[00:54] <PineappleCLock> Hello, I followed the server guide for installing rails, and I don't see how to actually serve the application. does apt-get install rails install Passenger / mod_rails ?
[00:57] <irv> is it possible to add a cron job to be executed as my user account but start when the system reboots?
[00:58] <irv> or woudl that be accomplished some other way?
[00:58] <umopaplsdn> Can someone help me with an Ubuntu Server installation?
[00:58] <irv> it's ZNC
[00:59] <umopaplsdn> The problem is that I need to use a USB to install the server, yet I cannot find the Ubuntu Server 12.04 on the Universal USB Installer. I have downloaded the ISO. Any suggestions?
[01:01] <zakk> irv: crontab -e as your user, afaik you can only do it on a time-basis but you could insert other logic into your script
[01:01] <zakk> umopaplsdn: I just did this yesterday, it's below Ubuntu Studio 12.04
[01:01] <zakk> umopaplsdn: it's like Ubuntu Server Installer 12.04
[01:01] <umopaplsdn> Okay, thanks!
[01:01] <umopaplsdn> I'll take a look.
[01:02] <umopaplsdn> Oh I see... Odd place to put the server installer
[01:02] <umopaplsdn> Thanks, found it.
[01:02] <zakk> yeah I was kind of surprised
[01:02] <umopaplsdn> See ya'!
[01:02] <zakk> l8r
[01:03]  * PineappleCLock redacts question
[01:03] <irv> zakk: hmm, like i have a cronjob doing it that i added as my user
[01:03] <irv> but it only launches ZNC when i login to the VM
[01:03] <irv> the cron line is */2 * * * * /usr/bin/znc >/dev/null 2&1
[01:04] <zakk> irv: I would just run that as a service via init scripting
[01:08] <irv> that was my next option, but this was the only thing outlined in the ZNC faq
[01:08] <irv> :P
[01:10] <zakk> I'm not familiar with znc, but it sounds like a service
[01:10] <zakk> so I'd run it and just have the init script su - to myuser
[02:38] <Eitan> hey guys. i am hammering a new ubuntu server with tons of connections, i was running centos for my application before with no problems... For some reason i am getting a lot of connection time outs when establishing connections both mysql 3306 and redis 6397... first i found that syn flood protection was blocking some stuff, so i managed to turn that off. but im not sure to see where else
[02:38] <Eitan> or why else connections are timing out as far as logs or perhaps if its a ulimit issue?
[02:41] <Eitan> any ideas fellas
[02:59] <aarcane> how can I tell the ubuntu server partitioner to use GPT partition tables?  It keeps wanting to make a legacy MBR partition table, and the usual utilities (parted, gdisk) aren't present on the installer.
[04:26] <uvirtbot`> New bug: #998504 in samba (main) "samba generates errors when installing. Also have problems with samba4. Tried uninstalling and reinstalling." [Undecided,Expired] https://launchpad.net/bugs/998504
[04:26] <uvirtbot`> New bug: #1000716 in php5 (main) "libapache2-mod-php5/php5-fpm post-inst script segfaults (exit status 139)" [Undecided,Expired] https://launchpad.net/bugs/1000716
[04:27] <uvirtbot`> New bug: #993657 in samba (main) "package samba 2:3.6.3-2ubuntu2.1 failed to install/upgrade: subproses skrip post-installation terpasang menghasilkan kesalahan status keluaran: 1" [Undecided,Expired] https://launchpad.net/bugs/993657
[04:27] <uvirtbot`> New bug: #994160 in memcached (main) "package memcached 1.4.13-0ubuntu2 failed to install/upgrade: ErrorMessage: subprocess installed post-installation script returned error exit status 1" [Low,Expired] https://launchpad.net/bugs/994160
[04:27] <uvirtbot`> New bug: #995449 in php5 (main) "cannot install package php5-intl on 12.04 (Precise)" [Medium,Expired] https://launchpad.net/bugs/995449
[04:27] <uvirtbot`> New bug: #995939 in mysql-5.1 (main) "package mysql-server-5.1 5.1.62-0ubuntu0.11.04.1 failed to install/upgrade: el subproceso instalado el script post-installation devolvió el código de salida de error 1" [Undecided,Expired] https://launchpad.net/bugs/995939
[04:27] <uvirtbot`> New bug: #996293 in samba (main) "Cannot login using active directory users ( Unknown id: $username )" [Medium,Expired] https://launchpad.net/bugs/996293
[04:49] <kyle6513> I'm having trouble getting php5 to execute scripts on an apache2 install, it's just displaying the scripts. Anyone who can help?
[08:23] <zastaph> I wonder why I didn't hear about the RAID write hole until now that I actually decided to go RAID
[08:23] <zastaph> now that I know about it, it's the first thing I would tell people if they asked me advice about RAID :)
[08:25] <zastaph> now I wonder how to deal with it
[08:25] <zastaph> would an UPS be sufficient?
[08:26] <ivoks> raid hole?
[08:26] <zastaph> http://en.wikipedia.org/wiki/RAID_5_write_hole
[08:27] <zastaph> I also understand now that ZFS was invented to deal with just that
[08:28] <ivoks> raid5 :)
[08:28] <zastaph> http://www.raid-recovery-guide.com/raid5-write-hole.aspx
[08:28] <zastaph> nah it's in all of them
[08:28] <ivoks> raid cards should allways have a battery
[08:29] <zastaph> ok, but what about Linux Software RAID ?
[08:30] <ivoks> ups, of course
[08:30] <zastaph> yeah, so I'm looking for one now
[08:30] <ivoks> this isn't a big problem
[08:30] <zastaph> and I presume for an UPS to have any significant impact, you need some kind of software to listen to the UPS and gracefully shutdown on errors?
[08:31] <zastaph> like http://www.apcupsd.org/
[08:31] <zastaph> but does it run in Ubuntu ?
[08:31] <ivoks> that's a definition of ups :)
[08:31] <zastaph> does APC deliver that software? i tried to read manuals, couldn't find info
[08:31] <ivoks> and this problem is not frequent
[08:31] <zastaph> frequent or not.. silent data corruption sounds pretty bad to me
[08:31] <ivoks> all major companies work with nut
[08:31] <zastaph> I choose RAID to secure my data, not to corrupt it :)
[08:32] <zastaph> nut?
