[00:14] <arosales> have a good weekend.
[00:23] <mgw> how could i create a 'file' whose content will always be the output of some cmd
[00:23] <mgw> ?
[00:24] <mgw> without using cron
[00:24] <mdeslaur> mgw: start a daemon that updates the file?
[00:25] <mgw> mdeslaur: I suppose…. any other ideas?
[00:26] <mdeslaur> a dbus service?
[00:26] <mdeslaur> I lack context
[00:26] <mgw> I have a directory
[00:26] <mgw> of files
[00:26] <mgw> I need a yaml file that will be always up to date
[00:26] <mgw> that lists the files in that directory
[00:28] <mdeslaur> hrm, unless I'm missing something, a cron job would be a good way or else it gets really complicated :)
[00:31] <mdeslaur> a daemon that has an inotify on the directory and that you communicate with using dbus or some other way?
[00:31] <mgw> mdeslaur: yeah, i was looking at inotify
[00:31] <mgw> I was hoping there was some fundamental file type I was missing
[00:31] <mgw> I'm just going to execute a script to do it
[00:33] <mdeslaur> mgw: named pipes?
[00:33] <mdeslaur> you still need a daemon on the other end though
[00:37] <vindav> Grsync you can count on for local drive back up to the your network in either its entirety or in segments.  Now to find out if it can do it on a scheduled basis or not, is the question.  Anybody know about how to go about that ?
[00:41] <mgw> mdeslaur: I just discovered that puppet (for which this file is required) has a prerun script
[00:41] <mgw> so I can create the file there
[00:41] <mgw> thanks for the ideas!
[00:42] <mdeslaur> cool!
[00:42] <mdeslaur> mgw: np
[01:40] <hallyn> aha - apparently, systemd REALLY wants /dev/shm to be mounted, or it won't do a thing
[01:42] <hallyn> yay, i have a f16 container working
[01:42] <hallyn> stgraber: ^
[01:42] <hallyn> now what to do with it?
[01:42] <hallyn> rm it i guess
[01:42] <stgraber> wow!
[01:45] <hallyn> took me hours to figure that out :)  it jsut wouldn't spit out any info at all.  and i thought strace was just being silly, but in fact it was telling me the problem :)
[02:16] <ravenel> evening everyone
[02:16] <ravenel> i have a question about an mdadm-based raid array im hoping someone might be able to point me in the right direction on
[02:34] <pmatulis> !ask | ravenel
[03:01] <ravenel> I have an mdadm-based raid array, and recently drive sda in this array failed. I replaced the drive, but now cannot boot--I get a "reboot and select proper boot device" error
[03:02] <ravenel> Im thinking this is likely from my BIOS, complaining because the new, bare sda drive does not have anything, including an MBR, on it
[03:02] <ravenel> Any suggestions on how to resolve? Can I just change the BIOS boot order without trashing the now degraded array?
[03:19] <pmatulis> describe the array
[03:31] <fluvvell> ravenel, this sometimes happens if the MBR is not written to both drives in your RAID array. as pmatulis said, describe the array to us. is it RAID1 ?
[03:32] <ravenel> the array consists of three disks, sda, sdb, and sdc
[03:32] <fluvvell> should find the first boot drive by itself. Mirrored?
[03:32] <ravenel> there are three arrays--one raid-1 for boot, another raid-1 for /home, and a raid-5 data partition
[03:33] <ravenel> when i setup the array, i did install grub to the MBR of all of the drives
[03:33] <ravenel> sda is the drive that failed and was replaced with a currently bare drive
[03:33] <fluvvell> if you can boot a recovery cd, you might get grub to write the MBR to sda
[03:34] <ravenel> would it auto-detect the arrays?
[03:34] <ravenel> should i just manually change the boot order in my bios to boot from sdb? or will that trash the arrays?
[03:35] <fluvvell> depends what you boot with. Won't trash the array unless you tell it to. You might need to apt-get install mdadm and tell it to auto detect then rebuild. You *will* need to partition the sda drive
[03:36] <ravenel> correct, once i can get it to boot, i will copy the partitions from sdb or sdc and re-add the drive to the array
[03:37] <fluvvell> I would love to see some auto recovery tools to use with mdadm, but you have to do a fair bit by yourself
[03:37] <ravenel> yeah :(
[03:37] <ravenel> ok, so going into the bios and telling it to boot from sdb wont hurt anything?
[03:39] <fluvvell> bios can't boot a drive that has 1. no boot active partition 2. no MBR
[03:41] <ravenel> right, but it (should!) have both of those things... i set them up when i built the array... not sure why the BIOS didn't try to boot from the next disk when the first failed, but thats a problem for another day :)
[03:41] <fluvvell> did you use grub?  Did you tell it to make the drive bootable?
[03:41] <ravenel> yeah
[03:42] <fluvvell> ok so have you set up the partitions?
[03:42] <ravenel> no, not yet, sda is still bare, fresh out of the box
[03:44] <pmatulis> ravenel: booting from a degraded array, that's not a given.  it's a cautious step to *not* do that automatically
[03:44] <fluvvell> ok well you do have to set them up. There is IIRC a couple of ways of describing the arrays, one as sda sdb, and another as UUID.  The UUID method is more robust if you move the drives into different slots
[03:45] <ravenel> gotcha
[03:45] <ravenel> ok, so here is what i see my next steps as--feel free to correct!
