[19:24] <roaksoax> hggdh: yes :)
[19:24] <hggdh> done :-)
[19:24] <lvmer> sarnold: aaahhhh.. it is on the working server nmbd -D,   but on 18..... it just says --color=auto nm
[19:26] <roaksoax> /win/win 8
[19:26] <roaksoax> err
[19:27] <sarnold> lvmer: hrm. :) I wonder what kept nmbd from running on 18?
[19:28] <lvmer> sarnold: no idea... I understand it is part of the samba package.... but how do I install it again?
[19:29] <jamespage> roaksoax, I think that makes alot of sense - so the principle charm says to the hacluster subordinate - ha this stuff
[19:30] <sarnold> lvmer: try running 'service nmbd status'
[19:30] <lvmer> sarnold: you think this is an easy fix? or should I just reinstall and then restore the configs? Because I started nmbd but samba still doesn't show and nmap on another computer shows 18's 139 & 445 ports still closed
[19:30] <sarnold> lvmer: and 'service nmbd restart' is probably a reasonable idea too; check /var/log/ to see if you can find the nmbd logs, they may be stuffed in a samba directory or something.
[19:30] <jamespage> roaksoax, if you do that through the container relation then it can also use that relation presence to determine what todo with restarts of services etc... and coor with the hacluster charm
[19:30] <lvmer> sarnold: stop/waiting
[19:31] <roaksoax> jamespage: yeah! I'll have the charm modified and test it out
[19:32] <lvmer> sarnold: /var/log/samba/log.nmbd
[19:33] <lvmer> sarnold: no network interfaces found
[19:33] <lvmer> sarnold: lib/interface.c:543(load_interfaces)
[19:34] <sarnold> lvmer: ooh very curious :) I wonder why -- misconfigured?
[19:34] <sarnold> does it have .17 address in its listen line or something?
[19:34] <lvmer> sarnold: I bet so!  http://paste.kde.org/607436/
[19:35] <lvmer> sarnold: where is the network interface file I should be looking at? xD
[19:36] <sarnold> lvmer: it might get it out of smbd.conf or whatever..
[19:37] <lvmer> sarnold: does this seem bad to you? http://paste.kde.org/607442/
[19:38] <sarnold> lvmer: does 'ip addr show' show you the 'em1' interface?
[19:39] <lvmer> sarnold: no... should it be 'eth0'
[19:40] <sarnold> lvmer: probably; whichever interface the machine should use. :)
[19:40] <lvmer> sarnold: how come .17 also has that command and no em1 interface?
[19:41] <sarnold> lvmer: no idea :)
[19:42] <lvmer> sarnold: well the ports on .18 are open now
[19:42] <lvmer> sarnold: but still no windows explore access
[19:43] <sarnold> lvmer: mmhm, that may take a while, i found sometimes the easiest way to force windows clients to see new systems was reboot them or maybe reboot the old master, force a new election
[19:43] <lvmer> sarnold: can you have 2 masters? I think that might be a problem for me.
[19:44] <sarnold> lvmer: only one master.
[19:44] <lvmer> in the smb config right?
[19:44] <lvmer> what if you have none?
[19:46] <sarnold> lvmer: they elect themselves
[19:46] <sarnold> lvmer: the 'master = yes' just says that that machine will participate in elections
[19:46] <lvmer> ok rebooting all
[19:47] <lvmer> sarnold: I put perfered master = no for one of them just incase. But domain master = yes for both
[19:47] <sarnold> lvmer: sounds good.
[19:48] <lvmer> sarnold: lol .17 doesn't work now lol
[19:49] <sarnold> lvmer: o_O
[19:50] <lvmer> sarnold: I set both to interfaces = eth0 em1
[19:50] <lvmer> sarnold: rebooting
[19:52] <lvmer> sarnold: no network interfaces found lol
[19:52] <lvmer> sarnold: nmbd did not start on either
[19:53] <sarnold> lvmer: ooof. :)
[19:57] <lvmer> sarnold: well the both have the same open ports from nmap. It's got to be an interface problem or something. .17 & .18 not accessible... but from the logs .17 thinks it is the master & has no errors.
[19:57] <lvmer> if my network interface is eth0 ? can't I use that for samba? why did it work when it was set to em1?
[20:01] <sarnold> lvmer: no idea there. :/
[20:22] <lvmer> sarnold: I got it! I redid the entire smb.conf & both servers work now. Thanks a ton for the trouble shooting help. There were definitely some network name & bind errors.
[20:54] <ZLoy> Hello one little question - i have server with ubuntu and want translate stream from my laptop with Traktor to that server. What software for streaming is better?
[20:56] <ZLoy> somthing like translate sound stream to internet from my dj rig..
[21:34] <subman> Any idea why my server would suddenly just stop serving web pages?  It was working just awhile ago.
[21:35] <SpamapS> subman: stop serving web pages how?
[21:35] <sarnold> subman: anything in the logs?
[21:35] <SpamapS> subman: like, 500 errors, or never responds?
[21:36] <subman> SpamapS, 507 and 508 codes found
[21:37] <SpamapS> never seen 50x's before
[21:37] <genii-around> Maybe /var/log/apache2/error.log   might hold some clues.
