[00:09] <rschmitty> sarnold: success! (went for the easy route) installed bsd-mailx first and I got the text normally.  It appears that when I have no mailx installed the apticron package selects the heriloom.  if you want to provide your answer on askubuntu i will accept :)  thanks again for your help!
[00:11] <sarnold> rschmitty: woo!
[00:15] <sarnold> rschmitty: okay, posted. thanks for the report back :)
[00:23] <rschmitty> sarnold: accepted :)  ...apparently rep is not shared between stackoverflow sites, I dont have enough on askubuntu to upvote your answer, but anywho, thanks again! time for food
[00:32] <GH0> sarnold, seems the issue is VNC. :/
[00:34] <GH0> Which is unfortunate, because I would think a shared VM running off a server wouldn't care about the window it is running in. Especially if it is going to be thrown into the background.
[00:54] <sarnold> GH0: yeah, 3d over vnc is schockingly poor
[00:54] <sarnold> it feels like the polite thing to do would just send a screen diff every RTT or something
[00:55] <GH0> Well, I e-mail RealVNC (don't hurt me for using proprietary stuff). Because honestly, there is no better way to achieve this. I am not running ESX which would be nice, but, I didn't install it that way at the time, and I can't really retrofit that now.
[00:56] <GH0> While 3D support exists over VNC in Windows, it doesn't seem to do that in Linux.
[00:56] <GH0> They just need to add more OpenGL extensions.
[01:07] <Patrickdk> heh, 3d support, I hardly even use 2d
[01:08] <Patrickdk> there is no need to use more than vt100 :)
[01:08] <sarnold> vt220 added something nice.. what was it..
[01:08] <Patrickdk> I normally get colors and 132char support with vt320
[01:08] <Patrickdk> not needed though
[01:16] <GH0> Oh, I do have another issue, why the hell would this happen when all of my devices are SATA II / Sata 3.0 Gbps capable? [   35.443984] ata3: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
[01:16] <GH0> Is there any way to figure out what is causing it to go to 1.5 Gbps
[01:21] <Patrickdk> heh?
[01:22] <Patrickdk> since when does, sata 3.0gbps capable == will be 3.0gbps
[01:22] <Patrickdk> ask the drive?
[01:22] <Patrickdk> bad cable, jumber on drive or card, sata card only supports 1.5gbps, ....
[01:23] <GH0> Doesn't mean it is 3.0 but the drives are all SATA II enabled, so I don't see why there is the issue of it jumping down in the first place.
[01:23] <Patrickdk> "I don't see why there is an issue" always results in, I didn't bother to look
[01:23] <Patrickdk> did you check any of the items I listed?
[01:26] <GH0> I checked the items listed around it when it occurs, but it is more foreign to me.  http://pastebin.com/TsVXcfpT  It shows up twice, once around the 2 second mark, and again in the 35 second mark.  With the one at the 35 second mark, I have no idea what any of that means. So no, I didn't look it up.
[01:29] <Patrickdk> so you didn't check anything I listed above?
[01:29] <Patrickdk> you only looked at dmesg output?
[01:30] <GH0> There is no jumper on either end of the device. Not entirely sure what cable to pull, but I just replaced everyone of them, the sata card supports 3.0
[01:30] <GH0> I only noticed this on my last boot up, and before I tear the system apart, I just wanted to make sure it wasn't something that would have been caused by software, or if there was a way to narrow down the issue prior to tearing the system apart.
[01:31] <Patrickdk> your cdrom supports 3gbps?
[01:32] <Patrickdk> seriously, there is no way a dvd drive is going get >150MB/sec
[01:33] <Patrickdk> http://www.samsung.com/ph/consumer/monitor-peripherals-printer/optical-disc-drive/dvd-rw-drive/SH-S223C/RSBF-spec
[01:33] <Patrickdk> show me on that page, where it says supports 3.0GBPS sata2
[05:06] <Babin> Hi all , I have installed Ubuntu server 12.04 and now i need to setup apt- chache-ng server how can i do it , help me for step by step procedures
[05:09] <Babin> I have installed Ubuntu server 12.04 and now i need to setup apt- chache-ng server how can i do it , help me for step by step procedures+
[08:11] <smadamr> hello, can anyone provide some info on best solving a full hdd issue? Is it best to add another HDD (if slots available) or to add network storage? Are there any resources I can learn about this?
