[00:01] <pyro1> alguem fala portugues
[00:01] <pyro1> ??
[01:27] <twb> Bah, screen can do everything minicom can do.
[05:17] <wizardslovak> does anyone uses webalizer?
[05:21] <wizardslovak> is there any software to check traffic, and more?
[06:53] <_bubsy> how to list installed apt-get package ?
[06:56] <_bubsy> ahh.. dpkg
[10:33] <_ruben> hmm .. does anyone know what the "p 0" part of tcpdump output means when showing tagged vlan traffic ?
[10:33] <_ruben> 11:28:17.547764 vlan 202, p 0, IP6 (hlim 255....
[11:37] <ghatak> Hi, i want to use ufw, however the only connection I got to server is via ssh, so to be safe I want to make sure when I enable ufw, it will kill my existing connection, how can I ensure that, i-e add rules somewhere that ufw will read when enabled?
[11:41] <ghatak> ok got it, /etc/ufw/before.rules
[11:45] <alex_joni> ghatak: when doing something like this I disable ufw by default on boot
[11:45] <alex_joni> then issue a reboot 20 before doing any changes
[11:45] <alex_joni> if all fails, the machine reboots and you can ssh again
[11:46] <ghatak> alex_joni: damn,
[11:46] <ghatak> alex_joni: i lost ssh
[11:46] <ghatak> alex_joni: but it clearly says in the rules file that add any custom rules in that file
[11:47] <alex_joni> ghatak: sorry.. never used ufw myself
[11:47] <alex_joni> I write my iptables script by hand
[11:49] <ghatak> i will have to call someone in the datacenter now to flush it :)
[11:55] <ghatak> -A ufw-before-input -p tcp --dport 22 -m state --state NEW,ESTABLISHED,RELATED -j ACCEPT
[11:56] <ghatak> what is wrong with this rule, ufw does not seem to load it
[11:59] <gamla_kossan> hi people. when wanting to restart dovecot, I get this:
[12:00] <gamla_kossan> http://fpaste.org/XHgn/
[12:00] <gamla_kossan> isn't this very odd?
[12:01] <gamla_kossan> /usr/lib/dovecot does not contain any "imap" file or folder though
[12:02] <gamla_kossan> maybe it isn't enough to just install the dovecot-common package?
[12:03] <gamla_kossan> yeah, that seems to be it =)
[12:50] <mrchrisadams> hi guys, if you run a basic 'aptitude install munin munin-node'
[12:50] <mrchrisadams> why would graphs and stats not be visible?
[12:51] <mrchrisadams> munin-node has been running on the box since last night
[13:25] <benc> I'm trying to understand vhosts and DNS
[13:26] <benc> can I use vhosts internally without configuring them in the DNS?
[13:27] <benc> nginx will listen to port 80 for http://example.com/vhost1/some/path and rewrite it to apache to port 800 http://vhost1.example.com/some/path
[13:38] <ghatak> ok having hard time with ufw, it is supposed to be simple hehe
[13:39] <ghatak> when I enable it, i loose all dns, applications can not lookup anymore
[13:39] <ghatak> the syntax is not very helpful either, can't seem to have it enabled easily
[13:39] <ghatak> help ?
[13:41] <pmatulis> ghatak: did you read the man page?
[13:42] <ghatak> pmatulis: i believe so
[13:46] <ghatak> ufw allow to any port 53 out
[13:46] <ghatak> this does not work
[13:47] <pmatulis> ghatak: do you have to specify the protocol?
[13:47] <jdstrand> ghatak: you put 'out' in the wrong place
[13:47] <trothigar> hi, I can't find in the man page of rsync what the option is to delete files from the target which don't exist on the source anymore?
[13:48] <jdstrand> ghatak: it should be 'ufw allow out to any port 53'
[13:48] <jdstrand> ghatak: see 'man ufw'
[13:48] <ghatak> jdstrand: ok i am running 8.04, so i think out is not supported on that
[13:48] <aubre_> I'm back from holiday, do the latest eucalyptus packages have working EBS support now? because if so I want to test it
[13:49] <jdstrand> ghatak: egress filtering via the cli command was added in 9.10. you'll need to update /etc/ufw/before.rules to do egress filtering in 8.04 LTS
[13:49] <ghatak> jdstrand: and of course omiiting out from the rule, causes the rule to be added as default inwards
[13:50] <ghatak> jdstrand: k, kewl, can I use application profiles to specify outgoing rules? they are used in 8.04 i beleive
[13:50] <jdstrand> ghatak: application profiles were introduced in 8.10, sorry
[13:51] <ghatak> ghatak: errmm... i must be running 8.10 then
[13:51] <ghatak> let me cheeek
[13:51] <ghatak> jdstrand: yes, i am running 8.10
[13:53] <jdstrand> ghatak: you can use application profiles, but not with egress filtering then. ie, you have to add iptables-restore style commands to /etc/ufw/before.rules to filter egress (outgoing) connections
[13:54] <ghatak> jdstrand: alrighty
[13:54] <ghatak> thanks
[13:57] <jdstrand> ghatak: btw, re ufw/ssh> you can add the rule before enabling it and the connection should not drop (but if for some reason it does, you can reconnect since the rule was loaded). eg:
[13:57] <jdstrand> sudo ufw allow OpenSSH
[13:57] <jdstrand> sudo ufw enable
[13:59] <jdstrand> ghatak: ^ will do what you want without having to tweak /etc/ufw/*rules
[14:01] <ghatak> jdstrand: no that did not work, I added the rule to /etc/ufw/before.rules and it dropped and connection and wont allow me back in, had to ask datacenter guy to drop me console to get it bcak
[14:01] <ghatak> jdstrand: got it sorted now though
[14:02] <jdstrand> ghatak: what I'm saying is that you shouldn't have had to touch before.rules at all. simply add the OpenSSH rule before enabling the firewall. I'm glad you got it sorted
[14:03] <ghatak> jdstrand: yea, thanks mate. Never used ufw before so was bit confused at start :)
[14:06] <Weasel[DK]> can somebody explain this ?