[03:45] <ravenel> 1) update bios to boot from sdb
[03:45] <ravenel> 2) when booted, format sda-- sfdisk -d /dev/sdb | sfdisk /dev/sda
[03:46] <qman___> boot from sdb degraded, create identical partition layout, add it into the arrays, and install grub to its mbr
[03:46] <ravenel> 3) add partitions back to arrays
[03:46] <ravenel> 4) install grub
[03:46] <qman___> no formatting is needed
[03:47] <ravenel> gotcha, thanks!
[03:47] <ravenel> well, now to give it a try... and hope that i did in fact install grub to sdb and sdc! wish i'd taken better notes when i built this thing a few years ago...
[03:47] <fluvvell> I've previously booted from a CD, created the partitions, installed grub to the new drive then rebooted into the degraded array.
[03:48] <qman___> yeah, you can do it live too
[03:48] <ravenel> fluvvell: is that any less risky than doing it off the booted array?
[03:48] <qman___> not by any measurable amount
[03:48] <ravenel> gotcha
[03:48] <qman___> the only risk here is if sdb fails
[03:48] <ravenel> right
[03:48] <ravenel> then im hosed :)
[03:48] <qman___> and both cases take roughly the same amount of time to replicate the data
[03:48] <fluvvell> agreed
[03:49] <qman___> and require the same disk activity from sdb
[03:49] <ravenel> right
[03:50] <ravenel> if i had to do it again, i think id make it raid-6 to protect against a drive failing during rebuild... with these huge drives, its a concern
[03:50] <fluvvell> from a live boot cd you could always backup important data without starting the installed system.
[03:50] <ravenel> the important stuff is all backed up already, thank god
[03:50] <fluvvell> ok
[03:50] <ravenel> backs up every night to another server
[03:51] <fluvvell> ravenel, are you cooling your drives directly?
[03:52] <ravenel> yeah, the case has a fan blowing right over them
[03:53] <fluvvell> great.
[03:53] <ravenel> ok, im going to go give this a try
[03:53] <ravenel> thanks for the help!
[03:53] <ravenel> hopefully i wont be right back with problems :)
[03:54] <fluvvell> feel free to feedback your success.
[04:43] <ndroftheline> hello
[04:43] <ndroftheline> i'm a US Peace Corps Volunteer serving in a rural area in the Philippines and I'm looking for some help with planning out how to best leverage the available technology here
[04:44] <ndroftheline> the school I work at recently received a donation of 11 desktop computers with quad-core processors and 4gb ram each.
[04:44] <ndroftheline> i've been doing research about multi-headed linux and i like it, but the school is highly reluctatnt to abandon support for the MS Office family of programs
[04:45] <ndroftheline> so i was thinking about turning one of those powerful systems into a linux server and running a virtual windows server. then the multi-headed linux clients could connect to the virtual windows server to use MS Office apps.
[04:46] <ndroftheline> does that sound manageable?
[04:52] <pmatulis> ndroftheline: effectively, the clients would be connecting to a windows server, regardless if it's being virtualized withing a linux system
[04:53] <ndroftheline> that doesn't bother me, i'd be using the linux server as a file server as well
[04:53] <ndroftheline> er, should that bother me?
[04:54] <ndroftheline> i guess i could just run the windows OS directly on the server hardware
[04:54] <ndroftheline> and use windows as the fileserver as well
[04:54] <ndroftheline> which is better?
[04:55] <pmatulis> ndroftheline: it depends on the workload.  if the windows server is consuming the vast majority of the hardware resources then it would be best to just run windows natively
[04:55] <pmatulis> ndroftheline: if i were you i would first try to get rid of windows entirely
[04:57] <pmatulis> ndroftheline: you could also run windows applications locally on each of the clients
[04:58] <ndroftheline> true, but the clients will be running linux - windows doesn't offer a feasible multi-head setup
[04:58] <pmatulis> if you're stuck with MS
[04:59] <pmatulis> ndroftheline: no.  run Linux and the Wine framework.  you would need to check to make sure the MS s/w versions will run on the Wine you will be installing
[05:00] <ndroftheline> my only problem with using wine is that my personal experience and the research i've done online indicates that many slightly less popular ms office apps aren't very stable under wine
[05:00] <ndroftheline> for example, my school has curriculum teaching Publisher to the students
[05:01] <ndroftheline> so i'd rather have it running as natively as possible, since the damn suite is unstable anyway, even running in a native environment
[05:02] <pmatulis> ndroftheline: nothing much more i can help with then.  connect to the windows server via some kind of remote desktop session
[05:05] <ndroftheline> yep that's the plan. but do you think it'd be better to run the windows server directly and also use windows as the fileserver or run windows inside linux and use linux as the file server?
[05:08] <pmatulis> ndroftheline: i already answered that question
[05:09] <pmatulis> 23:55 <  pmatulis> ndroftheline: it depends on the workload.  if the windows server is consuming the vast majority of the hardware resources then it would be best to just run windows
[05:09] <pmatulis>                    natively
[05:09] <pmatulis> and, yes, it is midnight here
[05:10] <pmatulis> ndroftheline: so you would need to test stuff out to actually answer the question
[05:25] <ravenel> guys, just wanted to let you know that it looks like my issues with my raid array are solved--thanks fluvvell, qman___, and pmatulis
[05:25] <ravenel> forcing the BIOS to boot worked, its now rebuilding!
[05:26] <ravenel> off to bed!
[06:21] <uvirtbot`> New bug: #934763 in mysql-5.1 (universe) "package mysql-client-5.1 5.1.54-1ubuntu4 failed to install/upgrade: trying to overwrite '/usr/bin/mysqlaccess', which is also in package mysql-client 5.5.20-2" [Undecided,New] https://launchpad.net/bugs/934763
[07:08] <ZorroT> is there an FAQ entry for "how to get language X to function with apache2 cgi handler" ?