[21:37] <SpamapS> you're all assuming apache. ;)
[21:38] <SpamapS> probably a good assumption, but still. :p
[21:38] <subman> Yes, apache, sorry
[21:38] <subman> Here is that last line in the access log:
[21:38] <subman> 173.206.102.89 - - [15/Nov/2012:16:37:36 -0500] "GET / HTTP/1.1" 404 508 "-" "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:16.0) Gecko/20100101 Firefox/16.0
[21:38] <genii-around> 507 is Insufficient Storage
[21:38] <subman> df
[21:38] <subman> oops
[21:39] <subman> Last line of error log:
[21:39] <subman> [Thu Nov 15 16:37:36 2012] [error] [client 173.206.102.89] script '/var/www/index.php' not found or unable to stat
[21:39] <subman> The index.php file is there
[21:41] <genii-around> Then likely "unable to stat"  might be due to these WebDAV errors.
[21:53] <subman> How to check?
[21:53] <subman> It was working literally an hour ago
[22:06] <subman> Could a bad .htaccess be causing this?
[22:09] <genii-around> subman: I would more suspect that you're just running out of room on your server.
[22:10] <subman> genii-around, On the hard drive?
[22:10] <genii-around> Yes.
[22:10] <subman> On / I have 36 G of which 22 G is free
[22:11] <genii-around> subman: What about /var ?
[22:13] <subman> genii-around, /var is on /
[22:13] <genii-around> subman: Does df -h show any partitions which are getting filled up?
[22:13] <sarnold> subman: how about ls -ld / /var /var/www /var/www/index.php   ?
[22:16] <subman> sarnold, http://pastebin.com/RQbTFqEb
[22:17] <sarnold> subman: ls says that file doesn't exist...
[22:17] <subman> The file is there
[22:17] <sarnold> at least permissions above it look sane enough.
[22:17] <sarnold> don't argue with me, argue with ls -l :)
[22:17] <subman> :)
[22:19] <subman> sarnold, oops
[22:19] <sarnold> subman: the + does mean that there are extended access controls on the directory; perhaps your user account and the apache user account have been forbidden to read or execute that directory?
[22:20] <subman> sarnold, http://pastebin.com/xZyfJqS4
[22:21] <sarnold> subman: okay; how about the directory access control lists? I _think_ that's something like getfacl /var/www -- but it's been ages since I've done that.
[22:22] <subman> sarnold, why in the world would any of this have changed?
[22:23] <sarnold> subman: half the time people notice unexpected changes, it's because a buddy was doing things and forgot to tell people / forgot to finish. a large part of what's left is folks noticing they've been in-elegantly hacked... the remainder are harder to explain. :)
[22:24] <subman> I'm wondering if hacked
[22:30] <subman> I guess del and reinstall?
[22:31] <ScottK> It's unlikely you're hacked.
[22:31] <ScottK> Figuring it out could be a good learning experience.
[22:46] <urthmover> When I add a ppa for postgres....how do I know that running apt-get install postgesql-9.1  is actually coming from the newly added ppa?
[22:47] <sarnold> urthmover: 'apt-cache showpkg postgresql-9.1' or 'apt-cache showsrc postgresql-9.1' will probably output reassuring information.
[22:48] <urthmover> well said sir.   I just man apt-cache  and found the apt-cache showsrc one   thanks for the quick answer
[22:51] <sarnold> urthmover: there may be another mechanism, via 'pinning', to ensure you always get from the ppa version, but apt was more designed for distributions than for ppas.. check out the apt_preferences(5) manpage's section on "APT's Default Priority Assignments" and see if that looks amenable.
[22:52] <sarnold> urthmover: "Pin: origin" may be useful.
[22:54] <subman> Well I'm stumped here and have no idea where to turn.  Any ideas for me to look?
[23:01] <urthmover> way cool sarnold
[23:02] <sarnold> subman: when all else fails you can attach strace to your apache and try to catch it in the act of failing; but that's a miserable way to spend an afternoon. You're sure there's nothing else instructive in the logs?
[23:03] <subman> Nothing that I can find
[23:12] <lvmer> I could use some minor help with samba. When I set one of my folders to $ security = user  && valid users = @group1 greg john     ..... I can't log into the file... with either greg john or anyone in group1.
[23:12] <subman> ok, now I am getting closer here!
[23:13] <subman> In my browser I type http://site/index.html but firefox returns /index.php was not found.
[23:14] <subman> It is looking for .php
[23:17] <sarnold> subman: do you have any horrible rewrite games going on to re-write .html requests into .php requests?
[23:17] <subman> sarnold, I'm looking into that right now
[23:21] <subman> It seems it is my .htaccess is destroyed.
[23:23] <subman> Can I just delete .htaccess?
[23:35] <sarnold> subman: depends what it is supposed to do. :) most sites don't use .htaccess files since they're read and re-read on every incoming request, compared with the standard apache config which is read once at startup
[23:36] <subman> sarnold, thanks, just reading up now.  Thanks for your help.
[23:41] <phillw> hi guys, a very quick one for you. Based on the quote "I love standards, anyone can create one", which standard for Cloud servers is becoming more prevalent? I ask not for your 'favourite', but what you good people see in the fight between "BetaMax & VHS", etc. ?
[23:41] <SpamapS> phillw: there's no "standard" for that.. just "providers" like EC2, Rackspace, HP, etc.
[23:43] <phillw> SpamapS: is OpenStack making any inroads or should I base my complete newness on a different standard?
[23:45] <SpamapS> phillw: There are currently 3 public openstack clouds you can get access to right now (HP, Rackspace, and Dreamhost, though that one is still limited I think)
[23:45] <SpamapS> phillw: there are also millions of dollars flowing into the OpenStack foundation, so yeah, I'd say its making inroads ;)
[23:47] <phillw> so, just like I had to make a leap over to gain 'Red Hat' certification for to be considered 'approved' in the commercial market, is using OpenStack similarly accepted?