[02:28] <hydrogen18> would anyone care to me help with a lvm problem?
[02:29] <twb> !anyone
[02:30] <hydrogen18>  Hello, I have a problem: I have a logical volume which has "#open" of 1 when I run lvdisplay. But looking in /proc/mount it is not mounted and 'lsof' does not return anything having it open. 'fuser -ki' does not prompt me to kill any processes when run against the logical volume. I am running ubuntu 10.04 server amd64
[02:30] <twb> hydrogen18: any processes wedged in D state?
[02:31] <hydrogen18> twb: how do I check for that
[02:32] <twb> ps auxf | grep -2 D
[02:32] <twb> There is a column where most stuff is "S" (idling), a few "R" (running), and occasionally "D" (wedged)
[02:32] <hydrogen18> it would seem not.
[02:34] <hydrogen18> there are definitely none in 'D' state
[02:35] <twb> Then I don't know
[02:35] <twb> occasionally stuff just gets like that and it's impossible to tell why
[02:35] <hydrogen18> so im just screwed?
[02:36] <twb> A reboot will definitely clear it, so that's your worst-case scenario
[02:36] <twb> I would also try umount -f
[02:36] <hydrogen18> I guess I'm doing a recovery then.
[02:36] <hydrogen18> its not mound
[02:36] <hydrogen18> mounted*
[02:36] <hydrogen18> i mean I guess im doing a reboot
[02:36] <twb> Are you sure it's not just that it's active?
[02:36] <hydrogen18> how do I check to see if it is active?
[02:36] <twb> lvchange -an /dev/foo/bar
[02:37] <hydrogen18> " LV dyson/swarm in use: not deactivating" what brought me here
[02:37] <hydrogen18> ''sudo lsof | grep swarm' is blank
[02:38] <twb> Dunno man
[02:38] <twb> You can try forcing it
[02:38] <hydrogen18> i don't think you can force an lv offline
[02:44] <hydrogen18> yeah...restarted, now marked as #open 0
[02:44] <hydrogen18> now im running a fsck on a 6.37 TiB filesystem...probably take all nite
[02:46] <twb> Yeah so maybe don't make multi-TiB filesystems
[02:46] <twb> And at least make them journalled or checksummed
[02:47] <twb> People who use LVM and then lvextend --size 100% need to be LARTs
[02:47] <twb> *LARTed
[02:54] <hydrogen18> There do not seem to be many other ways to glue physical devices together
[02:55] <twb> hydrogen18: well, I would hope that if you ahve that much data you have some kind of organization of it
[02:56] <twb> so e.g. make /srv/porn/interracial and /srv/porn/milfs separate LVs
[03:06] <Solaris444> does anybody know if 10.04 LTS server has native support for the LSI MegaRAID 9260-4i RAID controller?
[03:08] <qman__> Solaris444, I don't know, but it's pretty simple to get the answer to
[03:08] <qman__> find out what driver that card needs, what kernel version it was added in, and compare to the current lucid kernel version
[03:09] <Solaris444> Yes, I've tried to do that and can't get a straight answer.
[03:09] <Solaris444> LSI don't tell you on their website.
[03:09] <Solaris444> It just says "native support is included in some newer linux kernels, but we don't support it.  Use our driver here"
[03:09] <Solaris444> which is no help at all.
[03:10] <qman__> download the driver and find out what the name is
[03:10] <Solaris444> I've even contacted LSI directly and still couldn't get a straight answer.
[03:10] <qman__> then see if it's in the lucid kernel or modules
[03:10] <Solaris444> the kernel module is megaraid_sas
[03:11] <Solaris444> Which I believe is included, but the question is whether it supports that specific card.
[03:11] <qman__> well surely it should have a version or date
[03:11] <Solaris444> You would think so...
[03:14] <Solaris444> ok apparently the version on their website is 00.00.04.37
[03:14] <qman__> well, what I can tell you is the module on my file server is version 00.00.04.01
[03:15] <Solaris444> but that doesn't mean that was the first version of the driver to include support.
[03:15] <Solaris444> that's just the current version.
[03:15] <qman__> granted I haven't updated the kernel in a couple months
[03:16] <Solaris444> Well let me put it this way...
[03:16] <Solaris444> If I use the kernel module from their website...
[03:16] <Solaris444> I can't EVER update the kernel on the ubuntu box can i.
[03:17] <qman__> this should be helpful: http://lxr.free-electrons.com/source/Documentation/scsi/ChangeLog.megaraid_sas?v=2.6.32
[03:18] <Solaris444> incidentally, what kernel version are you running?
[03:18] <Solaris444> thankyou, reading now.
[03:18] <qman__> might be hard to find it in there, but that's the complete changelog
[03:19] <qman__> if you can find out where in it support was added, that will answer your question
[03:19] <qman__> currently running 2.6.32-32
[03:19] <Solaris444> I think the 9260 is a new generation controller.
[03:20] <qman__> that link is to my version, but it has links to all the others
[03:21] <Solaris444> hello hello, check this out
[03:21] <Solaris444> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/546091
[03:21] <Solaris444> looks like there was a bug but it appears to have been fixed in 2.6.32-26
[03:22] <qman__> lucky find
[03:22] <qman__> that confirms it
[03:23] <Solaris444> question though... if I *were* to use the kernel module on lsi's website...
[03:24] <Solaris444> I can't ever update the kernel, correct?
[03:24] <qman__> it would have to be recompiled with every update
[03:24] <qman__> and running production with kernel sources and compilers installed can be considered risky
[03:25] <qman__> Back in the Day (tm) I used to have to do that to run decent versions of nvidia drivers on my desktop
[03:26] <qman__> it was a bit of a nuisance in that case, can't imagine running a server that way
[03:26] <Solaris444> yeah, that's what i mean.  if I apply the kernel update, I'd have to recompile *prior* to rebooting, correct?
[03:26] <qman__> yes
[03:26] <qman__> well, prior to booting the new kernel
[03:27] <Solaris444> That's a really bad idea on a production mail server.
[03:27] <qman__> which it will try to do by default
[03:27] <qman__> and fail because of
[03:27] <Solaris444> right
[03:28] <qman__> you could always hack together some scripts to do it for you
[03:29] <qman__> but we all know how that goes
[03:29] <qman__> something that delicate would surely have just as many problems for a while
[03:38] <Solaris444> agreed.
[03:39] <Solaris444> unfortunately the final post in the bug claims it *isn't* fixed for the specific card the bug reporter has.
[03:39] <Solaris444> though it's newer than the reported card.
[03:43] <qman__> well, that kernel version is pretty old at this point, too
[03:44] <Solaris444> yeah, so i see.
[03:45] <qman__> there's been another re-roll of the disc since then, too
[03:45] <Solaris444> what version is the disc at now?
[03:46] <qman__> 10.04.3
[03:49] <qman__> not sure which kernel version is on the disc
[03:50] <qman__> my K6 is running 2.6.32-28 and claims 10.04.3, but that's also running the -386 kernel, and I don't know if it can claim minor version status without a kernel update
[03:51] <Solaris444> *nods*
[03:51] <Solaris444> I understand.
[03:52] <qman__> none of my servers are were actually new lucid installs, I upgraded them from hardy
[03:52] <qman__> so I could be dragging some baggage
[04:00] <twb> 13:06 <Solaris444> does anybody know if 10.04 LTS server has native support for the LSI MegaRAID 9260-4i RAID controller?
[04:01] <twb> Solaris444: does it show up as in lspci?
[04:01] <Solaris444> I don't have the system yet.
[04:01] <Solaris444> I won't purchase it if it doesn't.
[04:01] <twb> OK.  Well, if it's hardware raid, it will not matter because ubuntu will just see a disk (the virtual disk)
[04:02] <Solaris444> ?
[04:02] <Solaris444> It's a hardware raid card, but it still needs drivers...