[08:32] <ivoks> and support its development
[08:32] <zastaph> im a SOHO user
[08:32] <ivoks> nut, yes
[08:32] <ivoks> apt-cache show nut
[08:32] <ivoks> http://www.networkupstools.org/
[08:32] <zastaph> sudo apt-get install nut ?
[08:32] <zastaph> sudo service restart nut
[08:32] <zastaph> voila?
[08:33] <ivoks> you need to configure it, of course
[08:33] <ivoks> for your ups
[08:33] <zastaph> need to find compatible ups too I guess
[08:33] <ivoks> http://www.networkupstools.org/stable-hcl.html
[08:35] <zastaph> but why not http://www.apcupsd.org/ if I buy APC anyway
[08:35] <ivoks> that's also an option
[08:36] <zastaph> I don't need a big battery or anything, and I don't need uptimes at all.. i'm only buying UPS to avoid write hole and as surge protection.. so 1-2 minutes of shutdown time for 2 mini servers at less than 50W each
[08:37] <ejv> my apc is buggy on 10.04 lts, your mileage may vary
[08:37] <ivoks> i had no problems with eaton, powerware and blazer
[08:37] <zastaph> ivoks, can you narrow those 3 down to 1 recommendation? :p
[08:37] <ejv> the ups itself is fine, it just loses it's connection to apcupsd constantly
[08:38] <ivoks> zastaph: i can't - check them out, see what you can get and what prices are affordable for you
[08:38] <zastaph> with so many UPS options I figure I'd spend all day picking one :p
[08:39] <zastaph> the joy of linux.. reading about hardware support
[08:40] <ivoks> i've seen hardware that doesn't work with windows
[08:40] <ivoks> actually, i see it on a daily basis :)
[08:41] <zastaph> yeah but all hardware vendors write on their support page if they have windows or mac.. for linux you need to search elsewhere :)
[08:41] <zastaph> speaking of, did you see http://liliputing.com/2012/05/first-look-at-project-sputnik-dell-xps-13-ultrabook-with-ubuntu-linux.html
[08:42] <zastaph> I'd like that,but seems not quite finished/supported http://hwe.ubuntu.com/uds-q/dellxps/
[08:43] <ivoks> yes, i've seen it
[08:44] <ivoks> and, afaik, multitouch has been resolved since then
[08:45] <ivoks> lol that video just shows that trackpoint is unbeatable innovation :)
[08:46] <zastaph> Dell is on the top of the list for hardware certification http://www.ubuntu.com/certification/desktop/
[08:46] <zastaph> yet a laptop has to be customized for ubuntu and even that isn't finished :)
[08:47] <zastaph> i think hardware support is what prevents me from going full ubuntu
[08:47] <zastaph> for desktop that is :)
[08:48] <ivoks> that's a list of machines that were given to get certification
[08:49] <ivoks> it's in no way full list of working devices
[08:51] <zastaph> I also wanted to try to install it on a macbook: https://help.ubuntu.com/community/MacBook
[08:52] <zastaph> but reading that list of different side-effects on all versions of Macbook matched with all versions of Ubuntu made me change my mind :)
[08:53] <RoyK> anyone here ever setup sanlock with kvm?
[08:54] <zastaph> ubuntu for Slate computers also seems risky https://wiki.ubuntu.com/ARM/TabletList
[08:55] <zastaph> ivoks, seems powerware is eaton
[09:03]  * RoyK has a pandaboard with precise ;)
[09:06] <rbasak> smoser: sorry, I wasn't around. Ping me when you get in if you like.
[09:09] <zastaph> I presume you connect an ups to your servers with one power cable and one usb cable? so to notify 2 servers about power failure I need one that has multiple USB.. seems the cheap ones only has 1
[09:14] <zastaph> http://askubuntu.com/questions/107883/how-to-use-a-eaton-3s-700va-ups-with-ubuntu-server
[09:14] <zastaph> if my power goes, the network goes
[09:15] <zastaph> so I guess servers would shutdown alone if the network went
[09:23] <RoyK> zastaph: if the UPS can't hold power sufficiently long for the servers to be shut down, ten get a new UPS :P
[09:24] <zastaph> ahh the switch will also be on ups, but my internet router will be in another room.. but i guess that wont break the LAN
[09:28] <zastaph> but the router is the DHCP server
[09:28] <zastaph> so probably it will break something on the LAN
[09:31] <zastaph> oh well I only have 1 server with valuable files on.. the other server is a backup server, and I guess it wouldnt be disaster if it wasnt rock solid
[09:32] <zastaph> but then the UPS battery will always be emptied, if even some units attached aren't aware of the power fail
[09:34] <kuul> Hi i am trying to create clustered lvm but i can not pass "vgcreate --clustered y my group"  command on 10.04. am i doing something wrong?
[09:56] <uvirtbot`> New bug: #1025595 in tomcat7 (main) "tomcat7 test failed: AssertionError: ['tcp6 0 0 :::8080 :::* LISTEN 898/java '" [Undecided,New] https://launchpad.net/bugs/1025595
[10:09] <reisi> after enabling sssd via pam and nsswitch.conf shouldn't i be seeing all my users (from sssd domain[s]) with getent passwd?
[10:19] <reisi> with sssd debug level 0x1310 i cannot see anything related to authentication when i try to login into the system; any ideas?
[10:20] <reisi> auth.log has: pam_sss(sshd:auth): received for user reisi: 10 (User not known to the underlying authentication module)
[10:23] <jamespage> rbasak, around?  probably easier to discuss that jenkins thing
[10:23] <rbasak> jamespage: yes
[10:24] <jamespage> rbasak, just reading - does this stuff run in a jenkins job?
[10:25] <rbasak> jamespage: yes - the code you see creates the jenkins job with that as the contents
[10:25] <rbasak> jamespage: although, that's odd. It expects to be at the root of the checked out tree
[10:26] <rbasak> I'm not sure I follow what Al's doing any more
[10:26] <rbasak> Won't he end up two levels higher if he does that?
[10:26] <rbasak> ahs3: ^^
[10:26] <jamespage> rbasak, me neither - I think I'm lacking context as to what he's trying todo
[10:27] <jamespage> rbasak, do you inject config.xml files directly into the filesystem; or do you use the REST API through python-jenkins?