[14:06] <Weasel[DK]> root@linux01:/# ls -lh /var/log/lastlog
[14:06] <Weasel[DK]> -rw-rw-r-- 1 root utmp 554G 2009-10-14 14:30 /var/log/lastlog
[14:06] <Weasel[DK]> root@linux01:/# du -h /var/log/lastlog
[14:06] <Weasel[DK]> 32K     /var/log/lastlog
[14:14] <Weasel[DK]> nevermind... finaly i found out myself
[14:14] <Weasel[DK]> It is a sparse file, so its size
[14:14] <Weasel[DK]>        on the disk is usually much smaller than the one shown by "ls -l"
[14:14] <pmatulis> Weasel[DK]: yes, try 'du -sh --apparent-size /var/log/lastlog'
[14:15] <Weasel[DK]> pmatulis, gives 554G
[14:16] <Weasel[DK]> i wantet to do backup of /var/log/ with tar and was wondering why it took so long...
[14:22] <bogeyd6> !vmware
[14:22] <bogeyd6> !vmtools
[14:22] <bogeyd6> hrmph
[14:25] <th0mz> what is the pb with vmtools ?
[14:26] <bogeyd6> they didnt install by default
[14:26] <bogeyd6> couldnt remember the vmmouse package
[14:29] <ceb_user69842> Hi....anybody help me to install a eucalyptus cloud ?
[14:30] <pmatulis> Weasel[DK]: what backup program did you use?
[14:31] <Weasel[DK]> pmatulis, as i wrote ... tar    i fixed it with --exclude
[14:32] <pmatulis> Weasel[DK]: ok, better way IMO is to use --sparse option
[14:34] <Weasel[DK]> pmatulis, you are right... did't know it. i should pay more attention to whole man page ;)   Thx
[14:42] <Weasel[DK]> pmatulis, seems like tar still thinks it has to read 554Gb.... the output file stoped growing but tar is stille working hard
[15:11] <nijaba> Is there some documentation somewhere on how to connect elasticfox to uec?
[15:25] <aubre_> fyi I apt-get updated all my nodes and everything is working, I can launch m1.smalls now which I could not before, only problem was dhcpd didn't start the way it should have but once it did, I was golden
[15:31] <aubre_> when I deregister images, is there a way to keep them from being listed in the uec web interface? I've deregestered images, and even deleted the images from s3, but they still show up under my images tab with the option to re-enable them, which clutters up the interface.
[15:38] <aubre_> also, if I want to roll my own uec images will I be able to sell them through the store?
[15:43] <nijaba> soren: Is there some documentation somewhere on how to connect elasticfox to uec?
[15:43] <ashnazg> zul: ping
[15:43] <zul> ashnazg: hi
[15:44] <ashnazg> zul: hey dude, the symlink and zlib issues from yesterday still seem to be in the new package
[15:44] <ashnazg> zul: we've sent you new bugs for each
[15:44] <zul> ashnazg: yeah I saw I should get to it this afternoon
[15:45] <ashnazg> zul: cool cool... we're on efnet #karmic-php-pear working on it
[15:45] <zul> ashnazg: cool
[15:51] <smoser> aubre_, your deregistered and deleted images showing up would seem like a bug
[15:55] <ttx> smoser: hey -- could you FixRelease bug 444605 with a comment ?
[15:55] <smoser> hmm.. i thought i did that
[15:57] <ttx> zul: should we consider bug 434723 done ?
[15:57] <smoser> but i will.
[15:57] <zul> ttx: no it still fails
[15:57] <zul> im going to talk to pitti about it
[15:58] <ttx> zul: do that today, please
[16:01] <ttx> Everyone: Server bug day in progress on #ubuntu-bugs
[16:01] <ttx> https://wiki.ubuntu.com/UbuntuBugDay/20091014
[16:02] <smoser> ttx, 444605 is fix-released now
[16:03] <ttx> smoser: great :)
[16:28] <ttx> smoser: any opinion on bug 451358 ?
[16:28] <smoser> umm... sounds like a bug ?
[16:28] <ttx> yes, but where ?
[16:29] <smoser> you provide arch when you register
[16:29] <smoser> so either
[16:29] <smoser> a.) user error
[16:29] <ttx> ah-ah
[16:29] <smoser> b.) UEC losing that
[16:29] <ttx> c.) testcase error
[16:30] <ttx> smoser: explain "you provide arch when you register"
[16:30] <smoser> i dont know what the proper behavior is, but if you tried to register a i386 image as x86_64, on ec2 i think it would just work
[16:31] <ttx> euca-register has no arch parameter ?
[16:31] <smoser> but likely would fail when you try to run it
[16:31] <smoser> in bundle
[16:31] <smoser> you provide arch
[16:31] <smoser> from ec2-bundle-image help:
[16:31] <smoser>     -r, --arch ARCHITECTURE          Specify target architecture. One of ["i386", "x86_64"]
[16:32] <smoser> it is listed as "optional", but i think only because it assumes i386 if you dont tell it
[16:32] <smoser> i dont think it does any heuristics to have sane defaults.