[07:18] <ZorroT> any apache2 ppl awake?
[08:05] <datalay> about orchestra: cobler check can you help please, http://www.pastebin.com/w5iwAmPV
[08:06] <datalay> how can i solve it
[08:58] <RoyK> hi all. how can I reset a VM's config to allow duplicating it? thinking about UUIDs, mac address, ssh keys etc
[09:36] <datalay> http://ubuntuforums.org/showthread.php?t=1927459 <-- this is my quesion
[10:11] <datalay> http://ubuntuforums.org/showthread.php?t=1927459 <-- this is my quesion
[10:13] <RoyK> guess what I asked about is http://libguestfs.org/virt-sysprep.1.html
[10:13] <RoyK> datalay: you just said so...
[10:29] <datalay> RoyK, i installed ubuntu orchestra with dns&dhcp support
[10:29] <datalay> but i couldn't pass some of settings: when i typed: cobbler check .. u can see details in the post
[10:37] <datalay> problem: One or more repos need to be processed by cobbler reposync for the first time before kickstarting against them:
[10:37] <datalay> i dont know how should i process them
[10:50] <RoyK>  
[12:31] <uvirtbot`> New bug: #935004 in mysql-5.1 (universe) "mysql-5.1 version 5.1.58-1ubuntu4 FTBFS on i386 in precise" [High,New] https://launchpad.net/bugs/935004
[13:13] <uvirtbot`> New bug: #935071 in drools (universe) "drools version 4.0.7-0ubuntu1 FTBFS on i386 in precise" [High,New] https://launchpad.net/bugs/935071
[14:01] <uvirtbot`> New bug: #935190 in rabbitmq-server (main) "package rabbitmq-server 2.3.1-1ubuntu1 failed to install/upgrade: subprocess installed post-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/935190
[16:09] <Tixos> hey, a few days back someone helped me with apache on ubuntu, being that response times were VERY slow. He fixed the issue for me by telling me to set KeepAlive to 1, which eradicated the issue. Now its back with vengeance ! Can anyone help me diagnose?
[16:12] <Tixos> here is a pastebin output from 'ab -c 10 -n 100 http://serverip/'  ran from the server
[16:12] <Tixos> http://pastebin.com/69DS1fCh
[16:13] <Tixos> usually not much RAM is used, but currently  >    4037420    3851604     185816           total/used/free
[16:15] <Tixos> also mysqld using upto 20% CPU, which is higher than normal
[16:21] <uvirtbot`> New bug: #935442 in sheepdog (universe) "sheepdog version 0.2.3-0ubuntu1 FTBFS on i386 in precise" [High,New] https://launchpad.net/bugs/935442
[16:26] <Tixos> brb
[16:51] <uvirtbot`> New bug: #935490 in wss4j (universe) "wss4j version 1.5.8+svntag-1 FTBFS on i386 in precise" [High,New] https://launchpad.net/bugs/935490
[16:56] <uvirtbot`> New bug: #935516 in genshi (main) "genshi version 0.6-2 FTBFS on i386 in precise" [High,New] https://launchpad.net/bugs/935516
[17:47] <RoyK> hi all. anyone that knows a good test site for webserver performance testing?
[17:52] <jamesd> RoyK, join #ubuntu, #linux and a few others larger channels... post a link to your site and later ask the people who looked at your site how the performance was... you will get far better feedback than you would from a single site testing your site...  one host is easy.. 500 visiters from around the world is much more real world
[17:56] <RoyK> jamesd: wrong way around... I'm looking for a test load, as in a set of content to put on a server, to benchmark a few things locally - I guess I could you wordpress or something for the php test, but then, it won't be optimal because of all the possible performance issues with wp
[17:57] <jamesd> RoyK, you could use wget to mirror a site, not sure how good it is at fixing all the links.
[17:59] <RoyK> probably a bad idea :P
[18:09] <yakster> can ne one give me an idea for a simple OS that is as light weight as possible, the only needed feature is wireless (intel wifi) and a full page web browser…
[18:10] <yakster> like run google chrome in a standlone system, with no window manager
[18:12] <qman___> sounds like you're looking for a chromebook
[18:12] <qman___> but you don't need X for a web browser
[18:16] <yakster> I want it to be a web interface to sickbeard, couch potato, SABnzb, and such for the media center
[18:21] <qman___> if you're just trying to save some performance, you should try lubuntu
[18:22] <qman___> building a bare minimum system from scratch is harder than it looks and usually not worth the effort
[18:23] <qman___> I run lubuntu on my T23s, and they can actually play 720p video at a visually acceptable rate
[18:24] <qman___> pentium 3s with S3 graphics decelerators, in case you didn't already know
[18:25] <ch33z> ah finally
[18:26] <ch33z> Hello I have a few major issues
[18:27] <ch33z> I am trying to set up VirtualHosts which I do have 2 sites and I have set up the VirtualHosts like it says to on Apache but, it still wont save it because it cant find a directory for it?
[18:27] <ch33z> any ideas?
[18:27] <ikonia> ch33z: have you read the https://help.ubuntu.com server section on how to setup apache
[18:27] <ch33z> indeed sir
[18:28] <ikonia> ch33z: apache on ubuntu is setup slightly different than generic apache configuration
[18:28] <ch33z> o.
[18:28] <ikonia> ch33z: this is documented in https://help.ubuntu.com server section.