[10:27] <rbasak> API
[10:27] <jamespage> rbasak, right-oh
[10:59] <jamespage> zul, adam_g: did I see some crazy mad talk about adding nodejs to the dependencies of horizon?
[11:01] <zastaph> is it possible to buy a RAID BBU cheap?
[11:02] <ikonia> BBU ?
[11:07] <_ruben> depeonds on the raid controller obviously
[11:07] <_ruben> but they're easily close to $100
[11:10] <zastaph> I dont want to use hardware raid because if the card breaks I need to find an exact match
[11:10] <zastaph> but I thought about using just the BBU part of the card
[11:10] <zastaph> sounds complicated
[11:10] <zastaph> BBU = battery backed cache?
[11:11] <andygraybeal> lots of raid on ebay, right?
[11:11] <andygraybeal> just buy a popular used one
[11:11] <ikonia> what's a BBU ?
[11:11] <zastaph> is it possible to use an SSD for raid cache, as replacement for BBU?
[11:11] <zastaph> anything to prevent raid write hole
[11:11] <ikonia> zastaph: I wouldn't suggest that
[11:12] <ikonia> zastaph: you want a crappy host raid card, they are cheap, you want a quality true hardware raid controller, they are not
[11:12] <zastaph> I just want to setup a server and sleep well at night
[11:12] <ikonia> then buy a hardware raid card
[11:12] <ikonia> (that has good linux support, eg: 3ware, LSI)
[11:13] <zastaph> and if the card breaks in 4 years when they are discontinued?
[11:13] <ikonia> then replace with a different card
[11:14] <ikonia> or buy from an EOL supplier
[11:14] <zastaph> I read that you can't just get your data back if a hardware raid card breaks
[11:14] <ikonia> that isn't true
[11:14] <ikonia> are you doing a stripe or a mirror ?
[11:15] <zastaph> i don't know.. 2 servers with 4 disks.. probably 5/6 for backup server and 10 for file server
[11:15] <ikonia> well, the mirror certainly shouldn't be a problemn to revover from,
[11:15] <zastaph> 3ware/LSI over budget
[11:15] <ikonia> the stripe, I guess it depends on the mechanism it uses
[11:15] <ikonia> zastaph: then you need to accept the limitations of your budget
[11:15] <_ruben> safety costs money
[11:16] <_ruben> and no, using a bbu without a raidcontroller is not possible
[11:16] <ikonia> the good thing with the LSI/3ware you can upgrade your card model and the raid is still valid
[11:16] <_ruben> well, zfs with its cache on ssd is as close as one gets i guess
[11:16] <zastaph> so.. if I get an UPS, would it be safe enough?
[11:16] <ikonia> you don't even have to rebuild the array, I moved from an 8 series to a 9 series card without issue
[11:16] <ikonia> zastaph: it's up to you what you consider safe
[11:16] <_ruben> ups wont help if your system crashes
[11:17] <RoyK> zastaph: you don't lose your IP if the DHCP server goes down
[11:17] <zastaph> RoyK, yeah I figured that too :) just need to choose the right UPS
[11:18] <ikonia> errr, UPS isn't really the priority here.
[11:18] <zastaph> someone here must use UPS? and have good or bad experiences with ubuntu server
[11:18] <ikonia> unless I've missunderstood
[11:18] <ikonia> I use many small and enterprise UPS devices
[11:18] <zastaph> small
[11:18] <zastaph> not interested in uptime at all
[11:18] <ikonia> APC
[11:18] <zastaph> data safety all the way
[11:18] <ikonia> then why are you buying UPS
[11:18] <ikonia> I have two in my house
[11:18] <ikonia> if you are not interested in uptime
[11:19] <zastaph> to avoid raid write hole
[11:19] <zastaph> and power surges
[11:19] <ikonia> UPS will only give you safe shutdown in the event of power loss
[11:19] <_ruben> a raidcontroller with bbu would be more usefull than a ups
[11:19] <ikonia> yeah,
[11:21] <zastaph> seems complicated to install one in my hp microserver
[11:21] <zastaph> if there's even space for it
[11:21] <zastaph> I don't have enterprise needs.. I just want to protect my data
[11:22] <_ruben> as stated before, everything has its price :)
[11:23] <zastaph> ok i'll accept my limitations :) btw. are some RAID setups more prone to write hole than others?
[11:23] <zastaph> there gotta be a reason it's refered to as raid 5 write hole
[11:24] <zastaph> and I also read about software that can detect write holes.. perhaps I can accept the unlikely chance of having one, but then I would like to detect it early
[11:26] <_ruben> not using raid5 would be another option obviously
[11:26] <zastaph> yeah I thought about not using RAID at all
[11:27] <zastaph> but then 4 2TB disks suddenly seems like an awful waste :)
[11:27] <zastaph> i thought about some kind of rsync replacement
[11:27] <zastaph> to deal with mirroring
[11:29] <zastaph> but as I read it, all RAID types (except ZFS) has write holes
[11:29] <zastaph> unfortunately using ZFS requires you to deal with operating systems miles from ubuntu server in usability
[11:33] <ivoks> have you ever configured raid in your life?
[11:36] <zastaph> never :)
[11:36] <zastaph> but my theory is getting stronger
[11:36] <samba35> i am "trying " to understand openstack , i have test website installed and it was working as expected but after installed some packages and openstack my default web site is showing ubuntu  openstack dashboard page with username and passwd option how do i fix this  or how do i trace this
[11:37] <_ruben> raid5 isn't the only raid level....
[11:37] <zul> jamespage: if i can avoid it
[11:37] <zastaph> ruben, I know, and? as I wrote the other raid level has the write hole too
[11:38] <zastaph> ikonia, what UPS did you have, and do they play well with ubuntu server?
[11:39] <_ruben> zastaph: as i read it (only read the wikipedia article on it), it'd be limited to raid levels that use partiy, so raid1 and raid10 for instance don't suffer from it
[11:39] <_ruben> sure, corruption can still occur
[11:39] <zastaph> ruben, http://www.raid-recovery-guide.com/raid5-write-hole.aspx
[11:40] <_ruben> if a disk is starting to fail, you can get corruption as well, raid or not
[11:40] <zastaph> ok so what's difference betwene write hole and corruption?