[04:02] <twb> However if you want to be able to configure LSI while ubuntu is running, i.e. without rebooting into the BIOS, then you will need stupid userland tools, and they can be made to go but they're horrible and painful.
[04:02] <Solaris444> Booting into BIOS is fine.
[04:03] <Solaris444> I mean, if the array were to lose a disk, I'd take the system offline to do it anyway.
[04:03] <twb> Then I *think* it'll be fine, but I would still recommend using md RAID instead unless you definitely need the extra throughput
[04:03] <twb> Also you will want to buy a spare card in case the first one dies
[04:03] <Solaris444> But I did see a bug filed against 10.04 which claimed that some cards were not able to see the attached disks.
[04:03] <Solaris444> spare card?
[04:03] <Solaris444> that's a bit extreme isn't it?
[05:01] <idlemind324> hey. i have 5 static ips and i am setting up a router box if i assign 1 of the ip addresses in my range only traffic for that single address comes through. how do i tell it to listen for the whole range?
[05:10] <twb> Solaris444: if your one card dies, and you can't buy a replacement because they're not made anymore, then all your data is gone
[05:11] <twb> Solaris444: hardware RAID cards are not interchangable, whereas md RAID you are fine as long as you have a new enough Linux kernel on your rescue system
[05:13] <Solaris444> true.
[05:13] <Solaris444> performance isn't as good of course.
[05:13] <twb> Correct
[05:13] <twb> But if it's just for an office's word documents and stuff, it doesn't matter
[05:14] <twb> If it's the database server for, say, nintendo.com or ASX, that's different.
[05:14] <SpamapS> idlemind324: you need to assign the addresses as aliases in /etc/network/interfaces
[05:15] <idlemind324> spamaps: i can't do it through some kind of iptables rule? how is it done on cisco appliances and things of that nature?
[05:16] <SpamapS> idlemind324: no you can't because you need it to respond to arp requests
[05:17] <SpamapS> idlemind324: cisco appliances would also assign the addresses to the interface
[05:17] <idlemind324> k i'm tossing on aliases atm 1 sec
[05:18] <SpamapS> idlemind324: just make another interface..   eth0:1 .. eth0:2 .. with the address/netmask/broadcast
[05:18] <SpamapS> idlemind324: make sure its marked as 'auto' too so its brought up at boot time
[05:19] <idlemind324> for the aliases i shouldn't enter a gateway for them right?
[05:21] <idlemind324> ok made them aliases but i can't ping any of the alias addresses from an outside server but i can ping the ip i gave the physical interface
[05:21] <idlemind324> do i need to assign mac addresses to each alias?
[05:22] <twb> eth0:1 is ifconfig failspeak
[05:22] <twb> ip is where it's at
[05:23] <twb> http://paste.debian.net/133151/
[05:24] <SpamapS> twb: how exactly does that translate to /etc/network/interfaces ?
[05:24] <twb> SpamapS: you put it in interfaces(5)
[05:25] <twb> That paste I showed, that was straight out of /etc/network/interfaces
[05:25] <SpamapS> unmanaged inet manual .. hehe.. interesting
[05:25] <SpamapS> twb: have never seen that before
[05:26] <twb> "unmanaged" is instead of "eth0"
[05:26] <SpamapS> twb: the part I've not seen is manual setup..
[05:26] <twb> You can name your ifaces logically in /etc/udev/70-persistent-net.rules -- this is very useful when writing an /etc/iptab (firewall), because it means you don't need to use variables and thus don't need m4 or cpp or sh
[05:26] <SpamapS> seems more BSD-like. :)
[05:27] <twb> manual is just instead of static, so we can use ip instead of ifupdown's shitty old ifconfig
[05:27] <twb> ifupdown is written in noweb FFS
[05:27] <SpamapS> I know
[05:27] <SpamapS> its a nightmare
[05:27] <SpamapS> been digging around in it this release
[05:27] <Solaris444> twb it's a mail and file server for a graphical design company
[05:27] <idlemind324> http://paste.debian.net/133152/
[05:27] <twb> Solaris444: shrug
[05:28] <idlemind324> ^^ thoughts on why eth0:1 isn't working
[05:28] <twb> SpamapS: IMO it should be just rewritten from the ground up to be less stupid
[05:28] <SpamapS> twb: would you be willing to write that up for the server guide? seriously, much nicer
[05:28] <twb> SpamapS: but knowing canonical they will probably fold it into systemd or something
[05:28] <SpamapS> twb: systemd.. who said anything about systemd?
[05:28] <SpamapS> <--- Canonical :)
[05:28] <twb> Yes, but you lost all your upstart people
[05:29] <twb> to google or systemd, AFAICT
[05:29] <SpamapS> <--- upstart person
[05:29] <twb> OK, whatever
[05:29] <SpamapS> jhunt would be crushed to know that we lost him too
[05:30] <SpamapS> twb: upstart is, AFAIK, still the pid 1 for chromeos
[05:30] <SpamapS> twb: and we've already stated repeatedly that upstart isn't going anywhere before 12.04.
[05:30] <twb> OK, whatever
[05:31] <SpamapS> <-- sensitive about that
[05:31] <SpamapS> But yeah, ifupdown is crap and I'd love to replace it with something simpler.
[05:31] <twb> FWIW I'm sick of both upstart and systemd
[05:32] <SpamapS> I think the way we introduced upstart was a bit manic.. WAY too much explicit event listing for things that should all just start on runlevel 2.
[05:32] <SpamapS> And the shutdown.. makes me sad.
[05:32] <twb> I want some determinism and better debugging
[05:33] <twb> Last time I looked if your system hung during boot because mountall got its knickers twisted, you could eitehr get no debugging, or pass --verbose which made upstart emit thousands of lines a second and you can't read them before they scroll offscreen, AND they aren't written to printk so you can't see them by modprobing netconsole before pivot_root
[05:34] <twb> I don't really care about stuff like "but what if you want to suspend the system, go home, plug into a new network, and have stuff work" -- because I don't DO that with my rack full of servers
[05:34] <idlemind324> o wait give it a few minutes and tada the alias addresses start workin
[05:35] <SpamapS> twb: actually they are written to printk as of 1.3
[05:35] <twb> not in lts tho
[05:36] <SpamapS> Yeah, hopefully we won't have to make so many apologies for 12.04. ;)
[05:36] <twb> I'm sure you have been working hard to introduce new and exciting problems tho :P
[05:37] <SpamapS> Indeed, I think we call that "progress"
[05:37] <twb> I did an upgrade of my sid netbook the other day, and bluez, sudo and fbi were all broken
[05:37] <twb> twb SMASH!
[06:58] <twb> IIRC, d-i won't let you create a user called "admin"
[06:58] <twb> Why is this?  The user/group doesn't seem to exist in the static 0..100 area
[06:58] <_ruben> because it'd create a group admin, which is 'default'
[06:58] <_ruben> iirc
[06:58] <twb> But I can't see one in /etc/group either
[06:59] <twb> There is an "adm" but that's different
[06:59] <_ruben> i have admin on mine
[06:59] <twb> What gid ?
[06:59] <_ruben> probably created by some app .. it's 113
[06:59] <twb> OK, I will grep over the postinsts in my package mirror to see if I can find it
[07:01] <twb> Actually, can you save me a shitload of un-ar'ing and fgrep admin /var/lib/dpkg/info/*.p* for me?
[07:03] <_ruben> doesn't yield much useful stuff
[07:05] <twb> Mmm, I'm seeing that
[07:06] <twb> I did find a box with that group
[07:06] <twb> WTF, coworker is suggesting sudo
[07:06] <twb> Which I might not see because I'm using sudo-ldap...