[08:23] <TJ-> Is there a dedicated channel for XCP XAPI package issues?
[08:32] <TJ-> Any idea where the XCP-XAPI plugins (extauth, etc.) have got to - they seem to be missing from xcp-xapi?
[09:35] <zetheroo> How does one install Ubuntu Server 12.04 on ZFS?
[09:39] <acidflash> I don think that makes sense.
[09:41] <acidflash> zetheroo: ZFS is a file system, Ubuntu server is an operating system. you "cant" install Ubuntu server on "ZFS", you can use ZFS as an FS on Ubuntu server "after" you install it. Unless you mean format the disk using ZFS during the installation process, which I dont think is currently possible, since ZFS is not "officially" supported on linux. There is a spin-off, but not officially supported.
[09:42] <maswan> zetheroo: you add the https://launchpad.net/~zfs-native/+archive/stable ppa to install ZFS on Ubuntu
[09:43] <acidflash> maswan: you mis-understood his question, he wants to install the OS on ZFS, ie: during ubuntu server setup, have ZFS available.
[09:43] <zetheroo> so if I install Ubuntu Server on an ext4 partition and then later install the ZFS packages ... do I then convert my ext4 partitions to ZFS?
[09:43] <maswan> zetheroo: no, you don't
[09:43] <maswan> zetheroo: if you want to run zfs on the root filesystem you need to do some additional steps
[09:44] <zetheroo> so there is no way to get Ubuntu Server to 'live' on a ZFS filesystem?
[09:44] <zetheroo> I see
[09:44] <maswan> zetheroo: see https://github.com/rlaager/zfs/wiki/HOWTO-Install-Ubuntu-to-a-Native-ZFS-Root-Filesystem for instance
[09:46] <maswan> zetheroo: Mostly, I'd recommend staying with known stable and supported ext4/xfs for the OS and then use ZFS for data disks, but if you want to it is possible to do ZFS root too. Just not easy, and it might be prone to breakage.
[09:48] <yolanda> jamespage, zul: https://code.launchpad.net/~yolanda.robla/ubuntu/saucy/ipxe/ftbfs/+merge/171754
[09:48] <yolanda> applied the upstream patch
[09:49] <zetheroo> is there a way to install Ubuntu Server to a USB Flash drive so that it always lives there and boots from there?
[09:51] <maswan> zetheroo: shouldn't be that hard, but it will be very slow
[10:00] <sarthor> HI, ubuntu server 13.04 empty /etc/udev/rules.d/70-persistent-net.rules also one lan card is showing "p7p1" other LAN  is not listed
[10:00] <funkyHat> zetheroo: plug in the USB flash drive, run the installer from a CD or a different flash drive, select the flash drive as your /
[10:00] <sarthor_> f
[10:00] <sarthor_> f
[10:00] <sarthor_> HI, ubuntu server 13.04 empty /etc/udev/rules.d/70-persistent-net.rules also one lan card is showing "p7p1" other LAN  is not listed
[10:00] <sarthor> Sorry double pasted.
[10:02] <funkyHat> Is this because of the new consistent (as opposed to persistent) device naming thing in udev? I can't remember whether that was enabled by default on ubuntu server yet...
[10:02] <rbasak> biosdevname
[10:02] <funkyHat> I assume that the other card not appearing is a separate issue
[10:02] <funkyHat> Yes that
[10:04] <sarthor> now how to solve this problem. I can not configure my LAN cards,
[10:09] <TJ-> sarthor: Check for the missing Ethernet adapter, e.g. "lspci -nn" then check if a driver is loaded for it. E.g. if the Ethernet adapter is at 06:07.0 then "lspci -nnvvv -s 06:07.0" and look at "Kernel (driver in use|modules)"
[10:20] <sarthor> TJ-, both are 100% lan cards
[10:22] <sarthor> TJ-, and I can see both LAN cards in lspci
[10:22] <yolanda> jamespage, zul, about the server banners for postgresql, the check for version should be done in the client or in the server?