[18:28] <RoyK> debian/ubuntu uses a slightly different config scheme
[18:28] <ch33z> well, ive read it through and through i am merely trying to set up a simple task just websites
[18:28] <ch33z> a basic html website
[18:28] <ikonia> ch33z: what part are you stuck with ?
[18:28] <ch33z> virtual hosts
[18:28] <ikonia> what about it ?
[18:29] <ch33z> ie mulitple domains
[18:29] <ch33z> how do i set them up?
[18:29] <ch33z> ;0
[18:29] <ikonia> it's documented in the site I've just given you I believe
[18:29] <ch33z> and when I log off it says "not a qualified domain name 127.0.0.1"
[18:29] <ikonia> ch33z: that is because you have not configured a hostname/ip address for it to listen on
[18:29] <ch33z> I have 9 static ips and 12 websites
[18:29] <ikonia> ch33z: again this is documented in the https://help.ubuntu.com guide which you have said you have read
[18:30] <ch33z> i see..
[18:30] <ch33z> thanks... for something i guess?
[18:30] <ikonia> ch33z: have you read that website ?
[18:30] <ch33z> yup
[18:30] <ch33z> back to front
[18:30] <ch33z> front to back
[18:31] <ikonia> ch33z: ok - so what part of configuratin apache virtual hosts are you stuck with
[18:31] <ikonia> ch33z: as that site walks you through it, I'm reading the pages now
[18:31] <ikonia> so I'm stuck at how you could have read this site back to front yet missed the core apache config details
[18:31] <ch33z> how exactly do i save  the virtual hosts?
[18:31] <ikonia> ch33z: what do you mean "save"
[18:31] <ch33z> o
[18:31] <ch33z> well when doing sudo nano the directory
[18:32] <ikonia> in what directory ?
[18:32] <ikonia> what are you actually doing ?
[18:32] <ch33z> it wont save it in the directory even though i am in it
[18:32] <ch33z> apache conf
[18:32] <RoyK> ikonia: probably reading porn :p
[18:32] <ikonia> ch33z: tell me exactly what file you are editing
[18:32] <ikonia> ch33z: and in what location
[18:32] <ch33z> "no such directory" even though i am writing in it and saving it
[18:32] <ch33z> kk
[18:32] <ikonia> ch33z: exactly what file/path are you editing eg: /usr/local/test.txt
[18:32] <ch33z> usr/local/apache2/conf
[18:33] <ikonia> ch33z: ok - so that file is not a valid ubuntu configuaration file
[18:33] <RoyK> ch33z: you didn't read the ubuntu docs....
[18:33] <ikonia> and now I know you have not read https://help.ubuntu.com and you are wasting my time
[18:33] <ch33z> o
[18:33] <ch33z> hld on
[18:33] <ikonia> ch33z: I refuse to progress help with people who lie to me, so good luck
[18:33] <ch33z> sorry for that master.
[18:33] <ch33z> *bows*
[18:34] <RoyK> ch33z: stop that - we're trying to help here, but you haven't done as you say
[18:34] <ch33z> I shall return then possibly my good sir
[18:34] <RoyK> ch33z: and lies don't lead to anything good in particular
[18:34] <RoyK> shuddup
[18:34] <pangolin> ch33z: when asking for help it is unnecessary and unhelpful if you pretend to understand more than you do.
[18:34] <ikonia> RoyK: that's uncalled for, please don't tell people to shut up
[18:35] <ch33z> thanks, I will read more up on it then
[18:35] <pangolin> there is nothing wrong with admitting that you are lost and don't know what to do
[18:35]  * RoyK wonder what it takes for that to be 'called for', then
[18:35] <pangolin> RoyK: in the Ubuntu community, nothing.
[18:35] <ch33z> well, I have read this stuff and still dont get it.
[18:35] <ikonia> RoyK: there is no situation where it is acceptable to tell a user to shutup,
[18:35] <ch33z> ill read more though
[18:38] <ch33z> so ikonia I tried sudo nano apache2.conf
[18:38] <ch33z> do edit the virtual hosts there?
[18:38] <qman___> no
[18:39] <ikonia> ch33z: I'm not interested in progressing this further as I said, as I don't help people who lie/waste my time
[18:39] <ch33z> i see..
[18:39] <qman___> the virtual hosts are stored in their own configuration files in /etc/apache2/sites-available, as explained in the server guide
[18:39] <qman___> which walks you through, step by step, setting this up
[18:39] <ch33z> o
[18:39] <ch33z> thanks
[18:39] <qman___> please read it
[18:39] <ikonia> this is all documented in https://help.ubuntu.com site
[18:39] <ch33z> ah i think i found it...
[18:39] <ch33z> sudo cp /etc/apache2/sites-available/default /etc/apache2/sites-available/mynewsite
[20:18] <Tixos> hi qman___, was you here the other day when i was resolving a poor apache performace?
[20:24] <Tixos> it fixed for a day or so, but not request times are crazy slow again
[20:45] <Tixos> low RAM will cause apache to run slow?
[20:45] <Tixos> 500mb free is low?
[20:47] <Tixos> i need to find out if apache is using all that ram first
[20:50] <qman___> no, that's plenty
[20:50] <qman___> well, it depends on your load
[20:51] <qman___> but, for example, I run a very low load apache server on a machine with only 256MB total, along with several other applications, without issue
[20:51] <Tixos> hmm
[20:52] <Tixos> im out of ideas
[20:52] <Tixos> only thing that has changed was there was more RAM, it went down to 150mb before i restarted apache
[20:54] <Tixos> where can i begin to debug this?