[07:08] <twb> 113 and 116 seem too high for sudo tho
[07:08] <twb> Given sudo is installed quite early typically
[07:09] <twb> I reckon it's samba
[07:10] <_ruben> no samba on the box i checked
[07:10] <twb> it's always the group right after samba
[07:10] <twb> sambashare that is
[07:11] <_ruben> here it's after snmp
[07:11] <twb> OK
[07:11] <_ruben> both start with s... interesting? :)
[07:11] <_ruben> sudo follows rather closely ;)
[07:11] <_ruben> but doesnt seem to do it in postinst
[07:13] <twb> The sudo guess is that in our sudoers, we have both %sudo and %admin
[07:15] <_ruben> same here, but when looking at postinst, i only see it create sudo
[07:18] <twb> Oh well, I'm going to <beer> now, so don't stress over it
[07:20] <_ruben> Didn't intent to ;)
[07:51] <koolhead17> hi all
[09:00] <osmosis> my /boot is almost full. How do I remove unused kernels?
[09:09] <_ruben> just like you would any other package
[09:34] <bigjools> has anyone experienced an empty resolv.conf when using a static network interface in oneiric?
[09:37] <ikonia> bigjools: did you define the name serves in the static network ?
[09:37] <bigjools> yep
[09:37] <bigjools> this was working in natty until I upgraded to oneiric
[09:38] <ikonia> bigjools: if you try to look things up does it actually work ?
[09:38] <bigjools> no
[09:38] <ikonia> you set it up in gnome-network-maanger I assume ?
[09:38] <bigjools> not until I manually edit resolv.conf (which is wiped at next book)
[09:38] <bigjools> no, I edited /etc/network/interfaces
[09:39] <ikonia> why are you setting it there and not in gnome-network-manager ?
[09:39] <ikonia> have you disabled gnome-network-manager if you're using the interface file ?
[09:39] <bigjools> because I don't have  gnome-network-manager
[09:39] <ikonia> bigjools: what desktop are you using ?
[09:39] <bigjools> this is the server channel, right? :)
[09:40] <ikonia> my god, yes, so it is. I thought you where in #ubuntu+1 as you where talking about 11.10
[09:40] <bigjools> :)
[09:40] <ikonia> can you paste your interfaces file for me please (pastebin_)
[09:41] <bigjools> ikonia: http://pastebin.ubuntu.com/699610/
[09:41] <ikonia> thanks, one moment
[09:41] <ikonia> ok, have to be honest, never seen dns-nameservers or dns-search set in interfaces file
[09:42] <ikonia> I'd honestly remove that, and just put them in /etc/resolv.conf
[09:42] <ikonia> also see if you have network-manager installed (dpkg -l | grep network-manager_)
[09:42] <ikonia> (without the _ )
[09:42] <bigjools> I do, it seems
[09:43] <ikonia> maybe worth removing it, seeing as you're not using it
[09:43] <ikonia> (remove any potentialy issue)
[09:43] <bigjools> indeed, thanks for the advice
[09:43] <bigjools> I guess something changed in oneiric's n-m
[09:43] <ikonia> either that or the upgrade just enabled it
[09:43] <ikonia> and it wasn't enabled in your natty box
[09:49] <ersi> I have both dns-nameservers and dns-search in my interfaces file. Works great.
[09:49] <ersi> Oh, it won't work in Oneiric?
[09:50] <bigjools> seems that way
[09:50] <bigjools> maybe the directives changed
[09:55] <ersi> Or maybe you don't have resolveconf installed anymore
[09:56] <koolhead17> zul: i can see keystone.conf finally but its still broken and now there is no keystone manpage
[10:00] <rbasak> Should https://bugs.launchpad.net/ubuntu/+source/postfix/+bug/862915 be Importance: Medium? I've just set it to Low.
[10:05] <rbasak> To answer my own question, it just occurred to me that no, it has an easy workaround so that's Low by definition.
[10:31] <koolhead17> hey kirkland
[10:33] <ersi> rbasak: Indeed :)
[11:08] <kirkland> koolhead17: hi
[11:13] <Dulcin> Hi, I just installed postfix on ubuntu 11 server and I only want to use it to send mails with php. Can anyone give me some hints on how to make sure that nobody else can use the mailserver?
[11:14] <Dulcin> I read a little about SMTP Auth with postfix, but I'm not sure if this is required since I do not want users to have access at all
[11:17] <koolhead17> Dulcin: are other users on same system has ssh access?
[11:21] <Dulcin> yeah they do
[11:21] <Dulcin> but I dont care, those are only admins
[11:21] <Dulcin> basically I have a webserver, and it needs to send e-mails automatically
[11:22] <Dulcin> but it has to be secure, so nobody else should be able to use the smtp server
[11:22] <pmatulis> Dulcin: do not accept local mail (no local mail delivery) and do not accept to relay (see relay_domains)
[11:22] <pmatulis> Dulcin: see the official postfix documentation, it's not hard
[11:23] <Dulcin> Yeah I started reading that, but then I realised it was to securing authorized users only, so I got confused
[11:23] <Dulcin> as I didn't want any users :)
[11:24] <Dulcin> Ill check those points you mentioned and skim through the documentation a bit more to see if I can find what I need then
[11:24] <pmatulis> Dulcin: but you're supposed to always accept for postmaster and MAIL_DAEMON users (according to standards)...
[11:26] <Dulcin> I'm really new to this, so I can only guess what that means at this point :)
[11:27] <Dulcin> but I'll read through the postfix docs
[11:27] <Dulcin> I was reading this: http://www.postfix.org/SASL_README.html
[12:10] <john> i need some help i want logmein-hamachi on my server but it complains about lsd-core and i cant get it or even lsd-base
[12:13] <john> shadowdragon@ShadowDragonDen:~$ sudo dpkg -i logmein-hamachi_2.1.0.17-1_i386.deb
[12:13] <john> (Reading database ... 57941 files and directories currently installed.)
[12:13] <john> Preparing to replace logmein-hamachi 2.1.0.17-1 (using logmein-hamachi_2.1.0.17-1_i386.deb) ...
[12:13] <john> Unpacking replacement logmein-hamachi ...
[12:13] <john> dpkg: dependency problems prevent configuration of logmein-hamachi:
[12:13] <john>  logmein-hamachi depends on lsb-core (>= 3.0); however:
[12:13] <john>   Package lsb-core is not installed.
[12:13] <john> dpkg: error processing logmein-hamachi (--install):
[12:13] <john>  dependency problems - leaving unconfigured
[12:13] <john> Processing triggers for ureadahead ...
[12:13] <john> ureadahead will be reprofiled on next reboot
[12:14] <john> Errors were encountered while processing:
[12:14] <john>  logmein-hamachi
[12:14] <john> any one can help me figure this oen out?
[12:14] <john> one*
[12:14] <jussi> !pastebin | john
[12:14] <jussi> john: but it looks like you need to do: sudo apt-get install lsb-core
[12:16] <john> tried that and sorry ill try to use pastebin next time
[12:17] <john> logmein works on my desktop but try to install it on my server i just set up tonight and it gives me that problem
[12:17] <john> and wont let me get lsb-core for some reason
[12:24] <john> oh and after i run dpkg to isntall the deb file or use apt to it constantly shows me logmein needs lsb core
[12:24] <john> i cant do any thing with apt
[12:27] <pmatulis> john: so install lsb-core
[12:27] <pmatulis> john: what happens when you try?
[12:29] <john> http://paste.ubuntu.com/699716/
[12:30] <pmatulis> john: what release is this?
[12:30] <john> ubuntu server natty 32bit
[12:30] <pmatulis> john: maybe pastebin /etc/apt/sources.list
[12:36] <john> ok here it is http://paste.ubuntu.com/699719/
[12:37] <pmatulis> john: did you try 'sudo apt-get -f install' ?
[12:37] <john> yep still nothing
[12:37] <pmatulis> john: what was it's output?