[10:22] <yolanda> a telnet in postgres doesn't even show the version
[10:23] <yolanda> and a SELECT VERSION() already shows Ubuntu/Linaro
[10:30] <mikal> Do the packages for nova still use sqlite by default?
[11:07] <yolanda> what was the url for checking the launchpad errors for branches? i found that openssh branch is outdated
[11:14] <caribou> I have a Maas semantic question :
[11:15] <caribou> the settings pages states(around the proxy field) :
[11:15] <caribou> "This is used by the cluster and region controllers for downloading PXE boot images and other provisioning-related resources. It is not passed into provisioned nodes."
[11:15] <maxb> yolanda: Sounds like you want http://package-import.ubuntu.com/
[11:16] <caribou> The statement "It is not passed into the provisioned nodes" since the preseed does have the proxy defined
[11:16] <yolanda> maxb, thx
[11:16] <caribou> The statement is *false* is what I meant
[11:17] <rbasak> caribou: is it the same setting?
[11:17] <rbasak> (that gets passed through)
[11:18] <caribou> rbasak: yeah, If I define a proxy in the field, it makes it into the preseed of the node
[11:18] <rbasak> caribou: into apt's configuration, or just the installer? Sounds like the documentation's wrong to me, too, then.
[11:19] <caribou> rbasak: well, I need to comission one node to find out, but it's clearly listed in the node's preseed file :
[11:19] <rbasak> IMHO, the basic implementation of a proxy setting should use that proxy for *everything* where possible. I think that using the proxy only for selected things is a much rarer use case.
[11:19] <caribou> rbasak: d-i     mirror/http/proxy string http://192.168.124.3:8000/
[11:20] <caribou> rbasak: hmm, hold on, this is not what is defined in the proxy field
[11:20] <yolanda> found this problem for openssh: http://package-import.ubuntu.com/status/openssh.html#2012-07-04 12:48:07.322841
[11:21] <yolanda> bzrlib.plugins.builddeb.errors.PackageVersionNotPresent: openssh 5.8p1 was not found in <PristineTarSource at file:///srv/package-import.canonical.com/new/updates/openssh/oneiric/>.
[11:21] <yolanda> how can this be solved?
[11:21] <caribou> rbasak: my proxy definition is 192.186.124.1:9999/
[11:21] <caribou> rbasak: let me add another node w/o the proxy field filled up
[11:44] <caribou> rbasak: I get it into the preseed of the new node I just created even if the field is empty
[11:45] <caribou> rbasak: this might come from the pkg configuration; this proxy URL doesn't look right to me
[11:50] <caribou> rbasak: I must step out, will do more tests when I return
[12:15] <Jeeves_> Hi!
[12:15] <Jeeves_> Q: I've got an 'sync'-option in /etc/fstab for my nfs-mount. But it is ignored
[13:10] <zul> jamespage:  so neutron/quantum how do you think we should handle it?
[13:28] <yossarianuk> what is the best way of doing SSD + RAID + TRIM at present ?
[13:28] <yossarianuk> for ubuntu 12.04.
[13:30] <xnox> yossarianuk: do not use TRIM with RAID.
[13:30] <xnox> as it can lead to dataloss.
[13:31] <xnox> yossarianuk: plus it's fully available only in latest mdadm releases, not sure if it all is available in 12.04.
[13:31] <yossarianuk> xnox: does that include HWRAID as well as MDRAID ?
[13:32] <yossarianuk> xnox: its not
[13:32] <xnox> yossarianuk: hardware / fakeraid should be fine.
[13:33] <xnox> .... but then again check trim support on the hardware/fake raid solution you are choosing, it might not support trim.
[13:34] <yossarianuk> xnox: the HW doesn't - HP smart array - is RAID still o.k without TRIM.
[13:34] <xnox> yossarianuk: yes.
[13:35] <xnox> without TRIM all/any are fine. and you still get IO performance boost from SSD.
[13:35] <yossarianuk> xnox: thanks for the info !
[13:53] <resno> ive got a question about using rsync to backup. which is better to iniate from backup server calling to server or opposite?