[11:40] <zastaph> that link says RAID 10 inherits it from RAID 1
[11:40] <_ruben> write hole appears to be just one example of possible corruptions
[11:41] <jamespage> zul, that would be good....  its extremely fast moving so not really suitable for LTS style support IMHO
[11:41] <_ruben> ah, then it's just a name to indicate any "raid-related" corruption after all
[11:41] <zastaph> probably
[11:41] <_ruben> with raid5 a write failure could go undetected, with raid1 it would be detectable
[11:41] <zastaph> notice the list of tips to reduce the effects below
[11:41] <zastaph> UPS and  "Synchronize your array regularly"
[11:42] <zastaph> I wish he would specify how
[11:42] <ikonia> zastaph: I'm using APC, yes they work fine
[11:42] <zastaph> ikonia, nut or apcupsd
[11:43] <ikonia> apcupsd
[11:43] <zastaph> and you have 2 because they control 2 computers I guess? why not 1 to control all
[11:43] <ikonia> I have one because I have two setups, one in the cellar and one in the attic
[11:44] <ikonia> they power 2 - 3 devices each
[11:44] <ikonia> give approx 15 - 20 minutes of run time to the 2 - 3 devices and a clean shutdown and alert
[11:44] <zastaph> and i presume they only have 1 usb? so the other 1-2 devices gets notified by network from the master computer?
[11:44] <ikonia> correct
[11:44] <ikonia> there is a script on the master that will shutdown the other 2 and the master shutdown gets invoked from the unit
[11:45] <zastaph> how often do you change the battery
[11:45] <ikonia> not changed either in 2 years
[11:45] <ikonia> still showing %100 charge
[11:45] <zastaph> isn't 15-20 minutes overkill simply for shutting down?
[11:45] <ikonia> the two power outages I've had they have both held up exactly as expected
[11:45] <ikonia> the batter test option is still showing fine
[11:45] <ikonia> zastaph: it doesn't shutdown until it's a confirmed long outage
[11:45] <ikonia> I don't want to shut everything down if it's just a little blip
[11:46] <zastaph> right
[11:46] <ikonia> I think the threshold is 7 minutes-ish I set on the delay
[11:46] <ikonia> I can't remember, it's 2 years ago
[11:46] <zastaph> and surely you tested it :)
[11:46] <ikonia> yes, and I've had 2 unplanned outages that worked fine
[11:46] <ikonia> (I may have had more, but I can only think of two)
[11:47] <ikonia> I've had a few blip outages of 90 seconds or less that obviously didn't cause a problem
[11:47] <ikonia> the two outages that where long ones worked fine
[11:47] <zastaph> recall the model number?
[11:47] <ikonia> hang on
[11:47] <ikonia> I'll see if it's on the site
[11:48] <ikonia> although I'd still suggest you invest in a proper raid card over a ups
[11:48] <zastaph> don't think I have this option with my 2 atom based mini servers :)
[11:48] <ikonia> Pro 900
[11:49] <_ruben> must be tiny instead of mini servers, if there isn't room for add-in cards
[11:49] <ikonia> I do have a 3rd the pro 550, but that's not hooked up to computer equipment, that's powering my PVR and a few other odd house hold items
[11:50] <ikonia> I suspect the 550 would be perfect for your needs
[11:52] <zastaph> even just backups ES series (not pro) might be
[11:53] <ikonia> I'm not sure how the shutdown stuff works on the ES stuff, I wasn't convinced by it when I looked
[11:53] <ikonia> things may have changed
[11:54] <zastaph> it's terrible how many different models they have :)
[12:19] <zastaph> http://justwhatdoyouthinkyouredoingdave.blogspot.com/2012/06/ubuntu-nas-zfs-on-hp-microserver.html
[12:19] <zastaph> native linux ZFS filesystem ?
[12:25] <zastaph> As of June 2012, the port is in release candidate status for version 0.6.0, which supports mounting filesystems
[12:25] <zastaph> not ready for production i presume?
[13:05] <reisi> zastaph: i just tried out zfsonlinux.org for server usage; basically it's not there yet
[13:05] <zastaph> thought so
[13:05] <reisi> zastaph: if you want to try it head over to #zfsonlinux
[13:06] <reisi> zastaph: you'll need to use spl and zfs from the gentoo branch of ryao, and patch a custom kernel
[13:06] <zastaph> in other words, no thanks :)
[13:06] <reisi> zastaph: last is optional, but basically required as it could cause rare deadlock
[13:07] <zastaph> i could imagine not using raid at all
[13:08] <reisi> zastaph: i tried it with ryao's branch but without the kernel patch, everything worked out nicely, even with swap over zpool, but the discussion over the kernel patch lead me to believe it's rather essential
[13:08] <zastaph> just the word branch makes me say no thanks :) if the master branch isn't capable i'm not joining yet
[13:09] <zastaph> also.. ZFS is all about revisions .. with RAID there's no revision.. it's just RAID :)
[13:10] <reisi> if you mean RAID == mdadm there are revisions as well, but i guess those happen less "often" (not that zfs upgrades that often it'd seem)
[13:11] <xnox> well RAID = mdadm has multiple format revisions, there are dataformats as well e.g. ISMS/DDF, and hardware raid controllers...
[13:11] <ivoks> for someone that is so afraid about exceptionally rare problem, you are buying a server that will probably crash and burn before you ever see 'raid hole'
[13:11] <xnox> usually borked hw controller means you need to find the exactly same matching controller...
[13:12] <ivoks> if you have so precious data, which even backups can't restore, then you should buy a DC, not a microserver
[13:13] <ivoks> DC with bunch of servers, replication and hardware raid cards of $500+
[13:13] <reisi> apparently i jumped into the middle of the conversation..
[13:13] <ivoks> :)
[13:14] <ivoks> xnox: not exactly the same, but from the same vendor
[13:14] <Exopaladin> I like zfsonlinux for my file server at home, but wouldn't run it on a production server :P
[13:14] <xnox> ivoks: some vendors don't tell you which controllers are compatible, and if you bought one 10 years ago....