[12:38] <john> hmm i guess i should have tried it whout the lsb-core and stuff
[12:38] <john> couse now its downloading things
[12:39] <pmatulis> john: right
[12:39] <john> kinda new to the server thing i have only used ubuntu for about two years now
[12:39] <janimo> do the server developers using the ac100 arm netbook want some specific options present in the oneiric kernel? (DM, RAID, other things which normally would not make sense on such a machine)
[12:41] <john> there we go now its working thanks for being patient with me
[12:41] <pmatulis> john: k, keep on truck'n
[12:42] <janimo> ogra_, do you know which specific kernel devs have ac100 so I can ask them directly?
[12:42] <janimo> s/kernel/server/
[12:42] <ogra_> janimo, hallyn_, Daviey and zul i think
[12:42] <ogra_> there might be more i dont know about
[12:53] <zul> good morning
[12:55] <zul> janimo: yes :)
[12:57] <rbasak> zul, I was going to send you a patch/review request for my nova-compute patch but found that you've already done it. Did you base it on patch set 1 or 2 from gerrit, or did you get it from somewhere else? There are bugs in patch set 1, patch set 2 is the one I'm happy with.
[12:58] <rbasak> Also patch set 2 includes restarting the logger on nova-compute restart
[12:58] <zul> rbasak: im going to take patch set 2
[12:58] <rbasak> OK cool
[12:59] <rbasak> Two things that I haven't figured out how to test: rescue mode and live migration. But all the tests pass on patch set 2.
[12:59] <zul> rbasak: i need to update my lxc console patch but i found a problem with the lxc console patch yesterday to use console.fifo so that will be uploaded today after i figure it out
[12:59] <zul> k cool
[12:59] <rbasak> Also, it seems that if nova-compute is stopped while a guest dumps >64k to the console then the guest hangs
[13:00] <rbasak> If it dumps less than the logger picks it up out of the fifo buffer and it continues fine
[13:00] <rbasak> s/than/then
[13:00] <smoser> rbasak, so are you going to add ring buffer support to kvm next cycle ?
[13:00] <rbasak> smoser: yeah, that does seem like a better way to do it
[13:00] <rbasak> smoser: assuming everyone can agree on a ring buffer format, and libvirt will need to have support added too
[13:00] <smoser> the guest hangs ?
[13:01] <smoser> as in writes to /dev/ttyS0 hang
[13:01] <smoser> right?
[13:01] <rbasak> no, it seemed to hang the guest completely
[13:01] <rbasak> I need to retest to verify that
[13:01] <smoser> (not every one needs to agree, someone needs to do it, and thell others this is how it is!)
[13:01] <rbasak> in case it was something else that caused it
[13:01] <zul> smoser: lxc? or something else
[13:01] <smoser> it would seem strange to me that a write to a bit of hardware should stop the cpu for a machine.
[13:01] <rbasak> well the ringbuffer format I've implemented is how it is :-)
[13:01] <smoser> :)
[13:02] <smoser> rbasak, i agree.
[13:02] <rbasak> I think qemu might be assuming that the fifo reader never disappears
[13:02] <rbasak> I really should retest to verify though
[13:02] <smoser> zul, what issue did you have with your lxc console ?
[13:02] <KM0201> anyone know how to resolve thjis, trying to SSH my server, getting this error message on my client..   Warning: Permanently added '192.168.1.7' (RSA) to the list of known hosts.
[13:02] <KM0201> Permission denied (publickey).  (.7 is an ubuntu client... but i dont' know what file to modify to fix this)
[13:03] <smoser> you're getting permission denied, KM0201 . the other side of the connection is not letting you in.
[13:03] <KM0201> oh ok... hmm.. so this is a server prob, right?
[13:03] <smoser> the warning is telling you that the client system has seen this fingerprint of this server before, and so it is adding it.
[13:03] <smoser> KM0201, it would seem to be yes.
[13:03] <zul> smoser: that it displayed the console output in the log file rather than the command output
[13:03] <KM0201> ok.
[13:04] <smoser> KM0201, or your client isn't using the right key that you have configured.
[13:04] <KM0201> well, i completley purged ssh on the client, and still had the issue.
[13:04] <smoser> one way or another, 'ssh -v <host>' and logs on the server will tell you more.  the server doesn't tell clients why they cant connect.
[13:05] <smoser> zul, i dont follow.
[13:05] <KM0201> smoser: gotcha, thanks
[13:06] <zul> smoser: well when i run euca-get-console-output it displays nothing on the cli but it spews everything into the /var/log/nova/nova-compute.log
[13:06] <smoser> zul, well, it always does that.
[13:06] <smoser> stupidly, but with debug on user-data gets written to that log.
[13:06] <zul> smoser: anyways ill figure it out
[13:07] <smoser> er... s/user-data/console-output/
[13:10] <zul> smoser: this is what i mean: http://pastebin.ubuntu.com/699744/
[13:11] <smoser> it writes the word logfile.
[13:11] <smoser> nice
[13:11] <zul> anyways ill figure it out ;)
[13:15] <janimo> zul, yes as in DM/RAID needed on ? Any other configs?
[13:16] <zul> janimo: yes as in dm/raid needed...i would have a look at the linux-server config
[13:21] <qman__> I'm having a DNS problem. I'm trying to decommission one DNS server, and move a bunch of domains to the new one. The problem is, even though I have changed the IPs of my nameserver records, people are still asking the old one, weeks later. What am I missing? How can I find out?
[13:25] <pmatulis> qman__: dhcpd not updated?
[13:25] <patdk-wk> qman__, did you update your nameservers with your register for all those domains?
[13:25] <rbasak> qman__: you may need to change the zone on your old server to respond with the new NS records. Also try http://www.squish.net/dnscheck to check you've got everything set up correctly. I'm not sure where the right place to ask though, not sure that this channel is the right place.
[13:26] <patdk-wk> nothing like bad dns glue :)
[13:27] <qman__> I didn't change anything on the registrars, but the domains use 'ns1.mydomain.com' and 'ns2.mydomain.com'. I changed the A records for those to the new IP
[13:27] <patdk-wk> doesn't matter
[13:27] <patdk-wk> you own mydomain.com?
[13:27] <qman__> that's not the actual domain, sorry
[13:27] <qman__> I forgot that's a real site
[13:27] <patdk-wk> don't use other peoples domains as your own
[13:28] <qman__> what I meant was literally my domain
[13:28] <patdk-wk> doesn't matter
[13:28] <patdk-wk> glue has to exist somewhere
[13:28] <patdk-wk> and unless the glue is updated, it won't go to your dns servers
[13:29] <patdk-wk> everytime you change a dns server ip
[13:29] <patdk-wk> you have to login to the register of that domain and update the glue
[13:29] <patdk-wk> for the dns server domain name you use
[13:29] <patdk-wk> and that change can take 48hours (though normally only takes a few)
[13:29] <qman__> ah, ok
[13:30] <qman__> so I was missing a key step, thanks
[13:30] <john> here is a question with ubuntu natty server is there a way to make it auto login so i can have it launch somthing like a game server program
[13:30] <patdk-wk> john, isn't that what init.d? upstart? or even rc.local is for? :)
[13:31] <patdk-wk> hell, you could even use cron @reboot
[13:31] <john> rc.local wont run it
[13:31] <john> it seems like it dont want to work untill after every thing is loaded
[13:31] <patdk-wk> if it won't run it, that program is evil, and probably needs env thing set
[13:31] <slestak> hi guys.  anyone using the tmux package from the repo?
[13:31] <qman__> many game servers will only run interactive
[13:32] <patdk-wk> john, then add a sleep to it
[13:32] <slestak> I am getting a create session error.  Googleing gives me some leads but nothing ubuntu specific
[13:32] <qman__> I ran into this problem before, wrote up an init script which used screen
[13:32] <patdk-wk> qman__, true, but that is a fault of the person attempting to daemonize it, not the app itself really
[13:32] <patdk-wk> I mean, it's evil of the app, but can easily be worked around
[13:33] <john> qman how did you get it to work or did you
[13:33] <qman__> john, created a user to run it as, then put together a couple scripts
[13:33] <patdk-wk> john it probably wants to bind to a pty, and normally that doesn't exist except with a real console
[13:33] <qman__> I can pastebin those scripts for you but I have to look them over first
[13:34]  * patdk-wk remembers pty hell of using ssh from inside a freebsd jail
[13:35] <john> alright qman that could help I'm just trying to get it to be as automated as posible  this is my first server and its a dedicated minecraft  server for a start
[13:39] <qman__> if I run 'dig glue domain.com', and it resolves correctly, does that mean glue records aren't the problem?