[13:59] <TJ-> resno: Instinct says from the target, but whichever way it has to move data about the files found over the net
[14:03] <resno> TJ-: ya, a collegue opened up root ssh access and is doing rsyncs from remote to server... and i just wanted another opinion before changing it
[14:03] <TJ-> resno: Would make sense from organisational point-of-view, to have it all run from the back-up server though
[14:03] <TJ-> resno: One place for scripts, cron jobs, etc.
[14:04] <resno> but then having permitting root login access via ssh?
[14:04] <genii> If you have multiple machines you backup to a central point, having it done centrally there makes more sense
[14:06] <TJ-> resno: I wouldn't allow root access, I'd create a backup account with sufficient privileges (maybe part of the mythical 'wheel' group)
[14:06] <TJ-> resno: That way, you can restrict the back-up user in what they can do
[14:07] <resno> ok, thats a more resonable idea
[14:07] <genii> Alternately add specific commands for only that account in sudoers
[14:07] <genii> ( so they can run rsync, etc)
[14:08] <GH0> Is there any way I can verify that the DAMAGE extension is running on X?
[14:09] <resno> ok thanks TJ- genii
[14:09] <TJ-> You really should be using the rsync daemon though, which avoids the whole root issue entirely. See https://help.ubuntu.com/community/rsync#Rsync_Daemon
[14:10] <resno> heh, rsync daemon
[14:18] <yossarianuk>  I take it If I am not using TRIM I do not use the 'discard' mount option ?
[14:20] <patdk-wk_> yep
[14:21] <yossarianuk> patdk-wk_: if you were responding to me thanks !
[14:22] <patdk-wk_> well, you could use discard, but it would do nothing
[14:22] <patdk-wk_> discard enables TRIM on ata/sata, and UNMAP on scsi/sas
[14:22] <patdk-wk_> if your stuff doesn't support it, it will just ignore it
[14:23] <ikonia> yossarianuk: haven't we had this discussion already ?
[14:23] <patdk-wk_> ikonia, not in the last hour :)
[14:23] <ikonia> patdk-wk_: I think we did, in #ubuntu, about 25 minutes ago
[14:24] <ikonia> hence my surprise that the same thing is being repeated again with the same person
[14:24] <yossarianuk>  ikonia: I was just checking if I needed the mount option when not using trim
[14:24] <ikonia> what mount option ?
[14:24] <ikonia> 14:28 < yossarianuk> what is the best way of doing SSD + RAID + TRIM at present
[14:24] <yossarianuk> previous conversations made me realise to not use TRIM
[14:24] <patdk-wk_> :)
[14:24] <ikonia> and that's not "a mount option" thats the exact word for word question you asked in #ubuntu
[14:24] <yossarianuk> mount option 'discard'
[14:24] <patdk-wk_> actually it was in this channel too 50min ago
[14:25] <ikonia> 14:28 < yossarianuk> what is the best way of doing SSD + RAID + TRIM at present? For Ubuntu 12.04?
[14:25] <resno> ya, i just saw it before my quesition
[14:25] <ikonia> so that's the same question just being asked over and over again despite getting told the answer each time
[14:25] <patdk-wk_> needed a 4th opinion
[14:26] <resno> heh
[14:26] <ikonia> I don't massivly care, I'm just surprised that it's not clear
[14:26] <yossarianuk> no - i was confirming if I should use the 'dicard' option (fstab) when not using trim - due to earlier conversations I now realise with my HW I should not use TRIM.
[14:26] <ikonia> or wonderig if there is something that's not being explained well
[14:26] <yossarianuk> no its been explained fine.
[14:27] <patdk-wk_> using hardware raid, the trim will be ignored, and it shouldn't even be using trim, but unmap
[14:27] <yossarianuk> got it thanks
[14:27] <yossarianuk> and thanks for all people who have given advice..
[14:27] <yossarianuk> RAID+SSD seems a bit catch22 at present.
[14:28] <yossarianuk> (which is the best book of all time)
[15:22] <FunnyLookinHat> I'm testing the ubuntu-cloud images in my own Devstack environment, and they take a long time to boot because they're trying to mount /tmp and whatnot ... posting errors like "The disk drive for /tmp is not ready yet or not present. Continue to wait, or Press S to skip mounting or M for manual recovery" to the log.