[16:32] <ttx> apparently, it assumes x86_64 here
[16:32] <ttx> smoser: ok, so its the testcase that sucks
[16:33] <smoser> i'll check what the default is for ec2-bundle-image really quick
[16:33] <ttx> smoser: thx for your input
[16:34] <ttx> smoser: API reference says its not an option
[16:34] <smoser> it prompts you
[16:34] <smoser> ec2-bundle-image does
[16:35] <ttx> hah
[16:35] <ttx> so its a euca-bundle-image non-conformity
[16:35] <smoser> and for me, if i just hit return, default would be x86_64. but that is likely (i think i've seen that somewhere) just assuming it based on 'uname -a'
[16:35] <smoser> yes, thats what it does
[16:35] <smoser> verified via:
[16:35] <smoser> $ setarch i386 xc2 bundle-image --destination /tmp/x1 --image /boot/config-2.6.31-13-generic
[16:35] <smoser> using setarch, it says:
[16:36] <smoser> Please specify a value for arch [i386]:
[16:36] <smoser> without it, it defaults to x86_64
[16:44] <ttx> jsalisbury: I opened and closed bug 451358 -- fixed the testcase so that we coirrectly register i386 images
[16:45] <jsalisbury> ttx: thanks.  I'll re-run the test case.
[16:45] <jsalisbury> ttx: I also opened bug 451291.  Want me to close that one?
[16:45] <ttx> jsalisbury: yes, duplicate of mine, please
[16:46] <jsalisbury> ttx: will do
[16:47] <ttx> jsalisbury: note that it should have been opened against ubuntu/eucalyptus, not Eucalyptus upstream
[16:47] <ttx> tat's why I missed it
[16:47] <ttx> jsalisbury: I'm closing it
[16:48] <jsalisbury> ttx: thanks for the info.  I'll be sure to open future bugs in the correct place ;-)
[16:48] <jsalisbury> ttx:  thanks for closing it.
[17:06] <aubre_> Hmm, this new 11.6 inch netbook/laptop has a VT-x processor, albeit 32 bit http://ark.intel.com/Product.aspx?id=42791&code=SU7300 : http://howtohackstuff.com/2009/10/14/acers-116-inch-aspire-timeline-as1810tz-shows-signs-of-a-netbook-identity-crisis-acer/
[17:08] <aubre_> I know people were on the lookout for a cheap machine that you could use to setup a portable UEC configuration, at $600, seems like you could use a few of these to do that cheaply.
[18:07] <bventura> when ssh into my ubuntu server, I get a message on the MOTD that says "
[18:07] <bventura> oops says "6 packages can be updated, 11 updates are security updates"
[18:08] <bventura> but when I run apt-get update and then apt-get upgrade, i still get the same message when I log in the next time
[18:08] <bventura> whats up with that
[18:11] <bventura> OMG i guess I figured it out there's a command to update the MOTD i guess mine not updating automatically
[18:32] <kpettit> What is the easiest way to get a list of software installed on a ubuntu server so I can install the same stuff on a new machine?
[18:32] <kpettit> Something like a "apt-get install" list of software.
[18:38] <kpettit> Ahh, nevermind I found it.  http://ubuntuforums.org/showthread.php?t=261366
[18:41] <aubre_> is EBS functionality available in the latest karmic uec packages? I can do aeuca-describe-volumes and it shows a volume available, but when I try to attach it to an instance via euca-attach-volume the instance can't see it
[18:44] <aubre_> in the sc-stats.log I see Timestamp: 2009-10-14 12:32:12,471 Service: StorageController Version: 1.6-devel Volumes: 1 Space Used: 10737418240
[19:02] <leaf-sheep> ib
[19:02] <leaf-sheep> Hi folks. I have local repo that I'd like to use (to install systems). I used Ubuntu Minimal iso but that won't let me select the mirrors because there are only one on the list -- archives.ubuntu.com ?  Suggestion?
[19:09] <FFForever> ello
[19:09] <FFForever> how do i generate a new public key from a private key?
[19:16] <FFForever> anyone have a system with over a year of uptime
[19:16] <slampoud> ffforever: if you mean an openssh key, then ssh-keygen -e should do it
[19:16] <FFForever> slampoud, awesome D:
[19:16] <FFForever> :D*
[19:19] <SyL> I'm hardcrashing with the kernel 2.6.31-14 in under 3 minutes. any idea what would be causing this?
[19:29] <javaTN> hi
[19:29] <javaTN> anyone here good with apache configuration?
[19:30] <FFForever> javaTN, this is not #httpd and im not that bad but im by far not an expert :D
[19:31] <javaTN> well i want to re-configure my apache configuration so its less jumbled.
[19:31] <javaTN> let me pull up the configuration and upload it to paste bin to better show
[19:34] <javaTN> http://pastebin.com/m41fa0e0a that is my 000-default configuration file
[19:34] <javaTN> there has to be a better way, cleaner way, to get what i want done.
[19:36] <_ruben> over a year uptime? got a few of those
[19:38] <Bilge> How can I fox: kernel: [263151.175638] IPv6 addrconf: prefix with wrong length 56?
[19:38] <Bilge> fix*
[19:39] <_ruben> autoconf only works with a /64
[19:39] <Bilge> I don't know what autoconf is or why it's doing whatever it's doing
[19:39] <Bilge> I haven't configured IPv6 at all and I don't know what's generating the error
[19:40] <_ruben> that's the beauty of ipv6, it doesnt need configuration (on the client side)
[19:40] <Bilge> That doesn't stop it from spewing errors every other minute though
[19:42] <FFForever> _ruben, can i get a /proc/uptime from u?