[13:39] <patdk-wk> your using other peoples domains again?
[13:39] <patdk-wk> what is the problem ns server?
[13:40] <patdk-wk> and problem domain?
[13:40] <qman__> about 50 domains, the main domain is trsys.com
[13:41] <qman__> I inherited a mess, and I've never done this before, the domains are on several different registrars
[13:43] <patdk-wk> I am not getting any glue :(
[13:49] <patdk-wk> strange
[13:49] <qman__> I think I found it
[13:49] <patdk-wk> sometimes I get a 216.x ip, and sometimes a 184.x ip
[13:49] <patdk-wk> from the same dns server
[13:50] <qman__> 184 is the new one, 216 is the old one that's supposed to be going away
[13:50] <patdk-wk> maybe some of the root servers are lagging in their updates
[13:56] <qman__> john, this is the script I made up for srcds:
[13:56] <qman__> http://pastebin.com/SM4KJTdV
[13:56] <lynxman> morning everyone o/
[13:58] <qman__> some details you'll need, srcds is the user it runs as, ryan is my user so I can screen, and /home/srcds/srcds.sh is a script which runs the game server itself
[13:58] <iuytfr> hellot
[13:59] <iuytfr> anyone there
[14:01] <qman__> !anyone | iuytfr
[14:02] <iuytfr> who know if there is someone there ?
[14:03] <iuytfr> well
[14:04] <iuytfr> how to perform a complete hard disk check up ?
[14:05] <qman__> iuytfr, you can monitor hard disk SMART information with smartctl, part of the smartmontools package
[14:09] <zul> rbasak: whats the difference between the two patches?
[14:10] <reya276> How can I fix this issue http://pastebin.com/Z5pDBGNS ?
[14:10] <rbasak> zul: I've added tests, fixed some existing test failures (including a bug) and added restart of logging on nova-compute restart
[14:11] <zul> k
[14:11] <rbasak> Oh, and i8n in the RingBuffer error messages
[14:12] <iuytfr> i just installed smartmontools
[14:12] <iuytfr> what i have to do now ?
[14:14] <qman__> iuytfr, sudo smartctl -a /dev/sd?   where sd? is your drive name
[14:14] <qman__> read the manual for more options and details
[14:14] <reya276> How can I fix this error http://pastebin.com/Z5pDBGNS ?
[14:15] <reya276> for some reason mysql wont start
[14:15] <reya276> in the mysql channel I was told it was an Upstart issue
[14:18] <koolhead17> kim0: around?
[14:20] <kim0> koolhead17: yep
[14:26] <reya276> ok I figured out the cause of the error, is not upstart http://pastebin.com/xfpzJkh7
[14:29] <ersi> reya276: most common thing to forget when moving stuff from the default locations ;) (I've done that as well)
[14:33] <reya276> ersi, what is the thing I forgot?
[14:35] <reya276> ersi, I doubt that this is the issue, right now mysql wont start because it can't write to the new directories and I have chown it already but nothing
[14:38] <ersi> permissions
[14:39] <ersi> Your permissions are still wrong, or something else is prohibiting it from writing there
[14:39] <ersi> ls -l / and check the file permissions on /Data
[14:58] <CallidusDragon> ahh this is the one I wanted
[14:59] <CallidusDragon> Hello. is anyone around?
[14:59] <ikonia> yes
[15:00] <CallidusDragon> Im having some troubles with my home server. Cant get use to using ssh
[15:01] <CallidusDragon> its not that im command line inept, its simply that im slow as hell using it. I also have no clue to do basic things like download files (or transfer from my home comp for that matter)
[15:02] <ikonia> so what's your actual question ?
[15:04] <KM0201> probably wants a webui
[15:04] <CallidusDragon> would be usefull
[15:04] <KM0201> CallidusDragon: how often do you need to ssh a home server though?
[15:05] <CallidusDragon> .... multiple locations
[15:05] <KM0201> once it's all set up and configured, you shouldn't really ever need to
[15:05] <CallidusDragon> no direct hardware access
[15:05] <CallidusDragon> and im on my setup phase
[15:05] <ikonia> CallidusDragon: do you have a question, yes or no
[15:05] <CallidusDragon> Well considering the response I am getting, no.
[15:06] <KM0201> CallidusDragon: so your problem is, you type slow... and this is taking forever, or what
[15:06] <ikonia> CallidusDragon: you are rambling in #ubuntu and you're not really asking a question in here, it's quite frustrating as people give their time to help
[15:06] <ikonia> if you actually need help - ASK a question detailing your problem
[15:06] <CallidusDragon> I understand, however I am trying to form a question. And i dont see anyone else begging for help.
[15:07] <CallidusDragon> I have to configure my server to use ftp and its frustrating
[15:07] <ikonia> CallidusDragon: so form the question
[15:07] <CallidusDragon> id much rather have a gui and use other tools.
[15:07] <ikonia> CallidusDragon: ok, so you want to use FTP.
[15:07] <CallidusDragon> sftp in particular
[15:07] <KM0201> CallidusDragon: vsftpd is very easy to set up
[15:07] <ikonia> CallidusDragon: have you installed an ftp daemon yet ?
[15:07] <ikonia> ok, sftp is part of ssh
[15:07] <CallidusDragon> Define easy.
[15:07] <ikonia> CallidusDragon: have you fot ssh installed ?
[15:07] <ikonia> CallidusDragon: you don't want an ftp daemon
[15:08] <ikonia> you want to use sftp as you've said which is part of ssh
[15:08] <CallidusDragon> COnsidering i dont know what fot ssh is, no.
[15:08] <ikonia> have you got ssh installed on the server, yes/no
[15:08] <CallidusDragon> I believe so.
[15:08] <ikonia> CallidusDragon: ok, do you know how to install software from the package manager ?
[15:08] <CallidusDragon> yes
[15:08] <jpds> CallidusDragon: So you have sftp already set up.
[15:08] <CallidusDragon> then what is the command?
[15:08] <ikonia> CallidusDragon: ok, install the package "openssh-server" just to be sure
[15:09] <CallidusDragon> ive been using my ssh connection through putty for a week now.
[15:09] <ikonia> CallidusDragon: let me know when you've installed it/confirmed it's installed
[15:09] <CallidusDragon> I have an ssh server
[15:09] <ikonia> CallidusDragon: ok, so you should be able to sftp to that machine
[15:09] <CallidusDragon> I dont know how is the problem
[15:09] <ikonia> jpds: is the sftp line commented out by dwefadefault ?
[15:09] <ikonia> default
[15:09] <jpds> ikonia: No.
[15:09] <ikonia> (I don't have an ubuntu server to confirm)
[15:09] <ikonia> CallidusDragon: ok, so what's the issue when you try to sftp
[15:10] <ikonia> CallidusDragon: do you get an error/warning/issue
[15:10] <CallidusDragon> pull it back. How do enable an sftp session?
[15:10] <ikonia> CallidusDragon: it's already enabled by default
[15:10] <jpds> CallidusDragon: "sftp $IPADDRESS" is all I have to do.
[15:10] <ikonia> CallidusDragon: hence why we are asking you try it and report back
[15:10] <bastidrazor> sftp and scp are the same?
[15:10] <CallidusDragon> Scp was annoying as hell
[15:10] <CallidusDragon> i had to transfer 12 files
[15:10] <ikonia> CallidusDragon: focus
[15:10] <ikonia> FOCUS
[15:11] <ikonia> CallidusDragon: can you please try sftp and report back
[15:11] <CallidusDragon> ....