[13:16] <reisi> Exopaladin: exatcly; but i guess it'll be ready for server testing soon, i hear some are already running off btrfs as well :)
[13:17] <Exopaladin> I'd almost trust zfsonlinux more than btrfs, the fact that their latest news announcement says they now have an fsck that fixes *some* problems isn't exactly encouraging
[13:18] <ikonia> pretty hard to argue btrfs's stability when Oracle have deemed it production ready
[13:18] <zastaph> problem is I don't know how rare write hole is.. and I don't know how serious it is when it happens.. is it a hole that keeps digging itself deeper, or is it just a corruption that occurs on the active file being written to during power outtage ?
[13:20] <zastaph> Eaton Ellipse ECO line of UPS's look nice
[13:20] <zastaph> and they verified to me that their own software is ubuntu compatible
[13:22] <zastaph> http://pqsoftware.eaton.com/explore/eng/ipp/default.htm?lang=en
[13:38] <maplesoft>  I cant run a .jsp or .war     by tomcate.  i have mady /user/www as my apache dir for php and html files and place .jar and .war in that too.      https://help.ubuntu.com/12.04/serverguide/tomcat.html#tomcat-configuration       i cant get help from httpd or tomcat channels
[14:02] <jamespage> maplesoft, I think I understand your question
[14:02] <jamespage> tomcat won't by default look for war/.jsp files in /usr/www
[14:03] <jamespage> each war file (which can contain .jsp's) must be placed in /var/lib/tomcat6/webapps
[14:17] <maplesoft> jamespage hm... when i type mydomain.com:8080  it shows ....."it works. " so where is that file that sys that?
[14:19] <maplesoft> jamespage It works!
[14:19] <maplesoft> This is the default web page for this server.
[14:19] <maplesoft> The web server software is running but no content has been added, yet.
[14:19] <maplesoft> where is this file located?
[14:34] <maplesoft> the mysite.com:8080/index.jsp starts downloading . not executing.   so tomcat is not running. but it it is running and thats why i can see index.html on port 8080      ( mysite.com:8080 )  . what is happening?
[14:44]  * maplesoft waits
[14:48] <zul> maplesoft: check your logfiles they can you tell what is happening better than we can
[14:55] <zastaph> phew, took 5 hours to narrow down the options for UPS to a purchase :)
[14:59] <maplesoft> tomcat runs .html files in /var/www   on its default port of 8080 but does not runs .jsp files . instead it starts downloading them. why is that?
[15:08] <zastaph> no wonder noone fixed the raid 5 write hole yet, it's patented: http://www.google.com/patents/US7069382
[15:09]  * ogra_ is happy to live in times where you can patent a hole
[15:09]  * xnox wonders if my hole is patented....
[15:10] <ogra_> xnox, does it use multiple holes to merge to a big raid-hole ? then it might :)
[15:11] <xnox> funny you should say that
[15:12] <glance> zastaph: zfs / wafl solves raid 5 write hole another way, by cow
[15:13] <zastaph> yes but zfs on linux is "pre-alpha" according to someone, and "release candidate" according to others
[15:13] <zastaph> and I guess discussing opensolaris selections in here is not appropriate :P
[15:19] <jamespage> maplesoft, tomcat does not server files in /var/www
[15:19] <jamespage> are you sure that apache is not serving them?
[15:23] <maplesoft> jamespage apache is stopped
[15:24] <maplesoft> jamespage service apache2 status
[15:24] <maplesoft> Apache2 is NOT running.
[15:24] <maplesoft> jamespage please stick with me for some time
[15:24] <jamespage> maplesoft, and tomcat is?
[15:25] <jamespage> 'instead it starts downloading them. why is that?' would indicate that tomcat is NOT serving those files
[15:25] <maplesoft>  * Tomcat servlet engine is running with pid 19127
[15:25] <maplesoft> i have place files in /var/www
[15:25] <jamespage> maplesoft, 'It works!' is the default page for apache2
[15:25] <maplesoft> yes.
[15:25] <jamespage> which would be configured to serve files from /var/www
[15:28] <maplesoft> jamespage mysite.com:8080 gives me the file in /var/www
[15:29] <maplesoft> jamespage what is ihappening
[15:30] <jamespage> maplesoft, from the information you have provided I have no idea
[15:30] <jamespage> take a look in /var/log/apache2 at the access log - this might tell you something
[15:30] <maplesoft> jamespage ok. where the default file location for tomcat is confgured/changed?
[15:30] <hallyn> smb: hi, is nfs expected to work in quantal right now?
[15:31] <hallyn> i've got two instances running latest quantal.  /etc/exports has '/mnt <ipaddr>(rw,no_root_squash,no_subtree_check)', did exportfs -a, but the other host hangs when i try mount -t nfs -o rw ip:/mnt /mnt ...
[15:31] <jamespage> maplesoft, I think you are trying to think of tomcat like apache
[15:31] <jamespage> its quite different
[15:31] <hallyn> i know we've switch to nfs by default being nfsv4, but...
[15:32] <jamespage> everything gets deployed as web applications (.wars)
[15:32] <beandog> hallyn the client mounting stuff should see something in system logs
[15:32] <jamespage> maplesoft, which reside in /var/lib/tomcat6/webapps
[15:32] <beandog> or the server might, too
[15:32] <jamespage> maplesoft, the ROOT directory containing the context that maps to /
[15:32] <hallyn> beandog: doesn't.  just get 'timed out' after awhile on the client
[15:33] <jamespage> if you want to deploy .jsp's they must be contained in a web application under that directory
[15:33] <beandog> where?
[15:33] <jamespage> maplesoft, each web application sits on a 'context' - for example /var/lib/tomcat6/webapps/ubuntu-test.war would sit under the /ubuntu-test context by default
[15:34] <jamespage> make sense?
[15:34] <maplesoft> jamespage where do i need to put the .war file/
[15:34] <jamespage> maplesoft,  /var/lib/tomcat6/webapps/
[15:35] <jamespage> maplesoft, I think it can also just be a directory within that path - I don't normally deploy that way.
[15:35] <maplesoft> jamespage iam following http://www.igniterealtime.org/projects/openfire/plugins/webchat/readme.html
[15:35] <hallyn> d'oh, server reboot fixed it
[15:36] <maplesoft> i should not put it in /var/lib/tomcat7/webapps/ROOT        but in /var/lib/tomcat7/webapps/someDir?