[19:42] <_ruben> locate the router that's sending the ipv6 router advertisements, tell its owner to fix its configuration
[19:42] <jph_> I am in the middle of a clean install of Ubuntu Server, I have some RAID1 questions mainly about active and spare partitions would anyone be willing to meet me on a private chat and walk me through a few steps i am on a time crunch today. thanks
[19:42] <jjohansen> smoser: I have run through my EC2 test and got logs on all except 1
[19:42] <jph_> or should we just talk here?
[19:43] <smoser> :-(
[19:43] <_ruben> $ uptime ; cat /proc/uptime 20:43:09 up 462 days, 11:09,  2 users,  load average: 0.00, 0.11, 0.08
[19:43] <_ruben> 39956998.58 40422751.91
[19:43] <jjohansen> smoser: that one was on odd ball test, though using new ari and the ami's default kernel
[19:43] <FFForever> thanks :D
[19:43] <smoser> jjohansen, should we try with the beta kernels to see if you are just lucky today ?
[19:43] <smoser> oh...
[19:44] <jjohansen> smoser: I tried all combinations, of ari, aki
[19:44] <smoser> well i would think thats not really valid to use old ramdisk and new kernel. but if you booted, then it actually shouldn't have mattered.
[19:44] <jjohansen> smoser: I would say more I was unlucky yesterday
[19:44] <jjohansen> smoser: yeah, they all booted
[19:44] <smoser> but youthink your results yesterday were valid
[19:44] <_ruben> jph_: talking in here tends to be prefered
[19:44] <smoser> ie, that we've seen valid failure on this new kernel
[19:45] <smoser> but we still havent seen that failure on old beta kernel/beta ramdisk
[19:45] <smoser> you've looked at the changes between them ?
[19:45] <jjohansen> smoser: yep, I repeated them to many times and and used the same commands today, search back through bash history
[19:46] <jjohansen> I have looked and I don't see anything that should be causing this
[19:48] <smoser> so then we have either a.) random timing differences between new kernel and old kernel make us see failure
[19:48] <pyro1> what's the domain NIS ?
[19:48] <jph_> Ok thanks - Disclaimer this is my first solo install of Ubuntu-server AND a RAID.... Question: I am at the screen where it ask me to enter the number of active devices for the RAID1 array... It notes that a spare partition will only be used if a active devices fails.... I only have 2 Hard drives installed (both 100Gb) but I can free up another HD in about 14 days.... Do I need to have a spare?  If so can I lie to the system and say
[19:48] <jph_>  2 active and 1 spare and then just add the spare later? What dose this really mean
[19:49] <smoser> b.) we've just been fat dumb and happy with old beta kernel/beta ari , and we can probably see failure on console logs there too
[19:49] <_ruben> jph_: if you have 2 drives in it now, just go with 2 active and 0 spare .. you can configure the spare later
[19:50] <jph_> Ok but is says "NOTE: this setting cannot be changed later"
[19:50] <jph_> is that not really true
[19:50] <jjohansen> smoser: can we have an option c) blame Amazon :)
[19:50] <_ruben> jph_: not later in the partitioner part in the setup it means most likely
[19:51] <pyro1> what's domain NIS ?
[19:52] <smoser> network information services
[19:52] <smoser> if you dont know, you probalby dont need it
[19:52] <smoser> jjohansen, so what do we do?
[19:52] <jph_> Ok so if i free up or buy another 100Gb HD in the next 14 days or so i can add it as a spare and be good to go?  Also could you give me a short explanation as to what the spare dose? is it what would let me to rebuild after a damaged drive
[19:52] <pyro1> smoser: i need to know for my college
[19:53] <jjohansen> smoser: I am going to try and replicate running in each zone, maybe it is a zone problem
[19:53] <FFForever> is there a command to return just the space of a drive, and also is there a command just to return the used size of a drive
[19:53] <smoser> pyro1, http://en.wikipedia.org/wiki/Network_Information_Service
[19:53] <FFForever> i know df but thats to much to shift through
[19:54] <pyro1> smoser:  i know what's NIS, i want to know about domain NIS
[19:54] <zul> jjohansen: isnt that the default option?
[19:55] <jjohansen> zul: as in blame zones?  Or run instances in all zones
[19:55] <Bilge> _ruben: how can I just disable IPv6?
[19:55] <smoser> FFForever, assuming you mean size of partition, i'd just do 'df /path/to/mount' that just gives you 2 lines of output with pretty much what you want
[19:55] <zul> jjohansen: blame amazon
[19:56] <jjohansen> zul: ah, yeah it definitely is
[19:56] <smoser> pyro1, well i dont know what "domain NIS" is, and google doesn't seem to either. if google doesn't know it it doesnt exist.
[19:57] <_ruben> Bilge: wouldnt know, as im far from fond of disabling ipv6 anywhere
[19:57] <ScottK> Google will know about it shortly since this channel is logged online
[19:57] <pyro1> smoser: this exist because my teacher on college pass me
[19:57] <jph_> any thought?
[19:57] <pyro1> smoser:  but i dont find on google
[20:01] <jph_> also another question, Can I limit the size a directory is allowed to get?
[20:01] <jcastro> ScottK: you up for a plenary this time around?
[20:01] <_ruben> depends on the filesystem used .. dont know if any actually do support directory quotas
[20:01] <jph_> Like /home/user  Can I limit its size to 5Gb
[20:02] <ScottK> jcastro: I didn't think of anything yet.  Let me consider.