[15:11] <CallidusDragon> I dont know the syntax to use the command
[15:11] <ikonia> CallidusDragon: sftp ip.address
[15:11] <ikonia> as jpds suggested
[15:11] <ikonia> CallidusDragon: I believe putty installs an sftp client by default
[15:12] <ikonia> CallidusDragon: it's called "psftp"
[15:12] <ikonia> CallidusDragon: it's in the same location as your putty command
[15:12] <CallidusDragon> one second
[15:12] <CallidusDragon> I dont use putty as a command
[15:12] <CallidusDragon> im on a windows machine.
[15:12] <ikonia> putty IS a command
[15:12] <ikonia> it's in C:\program files\Putty
[15:12] <CallidusDragon> putty is a .exe
[15:12] <idlemind324> ahh run he said the naughty word c:\
[15:13] <ikonia> CallidusDragon: correct,
[15:13] <ikonia> CallidusDragon: so is psftp
[15:13] <ikonia> CallidusDragon: it's in the same pyscial location as putty.exe
[15:13] <rbasak> zul, smoser, soren_: yeah, if nova-compute isn't running when a guest dumps more than about 64k to /dev/console, the guest freezes and when I restart nova-compute it can't seem to do anything. That's pretty bad.
[15:13]  * CallidusDragon sighs
[15:13] <CallidusDragon> now i need to remove the rsa key fingerprint for the server. I tried sftping into itself
[15:13] <CallidusDragon> great.
[15:13] <ikonia> CallidusDragon: focus,
[15:13] <bastidrazor> there is 'winscp' that allows graphical copy paste to a server.
[15:13] <ikonia> CallidusDragon: no you don't
[15:13] <ikonia> CallidusDragon: just run the command psftp $ip-address
[15:14] <CallidusDragon> I am attempting to locate putty
[15:14] <ikonia> it's in c:\program files\putty
[15:14] <KM0201> if this is such an issue, why not just install an FTP client, and use filezilla to download/install files?
[15:15] <KM0201> sorry, download/upload
[15:15] <ikonia> KM0201: please - one thing at a time
[15:15] <CallidusDragon> it is not btw.
[15:15] <KM0201> ikonia: i understand, just syaing, if his issue is typing commands, then help him get around his issue... i dunno.. just doesn't make any sense how this is being handled
[15:16] <ikonia> CallidusDragon: then find it and come back to us
[15:16] <CallidusDragon> that is what im doing
[15:18] <zul> rbasak: so no to your patch?
[15:19] <rbasak> zul: well it's a trade off - that's the catch with my patch but I don't see a better way. The patch is better than the previous vulnerability IMO, given that nova-compute will generally stay up and restarts will be quick
[15:20] <zul> rbasak: ok ill leave it as it is
[15:20] <rbasak> zul: other options are the ones I said in the gerrit comment - but none are that simple. I suppose the simplest workaround would be to spawn off processes instead of threads
[15:20] <rbasak> (so they can outlive nova-compute)
[15:20] <rbasak> Daviey: ^^ - any comment?
[15:31] <Daviey> rbasak: give me 20 mins?
[15:31] <rbasak> np
[15:36] <KM0201> what's a good music player that can play a playlist from my server?
[15:36] <KM0201> movies, i just play them one by one w/ vlc.. music that won't work
[15:36] <CallidusDragon> What is the default install path for software?
[15:38] <CallidusDragon> ikonia im sorry you were less than helpfull
[15:38] <CallidusDragon> My issue was solved
[15:38] <CallidusDragon> it created a new one in the process im attemting to fix
[15:38] <ikonia> thats fine. Good luck then
[15:42] <Daviey> rbasak: Spawning off processes from threads, seems kinda scary in this case, no?
[15:43] <rbasak> Daviey: I would drop the thread use, spawn off one process for each machine whose responsibility it is to create a pid file in the instance directory and then do the fifo reading/ringbuffer writing. The nova-compute process would be able to read directly from the ringbuffer fine (perhaps with an flock). It would be able to kill the reader/writer process by looking for the pid file.
[15:44] <rbasak> On nova-compute restart it wouldn't have to do anything
[15:45] <rbasak> Perhaps it should check for pid files and spawn processes for missing ones though
[15:50] <Daviey> rbasak: Catching up on the comments
[15:52] <Daviey> rbasak: I'd quite like soren's opinion TBH.
[15:52] <Daviey> Considering he has been following it from the start.
[15:53] <Daviey> I'm not convinced too many filehandles is a concern TBH.
[15:53] <rbasak> which one's soren?
[15:56] <Daviey> rbasak: soren_ is soren_
[15:56] <Daviey> He hasn't commented on the review yet
[15:56] <rbasak> which opinion are you referring to?
[15:56] <rbasak> Oh
[15:56] <rbasak> You'd like his opinion, OK.
[15:56]  * rbasak should learn to read
[15:57] <Daviey> :)
[15:58] <Daviey> zul: In bug 529714 you dropped a patch introduced by us as it seemed to be included upstream.  Seems not, can you investigate as a priority?
[16:18] <philipballew> QUESTION: Can I set up a VPN server if I have a really crappy router
[16:22] <qman__> philipballew, a VPN server can be set up as long as you can port forward (in the case of openvpn, udp 1194 or alternate)
[16:25] <philipballew> alright. yeah, I have my server in a place where the router is the default att router qman__
[16:26] <cjs226> anyone using reiserfs4?  i ask as I'm looking for a filesystem that allows for filenames longer than 512bytes/characters
[16:29] <smoser> rbasak, you can probably reproduce that behavior in kvm, right?
[16:29] <qman__> cjs226, last time I heard anything about reiser4, it was still hopelessly destroying data
[16:29] <qman__> things could have changed since but I haven't heard about it
[16:29] <RiotCannon_Work> Anyone able to help me configure a mail relay on ubuntu server? I have been trying to get nullmailer to work for my nagios installation, only problem is no mail reaches my SMTP server
[16:29] <cjs226> qman_: ouch.
[16:30] <qman__> RiotCannon_Work, I recommend using postfix, because it's easy
[16:30] <qman__> just set it up as a satellite system
[16:31] <qman__> debconf walks you through it
[16:32] <rbasak> smoser: kvm on its own without libvirt or nova? I expect so.
[16:32] <rbasak> smoser: (after all I'm stopping nova-compute at that stage)
[16:33] <smoser> right. i'm just saying, for your testing you could cut out nova-compute.
[16:34] <smoser> to verify, just launch kvm guest, run dd if=/dev/zero of=/dev/ttyS0 and you're asserting it would lock up.
[16:35] <www2> hi i hef a problem with my apache server (httpd) and ipv6
[16:37] <RiotCannon_Work> qman__ to test if postfix is working, what to I do?
[16:37] <www2> the problem is that i run vew domines and if i login https://b.domine.org i get the cert of http://a.domine.org
[16:38] <qman__> RiotCannon_Work, telnet localhost 25 and write yourself a mail
[16:38] <www2> note this work corext in ipv4 but not in ipv6
[16:57] <topochan> bye people
[17:02] <zul> Daviey: yeahp
[17:06] <adam_g> zul: regarding glance, https://bugs.launchpad.net/ubuntu/oneiric/+source/glance/+bug/862844
[17:07] <adam_g> not sure which is preferred, http://paste.ubuntu.com/699911/ or http://paste.ubuntu.com/699916/
[17:09] <zul> adam_g: you tested with that postinst?