[15:22] <FunnyLookinHat> Any idea on how to avoid that ?
[16:04] <urthmover> I'm struggling trying to get vhosts to work with a default apache2 installation on a 12.04 server.  Does anyone have some spare time to help me debug this?
[16:05] <urthmover> root@prod2:/etc/apache2/sites-enabled# apache2ctl -S
[16:05] <urthmover> apache2: Could not reliably determine the server's fully qualified domain name, using 10.182.12.34 for ServerName
[16:05] <urthmover> VirtualHost configuration:
[16:05] <urthmover> wildcard NameVirtualHosts and _default_ servers:
[16:05] <urthmover> *:80                   is a NameVirtualHost default server iceadm.sparklingice.jpdmi.com (/etc/apache2/sites-enabled/iceadm.sparklingice.jpdmi.com.conf:1) port 80 namevhost iceadm.sparklingice.jpdmi.com (/etc/apache2/sites-enabled/iceadm.sparklingice.jpdmi.com.conf:1) port 80 namevhost prod.sparklingice.jpdmi.com (/etc/apache2/sites-enabled/prod.sparklingice.jpdmi.com.conf:1)
[16:05] <urthmover>  port 8 namevhost 10.1821.12.34 (/etc/apache2/sites-enabled/z000-default:1)
[16:05] <urthmover> Syntax OK
[16:06] <urthmover> yet I only see the first site no matter which URL I go to
[16:10] <shauno> urthmover: "port 8 namevhost 10.1821.12.34" is that a typo or straight from the file?  both that port and that IP look unrealistic
[16:11] <urthmover> must have been a pste messup
[16:12] <patdk-wk_> straight from the file, the dns name 10.1821.12.34
[16:12] <patdk-wk_> note it says namevhost, not ip
[16:14] <FunnyLookinHat> urthmover, probably a question for #httpd  - but have you setup a ServerAlias for each virtual host ?
[16:14] <FunnyLookinHat> You also don't need a namevirtualhosts config iirc unless you're doing some funky proxying
[16:14] <FunnyLookinHat> Lastly - make sure that you've enabled all of your virtual hosts ( ls /etc/apache2/sites-enabled/  should show all of the ones you want live )
[16:16] <urthmover> FunnyLookinHat: You're right I'm in there now actually thanks guys
[16:17] <FunnyLookinHat> urthmover, good luck :D
[16:17] <urthmover> FunnyLookinHat: thanks
[17:00] <thegoat> is there a flag i can turn on in my init script to see where it's bombing out?
[17:59] <shankstaBytes> i can connect to my server via ftp and i also can from an external ip address, but people from india can't
[17:59] <shankstaBytes> any idea why
[17:59] <shankstaBytes> im in the USA
[18:00] <sarnold> shankstaBytes: FTP is a horrible protocol. If the FTP server is behind one NAT firewall and a client tries to connect from behind another NAT firewall, there's a good chance it'll never work.
[18:02] <shankstaBytes> sarnold: but it works from me on a remote ip
[18:02] <shankstaBytes> sarnold: or you mean they are behind NAT?
[18:03] <sarnold> shankstaBytes: yeah, if your indian users are behind a NAT firewall..
[18:03] <sarnold> .. or even a poor regular firewall.
[18:51] <shankstaBytes> sarnold: odd because i had 2 indian users say they can't reach my ftp
[18:52] <shankstaBytes> sarnold: are you in the US?
[18:52] <sarnold> shankstaBytes: yes
[18:52] <shankstaBytes> just curious maybe you could test it
[19:01] <shankstaBytes> sarnold: they can connect to my database and apache just not ftp
[19:01] <shankstaBytes> should they use a VPN or something?
[19:01] <sarnold> shankstaBytes: a VPN could improve it, yes, but if you could just switch to sftp instead, that'd probably also solve it
[19:02] <shankstaBytes> sarnold: im using vsftp
[19:03] <shankstaBytes> because it allows me to chroot easily
[19:03] <shankstaBytes> i dont want the person to be able to execute any commands
[19:03] <shankstaBytes> just access files
[20:11] <TimRe> can anybody tell me why on 12.04.2 the dns numbers keep getting erased for after an hard shutdown of the server?