[20:02] <jcastro> ool
[20:02] <jph_> I was going to use ext3
[20:02] <jph_> but could change that now if you recommended something better
[20:03] <_ruben> jph_: never bothered with disk quotas myself
[20:04] <jph_> humm ok well thanks
[20:04] <Jagged> jph_: you could try using lvm
[20:08] <jph_> do you have a partitioning schema  you would recommend (I miss typed when i said 100Gb I have 120Gb HDs).... any thoughts on how I should cut that up
[20:09] <Jagged> if you go for lvm, make a 100 mb partition for /boot, then use the remainder for lvm
[20:10] <Jagged> http://en.wikipedia.org/wiki/Logical_Volume_Manager_%28Linux%29
[20:12] <jph_> I was just (still am) reading this http://www.ibm.com/developerworks/library/l-lvm/  is this source just as good, should i read the other?
[20:13] <Jagged> yours is more in-depth
[20:13] <Jagged> the wiki is just an overview of what lvm is
[20:16] <jph_> Ok I skimmed both, and your saying make a 100Mb for /boot and then make the rest of the HD a 119.90Gb lvm partition... Right?
[20:18] <Jagged> jph_: yeah, use the remaining 119 gb for lvm
[20:18] <Jagged> and create lvm logical partitions for you mount points
[20:18] <Jagged> http://tldp.org/HOWTO/LVM-HOWTO/
[20:18] <Jagged> that is extremely verbose on how to do lvm
[20:19] <jph_> and if i mess up guessing how big to make each logical partition i can just resize them?
[20:19] <Jagged> essentially
[20:20] <jph_> I have to make sure today because i will be on the road in 24hours and have to start this development via SSH... so thank you for taking the time to explain things i would normally read myself
[20:20] <Jagged> resizing might require a resize2fs command for ext2/3 filesystems
[20:21] <mushroomtwo> what about ext4?
[20:21] <Jagged> essentially what using lvm does is bypasses the partition table.  Whereas previously you would need to use parted or something similar to resize partitions, you can use lvm tools instead
[20:22] <Jagged> which won't require a reboot to re-read the partition table
[20:22] <jph_> great that sounds like what i need on this project.... ok is your 100mb recommendation considered "over kill or just right or bare min"
[20:22] <Jagged> overkill
[20:22] <Jagged> you can get by with 10-20 mb of /boot if you really want
[20:22] <mushroomtwo> I try to use 100
[20:23] <mushroomtwo> I've actually maxed-out 100MB on /boot before
[20:23] <mushroomtwo> though for Ubuntu, it seems okay.
[20:23] <Jagged> mushroomtwo: I don't have any experience with ext4
[20:23] <mushroomtwo> eh. it'll probably work.
[20:24] <Jagged> it depends on the implementation
[20:24] <Jagged> ext3 is just ext2 with a journal
[20:24] <mushroomtwo> are there any GTK+/Qt apps to handle partitioning/resizing in lvm?
[20:24] <Jagged> mushroomtwo: LVM Manager
[20:25] <jph_> hummmm ok so what if i play it really safe and go with 500Mb.... What file system should that be?
[20:25] <tonyyarusso> mushroomtwo: iirc gparted is supposed to have LVM support either now or very soon.
[20:25] <jph_> ext3?
[20:27] <Jagged> jph_: your boot partition doesn't need to be that big, but ext3 is fine
[20:27] <Jagged> jph_: you just need to stay away from filesystems which the grub version you install may not support (e.g. reiser4, xfs, jfs....)
[20:29] <mushroomtwo> if you use ext3, make sure to add "noatime" to the line in fstab
[20:29] <mushroomtwo> otherwise, use ext2
[20:30] <mushroomtwo> generally, I also include "noauto", because I don't want the partition being mounted unless I'm doing a kernel update.
[20:30] <mushroomtwo> ymmv
[20:30] <jph_> so select 'noatime' and deselect "relatime'
[20:31] <mushroomtwo> yeah.
[20:31] <mushroomtwo> unless you're ultra-paranoid about your system being compromised, it's okay to let the partition be mounted at boot.
[20:32] <mushroomtwo> especially if you have problems remembering to mount /boot before doing an upgrade.
[20:34] <jph_> ok so if i have it 'noatime' i have to mount it each time i update if i have it as 'relatime' then it auto mounts at boot
[20:34] <jph_> would there be any problems if i have to remotely re-boot the system via ssh?
[20:35] <Jagged> no
[20:35] <jcastro> mathiaz: have we ever supported Xen officially in -server?
[20:35] <Jagged> The only time you would have to mount /boot in mushroomtwo's example is when you do a kernel upgrade
[20:35] <mathiaz> jcastro: hm... well - no
[20:36] <mathiaz> jcastro: IIUC in karmic we support officially xen as a guest
[20:36] <Jagged> you can use dpkg to hold back your kernel that way all kernel upgrades are completely manual
[20:36] <jph_> and this is a good idea for system security right?
[20:36] <Jagged> yeah
[20:36] <mathiaz> jcastro: as this is what we run in EC2
[20:36] <jcastro> mathiaz: ok I just think this guy is confused (in the virtualization part) http://ubuntu-user.com/Online/Blogs/Marcel-Gagne-Orbiting-Planet-buntu/Sex-Live-and-Virtualization
[20:36] <mathiaz> jcastro: zul and smoser could probably give you a better answer
[20:37] <jcastro> thanks!
[20:37] <jcastro> sorry, I tend to default to you for everything. :p
[20:37] <zul> jcastro: domU yes dom0 no
[20:39] <mathiaz> jcastro: and domU is supported starting from which release?