[15:36] <maplesoft> jamespage ^
[15:37] <jamespage> maplesoft, probably - assuming 'it' is the webchat application and it is a war file
[15:38] <maplesoft> jamespage but any way. why is the file index.html in /var/lib/tomcat7/webapps/ROOT    not showing up on mysite.com:8080      and why the file index.html on /var/www show up when apache is stopped
[15:39] <jamespage> maplesoft, I really have no idea - have you tried restarting tomcat7?
[15:39] <maplesoft> mm. no
[15:40] <jamespage> it might be you are getting a cached version of something...
[15:40] <jamespage> just a thought
[15:45] <jamespage> utlemming, bug nominations accepted
[15:45] <utlemming> mr jamespage: most appreciated
[15:55] <maplesoft> jamespage well its pointing at /var/www
[15:56] <maplesoft> jamespage where is the config to change the default directory for tomcat7 ?
[16:01] <jamespage> ttfn
[16:21] <jamespage> rbasak, I think we need to re-consider the approach for the apache2 SRU/bug 988819
[16:21] <uvirtbot`> Launchpad bug 988819 in mod-proxy-html "[SRU] wrong path to libxml2.so.2 in mod_security - broken by multiarch enabled libraries" [Undecided,Confirmed] https://launchpad.net/bugs/988819
[16:21] <jamespage> if the minimal change to make stuff work is to drop the LoadFile stanza's in the config files that should be the fix.
[16:22] <rbasak> jamespage: for the SRU, and leave Quantal alone (same as Debian)?
[16:22] <jamespage> rbasak, yes
[16:22] <rbasak> jamespage: that's fine with me
[16:22] <jamespage> OK - lets check with SpamapS how we get whats in -proposed dropped....
[16:23] <SpamapS> jamespage: very easily
[16:28] <SpamapS> rbasak: are you saying that bug 988819 is "verification-failed" and that rather than uploading a new fix on top of that one, you'd like it removed from precise-proposed ?
[16:28] <uvirtbot`> Launchpad bug 988819 in mod-proxy-html "[SRU] wrong path to libxml2.so.2 in mod_security - broken by multiarch enabled libraries" [Undecided,Confirmed] https://launchpad.net/bugs/988819
[16:29] <rbasak> SpamapS: AIUI, it is difficult to verify, so the approach to fix is being reconsidered. So I guess that means I need to get the alternative done and uploaded to precise-proposed?
[16:30] <SpamapS> rbasak: its easier to track if we do it that way really.
[16:31] <SpamapS> rbasak: I'll mark it 'verification-failed' so it doesn't accidentally get sent to -updates
[16:31] <rbasak> SpamapS: OK
[16:31] <rbasak> SpamapS: I take it I still need to bump the version further though, as it's been in precise-proposed?
[16:32] <SpamapS> rbasak: yes
[16:32] <SpamapS> rbasak: -proposed counts as "being published"
[16:33] <rbasak> SpamapS: ack
[16:36] <nandersson> Hi, I would like to have some suggestions here. I am doing some packaging. I am installing a virtual machine (ova-format) using apt, but afterwards I would like to have the user automatically register that machine in Virtualbox. Is there a "best practice" as how to make the user aware after a package install that he must take action?
[16:37] <nandersson> The thing is that apt-get is run as root, and it is the user that must register the vm afterwards.
[16:38] <nandersson> I could probably hack something using /etc/bash.bashrc but I guess it wouldnt be standard, and it wouldnt look good. Quite the contrary it would be a pretty ugly solution.
[16:40] <nandersson> Is there some standard way to tell the user that he has to take action? Some kind of hook.
[16:42] <rbasak> nandersson: I'm not sure this is the right channel. Most server people don't use GUIs.
[16:42] <rbasak> nandersson: if I understand you, you have more of a desktop question?
[16:43] <rbasak> nandersson: or are you trying to do this on a terminal login? There is an motd system.
[16:48] <nandersson> rbasak, well, it is more of a packaging question. Let say that after I do "sudo apt-get install my-package" - I also want to touch files in /home/my-user/ directory. Well, Perhaps I could do it in my-package.postinstall and have the script iterate over all directories in /home and make the change, or set a flag...
[16:49] <nandersson> ...or have the script "su - my-user" and do the registration.... I guess that would do the job.
[16:50] <rbasak> nandersson: see https://lists.ubuntu.com/archives/ubuntu-devel/2012-July/035546.html which has to solve a similar problem
[16:50] <rbasak> nandersson: it won't work for you, but generally postinsts aren't really support to walk over home directories like that
[16:50] <jamespage> rbasak, I have another ARM kernel module request for omap4 (and others) - rbd
[16:51] <nandersson> rbasak, Yeah, it will be darn ugly but I think it would work
[16:51] <rbasak> nandersson: only when users are using a local db. If they're on LDAP it'll break
[16:51] <rbasak> nandersson: or if there's some kind of dynamic home directory mounting going on
[16:51] <nandersson> rbasak, yeah, you are right about that :-/
[16:52] <rbasak> anyway, I still don't think it's really a server question. Try #ubuntu-devel - you might get more relevant ideas!
[16:52] <nandersson> (y) I'll do that! Thanks a lot!
[16:52] <jamespage> nandersson, maybe a README.Debian might be a good approach - let users know which commands they need to execute to use the package fully
[16:53] <rbasak> nandersson: or perhaps #ubuntu-app-devel would be more appropriate here
[16:53] <nandersson> rbasak, I'll try both!
[16:53] <rbasak> !crosspost | nandersson
[16:54] <rbasak> jamespage: I'm not sure who to direct ARM kernel requests to, actually
[16:54] <jamespage> rbasak, me neither :-)
[16:54] <nandersson> jamespage, Yeah, if the users would be versed in Ubuntu... I am trying to do a "fool proof" solution here :)
[16:54] <rbasak> jamespage: perhaps smb, in the kernel section of the meeting?
[16:54] <rbasak> smb: who's the right contact for ARM kernel issues?
[16:55] <rbasak> smb: (which aren't vendor-specific)?
[16:55] <jamespage> rbasak, I just asked in -kernel
[16:55] <jamespage> SpamapS, did you get anywhere with the fastcgi-developers mailing list?
[16:57] <SpamapS> jamespage: not yet no
[16:59] <SpamapS> jamespage: just sent another message to the admin listed on the listinfo page.