[17:11] <adam_g> zul: yeah, both do the job but im not sure which method is preferred (just 0700'ing the whole directory or 0600 the individual files installed via .install)
[17:11] <zul> adam_g: i would say the 0700
[17:15] <adam_g> zul: hmm ok. ill put a proposal together in a min. as for the nova logfile_mode thing (bug #862816), should we just do the same to /var/log/nova/ until actually using the flag to set the modes gets fixed upstream? ive submitted a patch that is pending, but don't want to carry that unless it gets accepted
[17:15] <zul> adam_g: yes
[17:16] <zul> adam_g: can you check to see if logrotate wont screw with the permissions as well
[17:18] <adam_g> zul: ah, yeah ill test in a few
[17:18] <smoser> is here known data in /var/log/nova that is sensitive ?
[17:19] <smoser> i know of console-output being logged there, which is potentially moderately sensitive, but anything else?
[17:20] <soren> smoser: Yeah.
[17:20] <soren> smoser: sql passwords.
[17:20] <smoser> well, of course i assumed we were logging passwords
[17:21] <smoser> i meant anything *else*!
[17:21] <smoser> :)
[17:21] <smoser> i'd say thats important to fix then.
[17:21] <soren> Yeah.
[17:21] <soren> We don't currently have a "sensitive" flag type.
[17:26] <zul> soren: no but there is a flag to override the default log file permissions
[17:28] <hallyn_> zul: trying to reproduce our qcow-ext4 woes with openstack-lxc using a test script.  http://people.canonical.com/~serge/nbd-ext4-qcow-test.bash doesn't give me quite the same results, but it did screw up my laptop's kernel (hanging with sb_sem down or something - can't umount)
[17:28] <soren> zul: Yeah.
[17:29] <zul> hallyn_: how about running an instance openstack-lxc and then umount the /var/lib/nova/instances/instances-xxxx/rootfs
[17:29] <hallyn_> zul: i want to reproduce it without openstack
[17:30] <hallyn_> zul: though i think it requires snapshotted qcow2 to really break.  So I'm thinking qemu-nbd itself is the problem.
[17:30] <hallyn_> still testing
[17:30] <zul> k
[17:30] <hallyn_> (so kernel is probably waiting on a hung qemu-nbd task)
[17:31] <hallyn_> zul: smoser may not know it, but he has the problem too :)  i've seen it in his syslog :)
[17:33] <smoser> hallyn_, i never saw that.
[17:33] <smoser> hallyn_, do you have a bug opne on this ?
[17:35] <hallyn_> smoser: not yet
[17:35] <zul> Daviey: just doing a test build now (re: samba)
[17:35] <hallyn_> smoser: you never saw it, but it was on that node you gave me access to yeterdsay
[17:36] <hallyn_> smoser: i needed a better testcase before i could report a bug
[17:38] <RoyK> hrmf. just updated my home server to 11.04 and now it doesn't join the wifi
[17:41] <smoser> hallyn_, i dmesged and didn't see it.
[17:41] <Daviey> zul: cool
[17:42] <hallyn_> smoser: dmesg | grep 'EXT4' showed no errors about jounral or bh_sumit or anything?
[17:42] <smoser> i didnt' see it.
[17:42] <hallyn_> huh
[17:42] <hallyn_> ok.  i guess i lied
[17:42] <smoser> i looked, but maybe i looked in the worng place
[17:42] <smoser> who knows/
[17:43] <smoser> hallyn_, qemu-nbd is only in the disk write path for instances if instances are lxc
[17:44] <smoser> that said, i'm assuming qemu-nbd is basically the same code path as would be taken for the qcow backed driver in kvm
[17:44] <smoser> thats kind of its point in life.
[17:44] <smoser> so a bug in one would be probably in the other.
[17:44] <hallyn_> smoser: i've not had a kvm based node in months
[17:45] <hallyn_> of course i'm vaguely recalling we had a 'fix' for a nbd driver bug that noone really understood...
[17:45] <adam_g> zul: logrotate cronjob works just fine after chmod 0700 /var/log/nova/
[17:48] <zul> adam_g: cool
[17:54] <idlemind324> i'm using iptables -t nat -A PREROUTING -i eth0 -d x.x.x.x -p tcp --dport 22 -j REDIRECT --to-destination y.y.y.y:22 ... where x is external and y is internal
[17:54] <idlemind324> i get error "iptables v1.4.4: unkown option '--to-destination' but when i look in the man page it is there and is to be used in PREROUTING
[17:57] <cjs226> anyone using zfs?
[17:58] <ikonia> not really as it's not available on linux
[17:59] <TheEvilPhoenix> ^
[18:00] <hallyn_> heh, forgot a crucial '&' in the uploaded version
[18:06] <DrNick__> why would you need zfs when you have LVM2
[18:35] <zul> adam_g:  any closer to getting those fixes in?
[18:38] <adam_g> zul: yes, trying to submit proposals but launchpad is not working :|
[18:38] <adam_g> timeouts when i try submitting them
[18:38] <zul> bah
[18:38] <adam_g> =/win 24
[18:39] <adam_g> "Trying again in a couple of minutes might work."
[18:41] <zul> adam_g: "we are sorry we are unable to complete the call as dialed please try again later"
[18:41] <SpamapS> DrNick__: ZFS and LVM2 aren't even in the same galaxy ;)
[18:41] <DrNick__> yeah, lvm2 is much better :p
[18:42] <SpamapS> at being less capable, yes.
[18:42] <SpamapS> don't get me wrong.. I lurve lvm
[18:43] <SpamapS> just.. ZFS has bits in it that LVM just can't have, or probably won't ever have.
[18:43] <SpamapS> btrfs is a much better comparison.. but.. who knows when we'll get that
[18:43] <DrNick__> interesting, such as?
[18:45] <SpamapS> data checksumming for one
[18:45] <SpamapS> http://blogs.oracle.com/bonwick/entry/zfs_end_to_end_data
[18:46] <SpamapS> De-duplication is another nice one
[18:46] <SpamapS> http://blogs.oracle.com/bonwick/entry/zfs_dedup
[18:47] <DrNick__> both nice to have features, but i suppose with pretty specific applications
[18:50] <SpamapS> DrNick__: just massive scale. :)
[18:51] <SpamapS> DrNick__: the data integrity isn't exactly a specific application thing. It is intended to prevent those "WTF" errors that people never even know about.
[18:51] <DrNick__> yeah, would that have a performance impact tho?
[18:52] <adam_g> SpamapS: http://oss.oracle.com/projects/data-integrity/  looks like LVM/DM could be expanded to take advantage
[18:52] <nowashburn> anyone familiar with vsftpd?
[18:52] <DrNick__> yeah fairly familiar
[18:54] <nowashburn> awesome, first time setting up vsftpd (or any ftp daemon at that matter) and was wondering what the difference was between setting umask and file_open_mode
[18:55] <mdlueck> I have been trying to track down why 10.04 LTS servers do not automatically select kernel updates, but 10.04 desktop machines do? http://ubuntuforums.org/showthread.php?t=1807248
[18:56] <rich_> Need advice on location of lucid JeOS suite for vmbuilder use.  Anyone?   I'm getting : Invalid suite. Valid suites are: dapper gutsy hardy intrepid jaunty karmic
[18:57] <mdlueck> @rich_ Those distros are pretty old, maybe JeOS has been abandoned
[18:57] <ikonia> mdlueck: dead for ages
[18:58] <DrNick__> i've never needed to set file_open_mode, just changed the default umask for users to 022, or 077
[18:59] <nowashburn> yeah, ill be the only one using the ftp account and wanted default 755 permissions, i couldnt seem to do this with umask and needed file_open_mode. does that sound normal?
[18:59] <zul> adam_g: just nova and glance right? (im reviewing swift right now)
[18:59] <mdlueck> Anyone with feedback on if Ubuntu Server should auto-select / auto-detect kernel updates or not?
[18:59] <adam_g> zul: yea
[19:01] <DrNick__> a umask of 022 should give you what you want
[19:01] <mdlueck> Initially when 10.04 came out I was getting notified (apticron), but for the past several kernel updates I must manually select the new kernel. How can I get automatic back?