[20:54] <qman___> I'd start where you started before
[20:54] <Tixos> Benchmarking localhost (be patient)...apr_socket_recv: Connection reset by peer (104)
[20:54] <Tixos> lol
[20:54] <qman___> test to find out whether it's a problem with your server or not, by checking from a local net
[20:55] <Tixos> like ?
[20:55] <Tixos> i am running 'ab' and 'httperf' locally
[20:56] <qman___> and the problem persists?
[20:56] <Tixos> yes  >    Request rate: 0.1 req/s (13939.8 ms/req)
[20:56] <Tixos> if it was you who helped me before, it went down to 230ms/req
[20:57] <qman___> well, there has to be some setting that is different from default causing the problem, so step two is to compare your configs with bone stock configs and find what's different
[20:57] <Tixos> no, it was 100% stock, before i changed keepalive to 1, and upped maxclients to 256
[20:57] <Tixos> but i have put prefork back to stock now to test that, but its the same
[20:59] <qman___> well, if it's not configs, it has to be environmental
[20:59] <qman___> firewall settings, filters, something up with the hardware, conflicting software package
[20:59] <qman___> running anything like that?
[21:00] <Tixos> how can i check what exactly is using all my RAM? and no nothing like that, it was literally upped 5 days ago, and i installed nothing as such
[21:00] <qman___> top
[21:00] <Tixos> i did this   >   ps aux| awk '/apach[e]/{total+=$4}END{print total}'     and it outputs 30
[21:00] <qman___> I forget what the key sequence is but RTM to find the sort by mem usage
[21:00] <Tixos> i read top and free manuals
[21:00] <Tixos> but there are many apache processes
[21:01] <Tixos> with top and htop you can just click the %MEM and it sorts, but there are many processes
[21:01] <qman___> that's both normal and configurable
[21:01] <Tixos> i need a total
[21:01] <Tixos> could mysql effect apache page load times? nothing else i can think of :(
[21:02] <qman___> if you're loading data from a database, yes
[21:02] <qman___> static html pages, no
[21:02] <Tixos> hmm, mysqld is using 14-20% CPU
[21:02] <Tixos> and maybe the RAM too
[21:03] <qman___> that's definitely not normal
[21:03] <qman___> you'd have to be making thousands of queries per second to get that high on average hardware
[21:04] <Tixos> i probably am
[21:04] <Tixos> i get alot of hits
[21:05] <qman___> define "alot"
[21:05] <Tixos> million/day
[21:05] <qman___> then mysql is not really up to your task
[21:05] <qman___> test with a static HTML page
[21:05] <qman___> if the problem goes away, you have your answer
[21:05] <Tixos> it doesnt
[21:05] <Tixos> good point
[21:06] <Tixos> it doesnt, the default apache page still has bad request time
[21:06] <Tixos> about 5-8 seconds atm
[21:06] <Tixos> compared to a few milliseconds a few days vback
[21:06] <Tixos> question, i used this script  >  http://www.pixelbeat.org/scripts/ps_mem.py
[21:06] <qman___> could be a compound problem
[21:06] <qman___> how's your disk activity
[21:06] <Tixos> it is saying total MEM usage is only 400mb
[21:07] <Tixos> but free -m says 3GB is being used
[21:07] <qman___> +/- buffers/cache
[21:07] <qman___> that's the real application usage
[21:07] <qman___> the top number is total, and should be nearly full all the time if your system is performing optimally
[21:07] <Tixos> vmstat
[21:07] <Tixos>  1  0      0 457152 138800 2383584    0    0     0    11    2    6  1  0 98  0
[21:10] <qman___> that's saying 2.3GB is in cache
[21:10] <qman___> which is normal
[21:10] <qman___> buffers and cache should be using 'the rest' of your memory after applications
[21:10] <Tixos> so if MEM is getting as low as 150mb, thats not an issue? so im looking in the wrong place here?
[21:10] <qman___> run free -m
[21:11] <qman___> ignore the first line
[21:11] <qman___> the second line is what your applications are doing
[21:11] <Tixos> used 1GB
[21:11] <Tixos> free 2.9
[21:11] <qman___> yeah, memory's not an issue there
[21:11] <Tixos> hmmmmm
[21:12] <qman___> check your % iowait with top
[21:12] <qman___> if it's consistently high your disk subsystem is not fast enough
[21:12] <Tixos> wat is that value called?
[21:12] <qman___> it's at the top, %wa
[21:13] <Tixos> hmm i dont see that
[21:13] <qman___> third line
[21:13] <Tixos> PR/NI/VIRT/RES/SHR/S/%CPU/%MEM
[21:14] <qman___> above that
[21:14] <qman___> Cpu(s):  6.4%us,  0.8%sy,  0.0%ni, 92.7%id,  0.1%wa,  0.0%hi,  0.0%si,  0.0%st
[21:14] <Tixos> hehe sorry man
[21:14] <Tixos> Cpu(s):  1.9%us,  0.3%sy,  0.0%ni, 97.8%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
[21:14] <qman___> yeah, no problems there
[21:15] <Tixos> so its apache config thats the issue here?
[21:15] <Tixos> so i can safely put max clients to 256 etc again?
[21:15] <Tixos> although if its not helping...
[21:15] <qman___> either apache or something it's interacting with
[21:15] <Tixos>  1286 mysql     20   0  382m  44m 7120 S   20  1.1 392:48.30 mysqld
[21:16] <Tixos> 20% cpu, issue?