[16:59] <SpamapS> jamespage: If they're not responsive, I think we should suggest that the ceph team maintain a fork
[17:03] <jamespage> SpamapS, I've pinged the debian maintainer as well - you never know....
[17:05] <SpamapS> jamespage: good idea
[17:23] <pmatulis> strange, i install cobbler on 12.04 and syslinux is configured with 'Ubuntu MAAS'
[17:24] <roaksoax> pmatulis: yeah, we need to drop that :)
[17:24] <roaksoax> pmatulis: could you please file a bug and assign it to me?
[17:24] <pmatulis> roaksoax: yes, ok
[17:30] <\sh> kklimonda: ping freeipa..how can we add sudo-ldap support to the package without breaking the system a lot ? :)
[17:32] <uvirtbot`> New bug: #990162 in python-greenlet "armhf version is unusable -- task switches will cause sigsegv's" [Undecided,In progress] https://launchpad.net/bugs/990162
[17:35] <hggdh> Daviey: (following from -meetings): not really. We do not yet run automated arm tests.
[18:04] <stgraber> hallyn: so, I'm looking at implementing get_config_item in the API but the code doesn't seem to match what you mentioned earlier on IRC :)
[18:04] <stgraber> 21:04 < hallyn> returning the length, so you can do 'len = get_config_item(c, key, NULL); v = malloc(len+1); len = get_config_item(c, key, v)'
[18:04] <stgraber> though it doesn't look like I can actually get the length with the current implementation
[18:04] <stgraber> and looking at your tests, you seem to just be using fixed length char arrays?
[18:05] <hallyn> if key is valid you should get back len
[18:05] <hallyn> and the tests do
[18:06] <hallyn> what are you getting back?
[18:08] <stgraber> hallyn: hmm, can't see a test that matches my understanding of it. I was expecting to do len = get_config_item(c, key, NULL, NULL), then malloc and use get_config_item(c, key, value, len)
[18:08] <hallyn> i don;t malloc in the tests
[18:08] <hallyn> but that's irelevant :)  i print out the return values
[18:09] <hallyn> and they were correct lengths for me
[18:09] <stgraber> hmm, indeed, I just grepped and was surprised not to find a malloc in there, let me actually test my code, maybe it's working fine ;)
[18:09] <hallyn> theres 1 test passing in null v as well
[18:19] <Eitan> hello all
[18:20] <Eitan> i am hammering a new ubuntu server with tons of connections, i was running centos for my application before with no problems... For some reason i am getting a lot of connection time outs when establishing connections both mysql 3306 and redis 6397... first i found that syn flood protection was blocking some stuff, so i managed to turn that off. but im not sure to see where else or why
[18:20] <Eitan> else connections are timing out as far as logs or perhaps if its a ulimit issue?
[18:23] <stgraber> hallyn: http://paste.ubuntu.com/1097049/ <- I have the feeling I did a pretty obvious mistake but it's not really obvious to me at the moment...
[18:25] <aaas> any way i can benchmark my php performance?
[18:35] <hallyn> stgraber: and what does the first call to getitem give you?  what len?
[18:40] <stgraber> >>> test.get_config_item("lxc.utsname")
[18:40] <stgraber> got len: 4
[18:40] <stgraber> Traceback (most recent call last):
[18:40] <stgraber>   File "<stdin>", line 1, in <module>
[18:40] <stgraber> UnicodeDecodeError: 'utf-8' codec can't decode bytes in position 0-2: invalid continuation byte
[18:41]  * stgraber adds a printf to also dump the value before the unicode conversion
[18:41] <hallyn> stgraber: i wonder if you need a longer char * for PyUnicode_FromString
[18:42] <hallyn> stgraber: note that depending on what PyUnicode_FromString does, you ouht to free the char *value you mallocd
[18:42] <stgraber> hallyn: printf doesn't seem to be much more happy...
[18:42] <stgraber> >>> test.get_config_item("lxc.utsname")
[18:42] <stgraber> got len: 4
[18:42] <stgraber> got value: �1i
[18:43] <cwesterfield> Can someone tell me how ubuntu server is sourcing PATH? I need to add to it, but can't find it
[18:43] <stgraber> cwesterfield: /etc/environment?
[18:44] <cwesterfield> that's it
[18:44] <cwesterfield> google is sometimes a hindrance
[18:46] <hallyn> stgraber: what is lxc.utsname's actual value?
[18:46] <stgraber> hallyn: considering the garbage, it's pretty likely that my malloc is the part that's wrong but I'm not sure how/why
[18:46] <stgraber> hallyn: "test"
[18:47] <hallyn> stgraber: d'oh
[18:47] <hallyn> i shouldve looked closer at your pastebin :)
[18:47] <hallyn> hm no, stil looking
[18:49] <hallyn> stgraber: before calling the PyUnicode_FromString, can you do 'printf("key %s len %d value %p %s\n", key, len, value, value);
[18:50] <stgraber> >>> test.get_config_item("lxc.utsname")
[18:50] <stgraber> key lxc.utsname len 4 value 0x1b7da30 �M�
[18:51] <hallyn> stgraber: is the hostname actually 3 byts?
[18:52] <stgraber> hallyn: no, it's 4, though the malloc is for len + 1, so should be fine?
[18:53] <hallyn> stgraber: can you pass 'len+1' in to the second lxc_get_item call?
[18:54] <hallyn> stgraber: it's a bit confusing, but since i must have room for the trailing \0, i insist on having len+1 available to copy into
[18:54] <hallyn> confusing because i pass len back...  but i f i pass back len+1 that's confusing too
[18:54] <stgraber> hallyn: that worked indeed
[18:55] <hallyn> somebody with a silver tongue is gonna need to write a terrific manpage for this :)
[18:55] <hallyn> ^ note i disqualified myself
[18:55] <uvirtbot`> hallyn: Error: "note" is not a valid command.
[18:55] <stgraber> ;)

[18:55] <hallyn> ^ help
[18:55] <uvirtbot`> hallyn: (help [<plugin>] [<command>]) -- This command gives a useful description of what <command> does. <plugin> is only necessary if the command is in more than one plugin.