[20:39] <jcastro> mathiaz: I have no idea!
[20:39] <mathiaz> jcastro: hm - sorry wrong person
[20:40] <mathiaz> zul: and domU is supported starting from which release?
[20:40] <zul> jaunty i think
[20:40] <jph_> When i do all that guys it makes the remaining 119.5GB 'unusable' i have gone back and tried using ext2 and the same thing
[20:40] <mathiaz> zul: was this because DomU had been merged in mainline?
[20:41] <zul> mathiaz: yep basically you are on your own for dom0
[20:41] <jcastro> is there anyone that can send the guy an email? If he's writing for a magazine with ubuntu in the title we should make an effort to make sure he's got the right info, right?
[20:43] <zul> sure soren since he is the virtualization guy
[20:43] <jcastro> and he's probably asleep, perfect.
[20:44] <zul> now i must really be back later
[20:44] <jph_> I also tried setting up my lvm drive first by starting it at the end hoping i could just create the ext2 or ext3 out of what was left... and this makes the 531.4MB left unusable.....
[20:44] <jph_> remember this is a RAID1
[20:45] <Jagged> okay
[20:45] <Jagged> so, md0 = /boot?
[20:45] <Jagged> or is this fakeraid?
[20:46] <jph_> it is a software RAID setup using the install options
[20:46] <Jagged> okay
[20:46] <Jagged> so, 500 mb partitions on each drive for md0
[20:46] <jph_> i followed this http://knowledge76.com/index.php/Ubuntu_Server_Install_With_Software_RAID
[20:46] <Jagged> make md0 /boot
[20:46] <Jagged> remaining space for md1
[20:47] <Jagged> use md1 as lvm
[20:47] <Jagged> which should make the lvm menu available
[20:48] <jph_> I am not at the command line i am still in the installation wizard that you follow when you drop the Ubuntu-server live disk in
[20:48] <Jagged> yeah
[20:48] <Jagged> so
[20:48] <Jagged> at the partition tool
[20:48] <Jagged> drive 0: partition 1 = 500 mb
[20:48] <Jagged> use as raid
[20:49] <Jagged> drive 0: partition 2 = 119.5 gb
[20:49] <Jagged> use as raid
[20:49] <Jagged> same for drive 1
[20:49] <Jagged> create new raid
[20:49] <jph_> ok sounds like i have to undo the RAID I created... hold on thanks
[20:49] <Jagged> md0 = drive 0 part 1 and drive 1 part 2
[20:49] <smoser> jcastro, zul for the record: http://packages.ubuntu.com/search?keywords=xen&searchon=names&suite=all&section=all
[20:49] <Jagged> erm
[20:49] <FFForever> can someone give me the output of mount | grep -i "/dev/" | grep -v "none" | grep -v "pts" | grep -v "tmpfs"
[20:49] <smoser> we did have packages (universe) for xen dom0 in hardy
[20:49] <Jagged> md0 = drive 0 part 1 and drive 1 part 1
[20:50] <smoser> but seemingly also still have those in xen-hypervisor-3.3
[20:50] <smoser> err... in karmic (xen-hypervisor-3.3)
[20:50] <jph_> create MD Device
[20:50] <jph_> RAID1
[20:50] <Jagged> yep, and use your 500 mb partition on each drive
[20:51] <smoser> afaics there has never been anything in main. but guests on xen (domU) will "just work" because the kernels are pvops enabled
[20:51] <mushroomtwo> FFForever: /dev/sda6 on / type ext4 (rw,errors=remount-ro)
[20:51] <Jagged> jph_: You want to take this to PM?
[20:51] <smoser> maybe i'm missing something though
[20:51] <jph_> sure
[20:51] <jph_> thanks
[20:51] <FFForever> cool :D, it works right :D
[21:15] <mushroomtwo> anyone have a decent howto for openvpn on ubuntu?
[21:26] <bventura> mushroom: I found this to be the best howto for openvpn: http://www.openvpn.net/index.php/open-source/documentation/howto.html#pki .. I didn't need to mod anything for ubuntu it all worked verbatium from the openvpn guide
[21:28] <mushroomtwo> fantastic. the community documentation on ubuntu's site is pretty broken.
[21:29] <bventura> doh
[21:39] <jph_> need some more help, jagged was helping but we ran into a problem now that i have it fixed he has had to leave... I am setting my my first RAID1 system... here is a image of where i am at.. http://farm3.static.flickr.com/2494/4012511630_6e26092308.jpg     can someone walk me through what if anything i need to do to create lvm logical partitions for the mount points of my filesystem....
[21:40] <jph_> o by the way i have changed the RAID1 device #0 501.6MB partition to mount at /boot
[21:40] <jph_> from whats in that image
[21:42] <jph_> Anyone?
[21:47] <_ruben> "Configure the Logical Volume Manager" .. create a VG containing your md1 as PV .. create a LV for each partition you want to have .. configure each LV (choose filesystem, mountpoint, etc)
[21:48] <jph_> ok hang on let me do that --------Thankyou for stepping in i am on a dead line and need to get this going tonight
[21:50] <_ruben> wont be around much longer though, its nearing bedtime ;)
[21:50] <jph_> would you be willing to recommend some partition sizes this will be a basic LAMP server for a small group of people we will latter add a 1TB RAID for media storage
[21:51] <jph_> I have 119.5Gb to work ith
[21:51] <jph_> with
[21:51] <_ruben> you'll need to figure out/guestimate which parts will take the most space .. web content? databases? users's homedirs?