[20:14] <Pupeno> Any ideas how to make jetty see the installed slf4j? I'm getting this error when starting jetty, the package is installed: 501 [main] WARN org.mortbay.log - failed SolrRequestFilter: org.apache.solr.common.SolrException: Could not find necessary SLF4j logging jars. If using Jetty, the SLF4j logging jars need to go in the jetty lib/ext directory. For other containers, the corresponding directory should
[20:14] <Pupeno> be used. For more information, see: http://wiki.apache.org/solr/SolrLogging
[20:29] <superboot> Any hints on how to install virtualbox on a headless/X11-less server, or what channel would be best suited to get answers? I've added the Orical Virutalbox repo.
[20:29] <TimRe> anybody know the answer to my question
[20:30] <superboot> TimRe: I came in too late to see it. Post again if you want me to take a stab at it.
[20:30] <TimRe> can anybody tell me why on 12.04.2 the dns numbers keep getting erased for after an hard shutdown of the server?
[20:32] <superboot> TimRe: If you are talking about the entry in resolv.conf, I may have an answer.
[20:32] <TimRe> yes in the entry in resolv.conf
[20:34] <superboot> Ok, if I remember correctly it has to do with it being dynamicly set by "resolvconf". Try man resolvconf (I am too). If memory serves, you need to edit a file somewhere in /var/run...
[20:35] <TimRe> ok i did that
[20:35] <TimRe> what am I looking for in that file
[20:36] <superboot> TimRe: In the man file? You are looking for how to specify changes that will survive a reboot. Also I'll google too.
[20:36] <TimRe> ok
[20:39] <superboot> Ok. The file is /etc/resolvconf/resolv.conf.d/base (or head, or tail, depending on where you want to put it.)
[20:40] <TimRe> I seen that part but dont know what else to do with it
[20:40] <superboot> Entries in there will be like you put them in /etc/resolv.conf on a system without resolvconf(8).
[20:42] <superboot> TimRe: I was assuming you knew what to do with /etc/resolv.conf, since you were trying to edit it in the first place. If that was an invalid usumption, just add "nameserver 192.168.0.1" (or whatever your dns entry is, usualy your router's address (if it handles dns that is)).
[20:43] <superboot> So, anyone have an answer to MY question? ;)
[20:43]  * superboot chuckles
[20:43] <TimRe> well see everytime I do the /etc/resolv.conf and add dns nameservers in there save it then when I do either hard or soft shutdown it will erase it
[20:48] <superboot> Right. So you have to put it in the /etc/resolvconf/resolv.conf.d/base file... did you read what I wrote?
[20:50] <superboot> Also there is a nice hint at the top of the generated resolv.conf file stating to not edit the file as the changes will be removed. Unless of course I'm missing something (e.g. you don't have resolvconf and somehow are not realy writing the changes to the file. I doubt this)
[20:57] <superboot> TimRe: Any luck?
[21:36] <zwirc> Hi!
[21:37] <zwirc> I added two extra IP`s to my dedicated server, but when il add them to ubuntu server users cant reach them
[21:37] <zwirc> what may that be?
[21:37] <zwirc> some people can reach them , but not all
[21:39] <Patrickdk> zwirc, you have traceroutes showing this? from both ends?
[21:40] <Patrickdk> without that the only thing we can do to check
[21:40] <Patrickdk> is if you routing table, iptables, and ip addr are setup right
[21:42] <zwirc> hmm
[21:42] <zwirc> Only a few people can reach the server on one of the ips
[21:43] <zwirc> but the other ip, all can reach it
[21:43] <Patrickdk> you said that already
[21:44] <thumper> hallyn: you up?
[21:44] <thumper> hallyn: wondering about a good resource to read up on lxc-net
[21:45] <zwirc> but strange that its only for a few users
[21:45] <zwirc> iface eth0:1 inet static
[21:45] <zwirc>         address xxx.xxx.xxx.xxx
[21:45] <zwirc> thats the right way yes?