[21:16] <Tixos> i looked at my.cnf, but didnt see anything i could really change
[21:16] <qman___> what kind of CPU are you running
[21:16] <qman___> for example, if you've got lots of cores at a low speed, that would be a normal measurement
[21:16] <qman___> because it's 20% of one core
[21:17] <Tixos> ah, yes i think it is, quad core
[21:17] <Tixos> xenon, sec
[21:17] <qman___> same thing if you've got power saving enabled
[21:17] <Tixos> 8 cores with threading? htop says
[21:17] <qman___> try setting it to full speed if it isn't already
[21:17] <Tixos> yes that 20% isnt on 1 core
[21:18] <qman___> cat /proc/cpuinfo to verify
[21:18] <Tixos> thanks
[21:18] <Tixos> Intel(R) Xeon(R) CPU E31270 @ 3.40GHz
[21:19] <qman___> cpu MHz		: 2100.000
[21:19] <Tixos> cpu MHz		: 1600.000
[21:19] <qman___> if they're running at 800 or something that would throw all these numbers out the window
[21:19] <qman___> so about half speed, it's using 20% of one from eight threads
[21:19] <qman___> that's more normal
[21:20] <Tixos> argh?
[21:20] <Tixos> cpu MHz		: 1600.000
[21:20] <Tixos> CPU is set at about half
[21:20] <Tixos> sorry wrong window
[21:21] <qman___> so, it's using 20% of one eighth of your CPU's power, while it's at half speed
[21:21] <qman___> which is really not that much
[21:21] <Tixos> na, htop splits the cores
[21:21] <Tixos> shows about 2% on each
[21:21] <Tixos> most of the time 1%
[21:21] <qman___> it juggles the load around
[21:21] <qman___> but the thread percentages are based on one core
[21:22] <Tixos> hmm, what else is there left to check?
[21:22] <qman___> in any case, from those numbers, doesn't look like a hardware issue
[21:22] <qman___> unless it's a bizarre thing with the NIC
[21:23] <Tixos> i am with softlayer, they are pretty decent
[21:23] <Tixos> but i had same issue with ubuntu installed on another host
[21:23] <Tixos> not with centos, so i think still that its configs?
[21:23] <qman___> could be a kernel tuning thing too
[21:24] <Tixos> can i paste you my apache.conf?
[21:24] <Tixos> maybe its something obvious
[21:24] <qman___> is this on real hardware, or cloud/vps
[21:24] <Tixos> dedicated
[21:25] <qman___> ok
[21:25] <Tixos> its $200/m
[21:25] <Tixos> so half decent
[21:25] <qman___> yeah, just wanted to make sure
[21:25] <qman___> cloud/VPS systems can introduce all kinds of bizarre performance issues
[21:25] <Tixos> how can i select all from nano? :P
[21:26] <Tixos> ah ill just stick it in public_html lol
[21:27] <Tixos> maybe i need even more than the 256 max clients i said, i cant see how though?
[21:27] <Tixos> i put back to 150 and im sure its even slower, we are talking 12seconds page load
[21:27] <Tixos> just to load the .conf file
[21:28] <Tixos> http://pastebin.com/Wt44p5EK
[21:28] <qman___> well, a million requests a day, evenly spread, is 11/sec
[21:28] <Tixos> it was pretty much getting the same hits while DNS was changing over though, and it was fine 2 days ago
[21:28] <Tixos> lighting fast
[21:28] <qman___> since they're probably not evenly spread, and since each hit is probably several requests
[21:28] <qman___> you're un the hundreds per day
[21:28] <qman___> err
[21:28] <qman___> second
[21:29] <Tixos> i pastebinned my conf
[21:29] <qman___> oh, a DNS change?
[21:29] <Tixos> maybe its keepalive settings still
[21:29] <qman___> is your apache logging DNS names?
[21:29] <Tixos> hostnamelookups = off
[21:29] <Tixos> if thats what you mean
[21:29] <qman___> ok
[21:30] <Tixos> and the conf is just default apart from keepalive which i changed, as you can see
[21:32] <qman___> why's your keepalivetimeout 1
[21:32] <Tixos> someone here told me to set it to that
[21:32] <Tixos> when i did, it fixed the issue
[21:32] <qman___> did you try disabling keepalive altogether?
[21:33] <Tixos> i think right at the beginning, but it didnt have effect, should i try it ?
[21:33] <qman___> yeah
[21:33] <Tixos> my centos httpd.conf was set to off
[21:33] <Tixos> and that seemed ok
[21:33] <qman___> it's enabled by default but if that setting affected it
[21:33] <qman___> that's what to try
[21:33] <Tixos> what are drawbacks of turning off?
[21:33] <Tixos> let me switch off sec
[21:34] <qman___> forces clients to make more connections
[21:34] <qman___> the idea behind it is the client makes one connection, and it stays active for 15 seconds while it makes all the requests it needs
[21:35] <Tixos> erm
[21:35] <Tixos> locally
[21:35] <Tixos> Request rate: 5489.9 req/s (0.2 ms/req)
[21:35] <Tixos> :S
[21:36] <Tixos> outside  >  Request rate: 0.3 req/s (3680.3 ms/req)
[21:36] <Tixos> ah, because i just restarted
[21:36] <Tixos> locally >  Request rate: 0.2 req/s (4288.0 ms/req)
[21:36] <Tixos> so, once everyone has made connections again, it slows right down?
[21:36] <qman___> high, but better
[21:37] <Patrickdk> oh, having issues again?
[21:37] <Tixos> i think in 10minutes it will be back above 20k
[21:37] <Tixos> yea :(
[21:37] <Tixos> what about my prefork settings?