[21:51] <jph_> I have created the VG names 'Group1' and I am creating the root LV
[21:52] <jph_> well the users (there will be 5) will all be creating basic HTML pages which will be located at /home/user/public_html
[21:53] <_ruben> 120G of html .. wow :)
[21:53] <jph_> there will be a main home page for the group located at /var/www
[21:53] <jph_> ---- no just the HD we had on stock and paid for
[21:53] <jph_> I have no worries about having the space just don't know how to cut it up
[21:54] <_ruben> i'd go for smth like 10G for / .. 20G for /var/www .. might wanna reserve some gigs for /var/log (and mount it with noatime) .. put the rest up for /home
[21:56] <jph_> if i will have a database that will contain a message board should i make / larger?  also where does everything mount that i don't make a mount point for myself
[21:56] <_ruben> everything else = /
[21:57] <_ruben> if you plan on having large mysql databases .. you might wanna got with a seperate LV for /var/lib/mysql/
[21:57] <_ruben> (having 'dedicated' LVs for certain tasks allows you to tweak mount options for those tasks)
[21:57] <_ruben> like noatime for /var/log
[21:58] <jph_> ok well let me make these one at a time.. is there any reason / should ever be larger than 10g
[21:58] <_ruben> increases log performance, unless you care about when a log was last accessed
[21:58] <_ruben> most my systems have a 2G / :)
[21:59] <_ruben> and use about half of it
[22:00] <jph_> ok then 10g it is, i am a buding system admin, normally i have an expert here to help, wouldn't i want to know when a log was accessed incase someone was trying to cover there tracks ?
[22:00] <_ruben> if a certain directory grows out of proportion, add extra disks, add it to LVM, create large chunk for this data, migrate data, mount it at the right spot
[22:00] <_ruben> accessing logs = reading logs, not writing
[22:00] <_ruben> atime = read .. mtime = write
[22:01] <jph_> o ok
[22:02] <jph_> should i just make /var/log big or should i make /var/log  x number of gigs and then make /var/log/mysql its own thing?
[22:02] <_ruben> (which is why i love virtualization and/or shared storage .. darn easy to adjust diskspace to (new) requirements)
[22:03] <_ruben> well .. /var/log/mysql doesnt tend to get big (unless you're doing mysql replication which stores its logs there by default), /var/lib/mysql does (it holds the databases)
[22:05] <jph_> so should it be its own partition to start or should i just make /var/log big to start... if i am going to have a message board that will grow over time how big should this (these) be
[22:05] <jph_> I need a starting point and i guess i will learn and change things over time
[22:06] <_ruben> indeed ... and i prefer having /var/log on its own .. how big depends on the expected activity (and how often you rotate logs and how long you want to keep them)
[22:06] <_ruben> nothing's a requirement .. put all under / will work just as fine (grub wont work with /boot on lvm, which is why i keep /boot out of my lvm)
[22:07] <jph_> well in truth that will be up to thom (the guy teaching me how to do all this) I would guess i would keep logs for a while so he could show me trends over time
[22:07] <_ruben> if it's hard to guestimate the sizes of the various parts, keep the cutting up to a minimum
[22:08] <jph_> is 5g for /var/log over kill at its best or a good idea
[22:08] <_ruben> i'd say it'd be a nice start
[22:08] <jph_> and would that be about right to start out with for the /var/log/mysql
[22:08] <_ruben> i tend to use way less space in /var/log than i anticipate for .. never hurts to have some extra space there tho :)
[22:09] <jph_> right
[22:09] <jph_> if i have it might as well be safe...
[22:09] <_ruben> forget about /var/*log*/mysql unless you do mysql replication .. /var/*lib*/mysql could use its own LV for the databases :)
[22:10] <_ruben> and i have no clue about how big your databases will be :)
[22:11] <jph_> ok sorry for the typo... /var/lib/mysql wil have 2 databases and i don't foresee more than 4 ever... as of now it will have a user database for login info and contact info... and a messages database for the message board...
[22:11] <_ruben> sounds more like MB's than GB's .. i'd go for 10 or 20 G to be safe
[22:12] <jph_> 10 it is
[22:12] <jph_> again thank you for your time.. let me make that sql LV
[22:13] <jph_> humm ok so the rest should be home? anything we have forgotten before i make that one
[22:14] <_ruben> could always leave like 20G or so "free" incase you might need some space somewhere without having to add disks right away
[22:15] <jph_> ok good idea, will do
[22:15] <_ruben> the benefits of using lvm are that especially with some space to spare, you can re-distribute your space fairly easily
[22:16] <_ruben> (usualy with a bit of downtime, but beats having to reinstall from scratch :))
[22:17] <jph_> that would leave me with more or less than a 40Gig Home and about 26Gig of free space
[22:17] <_ruben> sounds fair enough i'd say
[22:17] <jph_> great lets see if it works
[22:18] <_ruben> and im gonna go see how things are in never never land .. off to sleep .. g'luck! :)
[22:19] <jph_> thanks again have a great night
[22:22] <bventura> my default apache install has /var/www set with root as owner and group.. is it OK to change it to be owned by www-data so that I can move files to it without using the root account?
[22:33] <tomsdale> need a little shell scripting hint - I can split up a command in multiple lines by using a backslash at the end of the line. Is there a way of somhow adding a comment behind the backslash? reason is that I want to document the options I use in an rsync command line by line
[22:42] <erichammond> bventura: Not requiring root for /var/www is recommended.  However, it would be best to use a normal user other than the one which is running Apache.  That reduces security risk from potential holes in Apache or your web application by limiting what files can be written to by the attacker.