[21:45] <zwirc> and auto eth0:1
[21:48] <Patrickdk> zwirc, insufficient info
[21:48] <Patrickdk> I told you what was required
[21:49] <hallyn> thumper: uh, mainly just https://help.ubuntu.com/13.04/serverguide/lxc.html
[21:49] <thumper> hallyn: ok, ta
[21:50] <thumper> hallyn: I read that in the config file for lxc we can limit max memory, can you limit the number of cores?
[21:50] <hallyn> thumper: as of saucy it's gotten just a bit funkier to support nested networking, but the changes are mainly in how hte package is shipped
[21:50] <thumper> hallyn: still focusing on precise
[21:50] <thumper> for juju
[21:50] <hallyn> thumper: you can use cpusets to pin the container to certain cpus,
[21:50] <hallyn> but you can't say "just use 2 cpus"
[21:50] <hallyn> you've got to pick them
[21:50]  * thumper nods
[21:50] <thumper> I think just limiting memory might be a good enough start for us
[21:51] <Patrickdk> isn't that going to cause numa issues?
[21:51] <thumper> right now I'm just using the default values
[21:51] <sarnold> more likely to help numa issues.. :)
[21:51] <thumper> which I'm assuming means "access to all?"
[21:51] <hallyn> what sarnold said
[21:51] <hallyn> thumper: yes
[21:51] <Patrickdk> sarnold, depending on how memory is allocated
[21:52] <sarnold> Patrickdk: always running on one of two CPUs is going to be better than running on any of N cpus..
[21:52] <sarnold> Patrickdk: especially since the tasks will be born, live, and die, entirely on those two
[21:52] <hallyn> i'm actually not sure how smart slub is - will it automatically pick memory closest to your cores?
[21:52] <hallyn> (assuming people actually still have numa... i thought numa was dead)
[21:52] <sarnold> hallyn: .. and yet everything is numa now, no? :)
[21:53] <Patrickdk> hallyn, everything for a long time has been numa
[21:53] <Patrickdk> basically all ddr3 systems are
[21:53] <Patrickdk> I don't know of a single multicpu system with ddr3 that isn't
[21:54]  * Patrickdk notes he should check what amd does, as I only really use intel
[21:54] <hallyn> but does the cpu just try to take care of you?  All I know is I'm on the libnuma list, and that list is dead :)
[21:54] <Patrickdk> no
[21:54] <Patrickdk> I know it's a big issue with mysql
[21:54] <hallyn> cool, i learned something.  i can end my day
[21:54] <Patrickdk> it will allocate memory from one side
[21:54] <sarnold> hallyn: haha :)
[21:54] <Patrickdk> till that memory is exaushsted then move to the other cpu ram
[21:54] <Patrickdk> but I dunno how it works for *smaller* allocations :)
[21:55] <hallyn> but yeah i'd guess that randomly bouncing around 2 of 8 cpus isn't going to help that.  if you're pinned to 2, you can presumably do some magic to get memory close to you?
[21:55] <Patrickdk> so mysql loaded up on a dual socket with 128gigs ram, will be numa bound with <64gigs allocated
[21:56] <Patrickdk> I was wondering, guess the real answer is, we don't know :)
[21:56] <Patrickdk> need research to answer it :)
[21:56] <Patrickdk> don't mean to totally derail :)
[21:58] <hallyn> yeah no derailment, i'm just shutting up until i do more research :)
[22:02] <JoeyJoeJo> When I installed my server I check off "Virtual machine host" at the end of the installation. That made a bridge interface but it also applies some iptables masquerading rules on boot that I don't want. How can I get rid of them?
[22:03] <sarnold> JoeyJoeJo: does /etc/default/lxc describe the bits you're looking to change?
[22:03] <JoeyJoeJo> No, I don't have that file
[22:08] <shodan45> I just ran a dist-upgrade on my 12.04.2 server & got a bunch of "error: found two disks with the index 1 for RAID md/0". Should I be worried?
[22:09] <shodan45> google only shows stuff about an old grub bug?