[19:02] <nowashburn> no matter what i set the umask, the default file creation would not include execute permissions
[19:04] <DrNick__> hm strange. did you restart vsftpd after making modifications
[19:05] <nowashburn> other than that, the only other thing im struggling with is default directory permissions. when the user i created creates a directory, the directory has no default permissions. must be a way to at least make directories writeable upon creation?
[19:05] <nowashburn> yes, i did restart
[19:06] <zul> adam_g: swift taken care of...debdiffs?
[19:07] <DrNick__> try setting file_open_mode to 0777
[19:09] <adam_g> zul:  generating now
[19:10] <nowashburn> changed file_open_mode to 0777, file upload is fine but still the issue with creating directories
[19:14] <philipballew> QUESTION: can i set up a way to get alerted if my server goes down or something happens to it
[19:15] <adam_g> zul: attached to both bugs
[19:15] <zul> bug numbers?
[19:15] <adam_g> zul:  bug #862816 and bug #  #862844
[19:16] <adam_g> zul: or if you want to just 'bzr merge' the linked branches, i guess that would work too?
[19:17] <zul> it should im not going to risk it right now though
[19:17] <DrNick__> what permissions are set on the directory?
[19:26] <zul> adam_g: applied the patches thanks
[19:27] <Daviey> zul: are you uploading today?
[19:27] <zul> Daviey: yeah
[19:28] <nowashburn> looking at the directory in zilezilla after creation it just says xxx. i guess that means nothing is set?
[19:28] <zul> swift is already uploaded
[19:29] <idlemind324> so say i wrote a handy file named firewall.sh that has all the rules i want setup for my firewall. where is the best place to put that for it to run during boot-up?
[19:36] <DrNick__> can you double-click it in filezilla to open it
[19:42] <Daviey> adam_g: have you come across http://dep.debian.net/deps/dep3/ ?
[19:48] <adam_g> Daviey: doh. ive read it, but i guess i should freshen up next time huh?
[19:58] <idlemind324> hmm firefox is making me angry
[19:58] <nowashburn> DrNick__: I cant open the folder, but i do have permissions to change the permissions
[19:58] <idlemind324> ie 9 displays my xml output with it's xslt
[19:58] <idlemind324> firefox does not
[19:58] <idlemind324> this has to be a first
[19:59] <DrNick__> what's the output of ls -l on the host
[20:00] <nowashburn> total 4
[20:00] <nowashburn> -rw-r--r-- 1 root root 1320 2002-12-05 00:58 jcameron-key.asc
[20:05] <zul> Daviey: nova uploaded
[20:07] <zul> Daviey: glance uploaded as well
[20:16] <Daviey> zul: super!
[20:28] <SpamapS> adam_g: definitely LVM could do that.. but I doubt it will be since BTRFS is slated to supersede the need.. I think. :)
[20:55] <hallyn_> smoser: i think nova's nova/virt/disk.py:_unlink_device() perhaps should first check if the device is mounted?
[20:55] <hallyn_> rather than pursuing a questionable kernel fix...
[20:56] <hallyn_> i suppose one could argue qemu-nbd should be doing that check
[20:58] <smoser> nova is the easiest place to fix that.
[20:58] <smoser> and its not going to be harmful
[20:58] <Daviey> Doctor, Doctor, my arm hurts when i twist it in a funny angle.
[20:59] <soren> nnnnnggggghhhhh....
[20:59]  * soren stabs user-mode-linux
[21:00] <hallyn_> Daviey: i'm guessing there are several ongoing conversations your comment could apply to?
[21:00] <hallyn_> smoser: all right... i'll do it in nova.
[21:01] <hallyn_> i get to target that bug at a 3d package, yay :)
[21:02] <Daviey> hallyn_: :)
[21:02] <smoser> hallyn_, you think just search through /proc/mounts for '/dev/nbdX' ?
[21:03] <Daviey> I wonder if we'll exceed 20 nova patches by release.
[21:03] <smoser> or /proc/mountinfo might be better.
[21:04] <hallyn_> smoser: yes
[21:10] <hallyn_> smoser: you know i might not even need to do that check.  it might just help if nova actually checked the return value of umount.
[21:13] <soren> hallyn_: Doesn't it=
[21:13] <soren> ?
[21:13] <hallyn_> soren: ah, no - i suspect utils.py will raise an exception
[21:14] <hallyn_> soren: that's why i only see it when i do euca-terminate-instances the *second* time :)
[21:14] <hallyn_> so it needs the check in _unlink
[21:15] <hallyn_> _unlink_device that is
[21:27] <Daviey> RoAkSoAx: around?
[21:28] <Daviey> adam_g / hggdh: You'll be happy to know the cobbler-enlist tool works from the cd menu.
[21:28] <adam_g> Daviey: nice!
[21:28] <hggdh> Daviey: \o/
[21:29] <hggdh> good work you both :-)
[21:29] <Daviey> hggdh: and you.
[21:29] <Daviey> adam_g: I haven't tried preseeding it as an option yet.
[21:29] <Daviey> I'm kinda worried where the hostname can originate from..
[21:30] <Daviey> I'm wondering if cobbler-enlist should generate a name based on MM-AA-CC-AA-DD-RRess if a null value is provided
[21:33] <robbiew> Daviey: that would "seem safe"...maybe with "ubuntuserver-" prepended
[21:33] <adam_g> hmm
[21:34] <hallyn_> zul: yay, no more ext4 errors
[21:34] <hallyn_> lemme just keep pounding this , uh, fine lady
[21:35] <Daviey> robbiew: node10.orchestra.target.waiting.for.juju.love.local
[21:35] <robbiew> umm...no
[21:35] <robbiew> lol
[21:36] <SpamapS> we all need some juju love
[21:36] <robbiew> maybe prepend orchestra-
[21:37] <SpamapS> I like just  ubuntu-XX-XX-XX-XX-XX-XX
[21:37] <robbiew> works too
[21:37] <SpamapS> Its a reasonable length, and pimps us without being conspicuous
[21:37] <robbiew> or...not allow null value for hostname
[21:37] <Daviey> it's not known that the target will be ubuntu :)
[21:38] <SpamapS> Daviey: even better, that will make them feel even sillier for not installing Ubuntu with it. ;)
[21:38] <Daviey> robbiew: Well, in order to automate this - it needs to generate a hostname on it's own, i think
[21:38] <Daviey> SpamapS: hah
[21:38] <robbiew> then ubuntu- it is
[21:38] <robbiew>  I agree with SpamapS :P
[21:38] <robbiew> they can always change it...the OS of course, not the name
[21:39]  * robbiew goes to workout....Insanity!!!!!!! \o/
[21:39] <SpamapS> Daviey: does this trigger from the PXE boot as well?
[21:40] <Daviey> SpamapS: That is the /plan/, but it's awaiting a preseed.
[21:40] <SpamapS> cause even w/o that.. its useful .. admins can carry around a usb key that they boot from to make this come up
[21:41] <Daviey> SpamapS: adding it to the ISO was for 2 reasons, try-before-you-commit usage & adds the ability to preseed for free (main target)
[21:41] <SpamapS> An OEM could even ship it as the default image on disk.. :)
[21:42] <Daviey> hah, RoAkSoAx was looking at creating a preseed for cobbler to ship, i suspect he didn't have time this week to get to that
[23:13] <sarthor> I Installed apache2 on my ubuntu server, with apt-get -y install apache2, then i put index.html file in /var/www/ index.html file have some arabic words, .. I can not see that arabic words when i am browsing the site... how to solve this problem, to be able to view all the arabic fonts.
[23:18] <idlemind324> sarthor what character set is the document saved as? is the html file DOCTYPE declared with that encoding as well? does your browser have those fonts available?
[23:45] <sarthor> I am new to linux. httpd room asked "find out what your distro renamed the httpd binary to, then.", I am using ubuntu-server 11.04, So what will be my reply there in httpd room?
[23:45] <SpamapS> sarthor: apache2
[23:46] <TheEvilPhoenix> ^