[22:43] <erichammond> tomsdale: One approach would be to build the command line by line in a variable:
[22:43] <erichammond> cmd="$cmd --option" # comment
[22:43] <erichammond> and then run it:
[22:43] <erichammond> $cmd
[22:44] <tomsdale> erichammond: sounds easy eno
[22:44] <tomsdale> ugh - thx let me give it a try
[22:46] <erichammond> tomsdale: Or just add a big comment above the command.  This would make the command easier to read for people who know what the options are for.
[22:46] <tomsdale> it's just rsync has so many options - and I would like to be able to easily remove or add options in the script file.
[22:46] <bventura> erich, that makes a lot of sense, thanks for the info
[22:49] <jph_> I just installed Ubuntu Server on my first RAID ever built solo (well with a lot of help form people here)  Is there a way to check that the RAID is writing to both drives correctly and creating a full mirror, or is it one of those if it works it works things
[22:51] <erichammond> tomsdale: Yep, I'm a fan of using backslashes to break long commands into multiple lines so options are easy to read: http://alestic.com/2009/06/ec2-ebs-raid
[22:54] <bventura> jph I'm not an expert on raid but I believe what you want to do is lookinto the 'mdadm' command
[22:55] <bventura> it'll report status + you would use that to repair mirror, etc
[22:57] <tomsdale> erichammond:  you seem to know a good deal about raid. I'm writing this rsync script to clone a running server on software RAID-1 into a virtual machine (no raid on the vm) Do you think the Raid will pose a problem?
[23:06] <erichammond> tomsdale: I am in no way a raid expert.  I have been able to collect research from the experts to put together the exact commands which work for a particular situation, and I have the willingness to publish documentation about it, but that's about it.
[23:07] <tomsdale> personally I think there should be no problem but I have no experience with RIAD. The filesystem doesn't know it's on a RAID and I will adjust the fstab on the destination system.
[23:07] <erichammond> tomsdale: That said, I think any form of RAID should be transparent to the process you describe.
[23:07] <erichammond> right
[23:07] <erichammond> er, not sure you should be rsyncing the fstab...
[23:07] <owh> Salutations all. Anyone know how to solve the issue of doing an atomic file transfers using ftp?
[23:10] <tomsdale> I probably put it in the exclude or change it in an chroot environment afterwards. Already cloned a test image across vms. All you had to adjust was the fstab and the grub/menu.lst.
[23:11] <tomsdale> with the UUID's.
[23:15] <FFForever> what do you guys think of this so far?, any other information you would want to see (its going to be a system monitor in the long run...)
[23:15] <FFForever> http://chr1831.com/qstats2/
[23:17] <owh> FFForever: IIRC there is going to be (or already is) a system status page on login. I'm pretty sure that it has a means to extend the things it reports on, so you might be duplicating your efforts.
[23:18] <FFForever> owh, this will be for hosting systems mainly where people don't have a head or sometimes even a dedicated server
[23:19] <tomsdale> FFForever: do you get the information remotely?
[23:19] <owh> FFForever: The login reference was just to jog my memory - not that it brought up more than I already told you - but my point was that there is already a system monitor infrastructure in place.
[23:20] <FFForever> tomsdale, that will be in version 3 or so i hope for remote monitoring
[23:28] <tomsdale> FFForever: for remote monitoring I'm playing around with Nagios atm. maybe a little to big for my needs but it seems to be very extendable.
[23:28] <FFForever> tomsdale, free?
[23:28] <bventura> also - try zabbix I LOVE that software, also free
[23:29] <FFForever> bventura, I HATE U!!!!!!!!!!!!!!!!!!!
[23:29] <FFForever> i was going to basically write zabbix...
[23:29] <tomsdale> yes - both are free.
[23:30] <bventura> doh
[23:31] <FFForever> ohhh well i am still going to finish my stat script i need the php experience
[23:31] <owh> :)
[23:31] <bventura> hey speaking of PHP try smarty if you haven't seen it, it's really cool, a templateing engine
[23:31] <bventura> another one of my favorites
[23:31] <FFForever> smarty is a pitfa
[23:31] <tomsdale> zabbix actually looks nicer than nagios at least from the layout
[23:31] <owh> There is also something like php-system-status, not sure exactly what it's called - also free :)
[23:38] <tomsdale> and apparently there is opennms - just checking it out - nice how you can creat your graphs on the fly. I'm a sucker for graphs :-)
[23:39] <bventura> zabbix has got mad graphing tomsdale, you can graph anything, it's got a wizard for making them + you can make a whole layout page where you can combine graphs with status indicators or raw data
[23:39] <bventura> you gotta try an install if you haven't already
[23:41] <tomsdale> I think I will, looks great - I already installed nagios and munin on my new hosting server but I'm still evaluating software
[23:42] <bogeyd6> kudos on choosing nagios tomsdale
[23:42] <bogeyd6> now lemme throw a wrench into your equation
[23:42] <bogeyd6> tomsdale http://www.centreon.com/
[23:45] <tomsdale> haha - there is just too much software out there. I think there is also a more community centric nagios fork on the way http://www.icinga.org/
[23:49] <tomsdale> bogeyd6: from what I understand centreon is an addition to nagios?
[23:50] <leaf-sheep> I'm trying to set up PXE for first timer. Could somebody verify with my processes? I don't know what I'm missing out on.  I have dhcp3-server installed, I had placed the PXE netboot files in /var/lib/tftboot and set up link for /var/www/ubuntu.