[22:14] <noobuntu> wondering what the " & " at the end of a terminal command does? ..i see it like ex. firefox & ... gedit &
[22:14] <sarnold> noobuntu: it runs the command "in the background"
[22:14] <noobuntu> cool..thanks soo much for the quick response
[22:15] <sarnold> noobuntu: shells with job control let you manipulate multiple jobs at once; most programs will suspend when you hit ^Z, and then you can either have them continue running in the background with "bg %1" or resume them in the foreground with "fg %1" -- where the job number can be found by typing "jobs"
[22:15] <sarnold> heh
[22:17] <hallyn> JoeyJoeJo: virsh net-autostart --disable default
[22:23] <JoeyJoeJo> hallyn: That worked, thanks
[22:25] <hallyn> np
[22:41] <tboat> hey all, so I recently attempted to migrate my server from one HDD to a larger one using clonezilla.  The transfer failed a few times, so I just rebuilt the server.  However, I still need to get my user files off of the old HDD.  When I connect the HDD via external USB, it only shows /boot partition.  How can i get to /home?
[22:42] <bobek_balinek> Hi, my ubuntu server 12.04 is ramming 100% CPU on process 'tty1' for good hours, strace just logs:" trace(0, "") = 0 " in an infinite loop. Killing the process just stops it for couple of hours, then its back on full CPU again.
[22:50] <sarnold> tboat: check dmesg to find out which partitions the kernel knows about; then use mount /dev/sd<whatever> /mnt/some_directory to mount the other partitions, and find your data..
[22:50] <sarnold> bobek_balinek: that's odd. I've never heard of those process names before, and man 2 trace shows nothing.
[22:51] <sarnold> bobek_balinek: what does the process hierarchy look like? what is its ppid, and what is _that_ process?
[22:51] <sarnold> bobek_balinek: does /proc/pid/exe or other /proc/pid/* file show you anything?
[22:53] <bobek_balinek> sarnold: so what I'm getting in htop is: /sbin/getty -8 38400 tty1, i killed the process so right now its using no cpu. how do i execute /proc/pid ?
[22:53] <sarnold> bobek_balinek: ahhhhhhhh
[22:54] <tboat> sarnold: when i attempt to mount the partition, it claims i need to add the disk type (ext2, ext3, etc) which i do with -t extx, yet it claims it can't mount
[22:54] <sarnold> bobek_balinek: /proc/pid is just a short-hand; the kernel exposes information about processes in the /proc/ directory. check out ls -l /proc/self/  for a very-quick view of what I mean..
[22:54] <bobek_balinek> sarnold: if that helps: here's a link to someone who's had that problem but its still unsolved http://askubuntu.com/questions/174710/sbin-getty-process-causing-100-cpu-utilization
[22:54] <sarnold> tboat: are there any additional errors in dmesg?
[22:56] <tboat> sarnold: wrong fs type, bad option, bad superblock
[22:57] <tboat> tboat: just got it sorted though, files I want are on an LVM partition, so just needed to look up a few extra steps to mount the LVM, thank you though!
[22:57] <sarnold> bobek_balinek: _crazy_. I suggest filing a bug; ubuntu-bug can take a pid, that's probably the best approach.
[22:57] <sarnold> tboat: oh! woohoo. :)
[22:58] <tboat> now i just need to save up for a backup server for my main server lol, and never deal with this again.
[22:58] <bobek_balinek> sarnold: cool, thanks man :)
[22:59] <sarnold> tboat: good plan. :)
[22:59] <tboat> so hypothetically if i wanted to add a user with scp access to their own partition with their home on it, could that be done?
[23:07] <sarnold> tboat: sure, adduser can set a home directory somewhere unusual quite easily
[23:07] <tboat> cool, then just a matter of making the partition mount on boot
[23:07] <tboat> what file type should i use for that partition?
[23:07] <sarnold> tboat: right, that's /etc/fstab
[23:08] <tboat> or would i add this through LVM?
[23:10] <TimRe> superboot so I have to put the nameserver address into /etc/resolvconf/resolv.conf.d
[23:10] <TimRe> bc that file is blank
[23:56] <banzounet> Hey guys, where is the log file of apache ? cause when I try to reload the config (service apache2 reload) it fails, and I don't know where to look