[21:37] <Tixos> they are default ubuntu
[21:37] <Patrickdk> are you using php?
[21:37] <Tixos> i am, but i am testing against a static HTML page
[21:38] <Tixos> just a page with 'test' written on it
[21:38] <Tixos> its a virtualhost, if that matters?
[21:38] <Patrickdk> na
[21:38] <Patrickdk> one thing that can make apache work a lot nicer
[21:38] <Patrickdk> is to change php over to fastcgi, and change apache over to mpm-worker
[21:39] <Patrickdk> but that isn't a *simple* thing to do
[21:39] <Tixos> i read it, but not sure if that was the issue, and yes it looked compicated
[21:39] <Patrickdk> well, it will fix some of your issue
[21:39] <Patrickdk> static pages wouldn't be slow anymore
[21:39] <Patrickdk> cause php pages won't get in the way
[21:39] <Patrickdk> since they will be offloaded to php
[21:39] <Patrickdk> instead of using the php build into apache (mod_php)
[21:40] <Tixos> i pastbinned my apache.conf Patrickdk, and qman has ruled out alot of things already if you want to scrollback a little
[21:40] <Patrickdk> but I thought we tuned it up some
[21:40] <Tixos> yea, its back with vengance
[21:40] <qman___> yeah, it could be that all your processes are busy handling php requests and thus can't return static pages quickly enough
[21:41] <Tixos> so tune PHP?
[21:41] <Patrickdk> that is what it was the other day atleast
[21:41] <Patrickdk> well, between that, and keepalive keeping apaches busy being idle
[21:41] <Tixos> or prefork?
[21:41] <Tixos> keepalive is off now
[21:41] <Patrickdk> ya, so now it can only be two issues
[21:41] <Patrickdk> php taking too long
[21:41] <Patrickdk> or clients downloading webpages too slow
[21:41] <Patrickdk> increasing prefork stuff would fix that
[21:41] <Patrickdk> if you have enough memory to handle it
[21:42] <Tixos> what is max i can set to?
[21:42] <Tixos> it was on 256
[21:42] <qman___> if php is the problem, I think turning maxclients way up would make it slightly faster
[21:42] <qman___> though not exactly the right solution
[21:42] <Patrickdk> yep
[21:42] <Tixos> im open for testing :)
[21:42] <qman___> like a couple thousand
[21:42] <Patrickdk> quick and simple fix, then handle looking into php more :)
[21:42] <Tixos> wtf
[21:42] <Tixos> i read over 256 is dangerous
[21:42] <Tixos> lol
[21:43] <Patrickdk> I run over 1000
[21:43] <Patrickdk> but it depends on your memory
[21:43] <Tixos> should i try and monitor ram?
[21:43] <Tixos> its 4gb ddr3
[21:43] <Patrickdk> take each one will use 20-30megs (cause of php)
[21:43] <qman___> you've got over 2GB available
[21:43] <qman___> so
[21:44] <Patrickdk> as long as you don't swap, you can increase :)
[21:44] <qman___> well, what's your memory at with 256?
[21:44] <Tixos> i cant remember
[21:44] <Tixos> let me switch over
[21:44] <Tixos> what else do i need to up apart frmo maxclients?
[21:44] <qman___> just the clients
[21:44] <Tixos> before i put childerequests to 4000, i just copied my centos config
[21:44] <Tixos> and startservers to 8 etc
[21:44] <qman___> the reasoning is, each client handles a php request, then waits o nit
[21:44] <qman___> so more clients means more simultaneous php requests
[21:45] <qman___> if php is the problem this will improve performance
[21:45] <Tixos> ok lets see
[21:45] <Tixos> MaxRequestsPerChild   0
[21:45] <Tixos> leave this?
[21:45] <Patrickdk> that means never kill/restart one
[21:46] <Patrickdk> that is better, IF you have no memleaks
[21:46] <Tixos> right, so i want to monitor ram?
[21:46] <Tixos> -/+ buffers/cache:       1022       2920
[21:46] <Tixos> so i should see used grow to say 2GB?
[21:47] <qman___> ok, and from your numbers before, around 500MB of that is apache
[21:47] <qman___> so four times the number you're using should be save
[21:47] <qman___> 1024
[21:47] <Tixos> let me try that script again
[21:47] <Tixos> 330.0 MiB +  30.4 MiB = 360.4 MiB	apache2 (165)
[21:47] <Tixos> 165 processes?
[21:47] <qman___> ok
[21:47] <Tixos> lets give it 5
[21:48] <qman___> 165 is using 360, 1000 should be safe
[21:49] <Tixos> 419.2 MiB +  36.2 MiB = 455.4 MiB	apache2 (208)
[21:49] <qman___> remaining consistent
[21:50] <Tixos> upto 217 now, so it doesnt need the 256?
[21:50] <Tixos> or it will grow and grow?
[21:50] <qman___> it grows as needed, based on the other settings
[21:50] <Tixos> i see
[21:50] <qman___> the fact that it is growing means you're placing a load on it that needs more clients
[21:51] <Tixos> thanks guys, learning is the best part :)
[21:51] <qman___> which is also consistent with the theory
[21:54] <Tixos> ok, upto 240 now,
[21:55] <Tixos> 532.6 MiB +  43.7 MiB = 576.3 MiB	apache2 (256)
[21:56]  * michael_tn waves to all
[21:56] <michael_tn> anyone around to talk about ochestra/juju?
[21:58] <Tixos> Patrickdk: / qman___ i must also increase ServerLimit ? which isnt even there on default ubuntu config
[22:00] <qman___> from what I'm reading, yes
[22:00] <Tixos> ok, i will try 1024 and see :S
[22:00] <Tixos> lol
[22:01] <Tixos> Patrickdk: what did you say you was using? and how much ram ?
[22:01] <Tixos> maybe i need to look into worker MPM, but thats for another day i think
[22:01] <Tixos> think it requires reinstall of apache
[22:01] <Patrickdk> not a reinstall, but pkg change
[22:01] <Patrickdk> and you can't use mod_php with worker
[22:02] <michael_tn> nod, package/config change should set that
[22:02] <Tixos> ok, if i set to 1024
[22:02] <Tixos> should i learn StartServers etc?
[22:02] <Tixos> or increase those
[22:02] <Patrickdk> just do startservers at like 100 or so
[22:02] <Tixos> really? lol
[22:02] <Patrickdk> that is just how many it makes, when you START apache, like on server reboot
[22:02] <qman___> yeah, it clearly needs them
[22:02] <Tixos> i feel something is going to pop :P
[22:03] <qman___> should make it run faster off the bat
[22:03] <Patrickdk> if your running 300, a 100 to start is more than low :)
[22:03] <Tixos> okies :)
[22:03] <Tixos> boom!
[22:03] <Tixos>     MinSpareServers       5    +    MaxSpareServers      10
[22:03] <Patrickdk> I'm running worker, with 8 forks, and 150 threads each
[22:03] <Patrickdk> probably want to widen that window some
[22:04] <Patrickdk> minspare, 50, maxspare 100?
[22:04] <Patrickdk> if you really need hundreds of them
[22:04] <Tixos> i duno
[22:04] <Tixos> :O
[22:04] <Tixos> lol
[22:04] <Tixos> lets try 25/50 ?
[22:04] <Patrickdk> ok
[22:04] <Patrickdk> that means there will always be 25extra ones ready to use
[22:04] <Patrickdk> and no more than 50 extra ones
[22:05] <Tixos> so you think higher if i have such a high maxclients?
[22:05] <Patrickdk> it depends on your client usage
[22:05] <Tixos> maybe they will use all 1024?
[22:05] <Patrickdk> if they all spike at once, or if it's more even
[22:05] <Tixos> o right
[22:05] <Tixos> lets go 25/50 and see
[22:06] <Tixos> and leave this, even though centos has 4000 set here      >    MaxRequestsPerChild   0
[22:07] <Tixos> ok ill update you in an hour or so, should it really fill the 1024?
[22:07] <Patrickdk> if you have other issues, sure :)
[22:07] <Patrickdk> and if it's that bad
[22:07] <Tixos> so it should never hit the 1024?
[22:09] <Tixos> that would be 4000 hits per second? @ 250ms per request? maybe im totally wrong :P
[22:33] <Tixos> Patrickdk: looking ok so far,  757.0 MiB +  58.9 MiB = 815.9 MiB	apache2 (359)
[22:48] <yakster> can someone give me a hint how to mount a USB HDD connected to Airport Extreme base station?
[23:17] <RoyK> [offtopic] what are the ubuntu irc guidelines for handling people that behave like either drunks or young teens, threatening with different unpleasent scenarias?
[23:17] <RoyK> off the channel, after a small discussion _in_ the channel
[23:23] <SpamapS> RoyK: That would be a violation of the Ubuntu CoC, and usually is dealt with by first asking them to stop, and then asking them to leave by way of the community council. If they haven't signed the CoC.. I'd think a kick/ban is appropriate.. but I don't know official policy. :)
[23:36] <Cerin> hi, I just installed 11.10 on a barebones Asus RS100-E7 with integrated Aspeed graphics, and I can't get Xorg to start. All I get is the error "screens found, but none have a usable configuration. fatal server error: no screens found"
[23:36] <Cerin> I've been Googling, but I can't find any resolutions. Does anyone know where I should look for a fix?
[23:49] <SpamapS> Cerin: wrong channel? Xorg and servers aren't really something that should be used together. :)
[23:49] <Cerin> SpamapS: I don't disagree. I only want to confirm it works, so that it's an option.
[23:50] <SpamapS> Cerin: perhaps #ubuntu or #ubuntu-desktop would have better answers.
[23:50] <Cerin> #ubuntu didn't have any suggestions
[23:50] <SpamapS> Cerin: really, servers should never have X on them. :)
[23:50] <Cerin> since it's an xorg specific issue, I thought I'd ask here
[23:51] <SpamapS> Cerin: right, #ubunt-desktop has X experts usually
[23:51] <Cerin> that's not really helpful :)
[23:51] <Cerin> thanks
[23:51] <SpamapS> Cerin: I'm trying to help you, by telling you, X will make your server less stable, and so, should not be installed.
[23:55] <Cerin> SpamapS: yeah, I know. and that's probably what I'll do. I'm still a little new to a pure command line environment, so I was going to use X as a crutch for a while before I felt more comfortable, but since this issue looks unresolvable, I might just bite the bullet
[23:56] <SpamapS> Cerin: what exactly will you get in the GUI that you are afraid to lose?
[23:56] <SpamapS> Cerin: I'm genuinely curious
[23:57] <GhostFreeman> If you're so concerned about not having a good degree of control w/o a GUI, there are tools like cpanel on Packages you can install to help admin your server
[23:58] <SpamapS> ugh
[23:58] <SpamapS> not cpanel.. ;)
[23:59] <GhostFreeman> there's also Launchpad
[23:59] <GhostFreeman> really though the Server Guide is concise about teaching you the commands you need and how to use them