[18:55] <hallyn> ! help
[18:55] <hallyn> dang how many bots we got here
[18:56] <stgraber> maybe adding an example of the malloc to the tests might help, I'm assuming I won't be the only one to copy/paste from it :)
[18:56] <hallyn> stgraber: ok will do so right now
[18:58] <stgraber> what's looking quite weird is the check (get_config_item(c, key, value, len + 1) != len) when len is the value you get from an identical call to get_config_item ;)
[18:58] <stgraber> but yeah, it makes sense when you know ;)
[18:58] <hallyn> luckily everyone will use the python api :)
[19:03] <hallyn> stgraber: all right, pushed.  hopefully it passes.  (no vm at the ready to test on)
[19:06] <stgraber> hallyn: can't remember, did we settle on having a way to query the list of keys set for a given network entry or not?
[19:10] <hallyn> stgraber: not exactly, so for now you can look at the lxc.network.0.type and know what's valid based on that
[19:10] <hallyn> hardcoded, but then i don't expect veth.peer to every become a valid key for macvlan, for instance :)
[19:10] <stgraber> ok
[19:11] <hallyn> stgraber: if you feel better about it, i can go ahead and ipmlmenet c->getkeys(),
[19:11] <hallyn> and have 'lxc.network.0' return the valid list
[19:12] <stgraber> hallyn: a few quick notes that I noticed while testing .get_config_item("lxc.network") is terminated by a \n, is that intended (I'd have expected \n separated, not \n terminated)
[19:12] <hallyn> (c->getkeys(c, "lxc.network.0", value, len)
[19:12] <stgraber> hallyn: +1 for getkeys() it'd make my life quite a bit easier ;)
[19:12] <hallyn> stgraber: i waffled on the trailing \n.  you want me to skip that for the last one?
[19:12] <stgraber> (or rather, make my future life easier by not having to update the hardcoded logic in python)
[19:12] <hallyn> that'll mean that if there is only a single entry, there'll be no \n
[19:13] <hallyn> it seemed inconsistent to me, but maybe it works nicer with str.split('\n') <shrug>
[19:13] <stgraber> hallyn: yeah, I'd prefer no trailing \n as it'd indeed work better with split() :)
[19:13] <hallyn> all right let me look through some remaining unreproduced bugs then i'll implement those two
[19:14] <hallyn> bbl
[19:14] <hallyn> stgraber: perhaps we should keep a wiki page with api todo's
[19:14] <stgraber> hallyn: I was also somewhat surprised to see that lxc.cap.drop gets me a space separated list of values instead of separate lines like .network and it also seems to be fine having duplicate entries :)
[19:15] <stgraber> hallyn: though I guess this kind of makes sense as it can be stored either as a single line or multiple lines in the config
[19:15] <stgraber> +1 for wiki
[19:15] <hallyn> right
[19:15] <hallyn> i did space separate bc it works in config :)
[19:17] <stgraber> can you make it \n separated? I think it'd make it less confusing for people doing
[19:17] <stgraber> get_config_item("lxc.cap.drop")
[19:17] <stgraber> getting = cap_sys_admin cap_net_admin
[19:17] <stgraber> then doing set_config_item("cap_net_admin")
[19:17] <stgraber> and expecting it to be set to "cap_net_admin"
[19:18] <stgraber> instead of "cap_sys_admin cap_net_admin cap_net_admin"
[19:18] <stgraber> making it multi-line would be a good indication that you need to clear_config_item() first
[19:19] <hallyn> stgraber: https://wiki.ubuntu.com/lxc/apitodos
[19:19] <hallyn> stgraber: ok
[19:27] <\sh> did anybody played around with freeipa-client lately? ipa-getkeytab doesn't work :*
[19:36] <stgraber> hallyn: is there an easy way for me to get the list of all keys accepting multiple values and requiring the use of clear_config_item?
[19:37] <hallyn> stgraber: nope
[19:37] <stgraber> hallyn: my plan is to have any of these key return a python list and have set_config_item(key, [list]) automatically call clear_config_item and set_config_item for each of them
[19:38] <hallyn> (apart from looking at src/lxc/confile.c:lxc_clear_config_item())
[19:39] <hallyn> stgraber: perhaps get_item("lxc") should return a list like "cgroup 1", "caps 1", where a 0 would mean not a list?
[19:39] <hallyn> don't really like that, but not sure what better to give you
[19:39] <stgraber> hallyn: oh, actually, there's an easy trick but that's going to make you revert one of your current change :)
[19:39] <stgraber> hallyn: can you always add a trailing \n to these that are a list?
[19:40] <hallyn> yes.  note though that not everything returning a list can be cleared
[19:40] <hallyn> i.e. lxc.network.0.ipv4
[19:40] <stgraber> hmm, why?
[19:41] <hallyn> i dunno, the conf stores it as a list
[19:41] <hallyn> so aparently you can have multiple values
[19:42] <stgraber> so if lxc.network.0.ipv4 returns "1.2.3.4\n4.5.6.7\n" how can I set it to "8.9.10.11"?
[19:42] <hallyn> clear all the networks
[19:43] <hallyn> more to the point what does it mean to have 2 values
[19:44] <hallyn> it may just be a relic of his early attempts at wrestling with the list api
[19:44] <stgraber> two IPs on eth0?
[19:44] <hallyn> yes
[19:44] <stgraber> I'd expect it to be the equivalent of calling "ip -4 add <ip> dev eth0" twice
[19:45] <hallyn> and what does that do?
[19:45] <stgraber> put two addresses on the same interface
[19:45] <RoyK> hrmf - http://paste.ubuntu.com/1097181/ <-- I have a job moving md0 from raid-5 to raid-6, but I have no idea whatsoever what's causing the traffic on the root (md3/sdh) - any idea what that might be?
[19:46] <stgraber> which is a lot more common with IPv6 than IPv4, but still a pretty common usecase
[19:46] <hallyn> stgraber: so you'd like a clear_item("lxc.network.0.ipv4") ?
[19:47] <stgraber> ideally, I'd like clear_config_item on anything that's a list (where I know whether it's a list or not based on the last char of the matching get_config_item being \n)
[19:47] <hallyn> stgraber: so should get_keys() just return "key1 is_list\nkey2 is_list\nkey3 is_list" (is_list = 'Y' or 'N')?
[19:48] <hallyn> all right i can put a \n after the last one
[19:49] <stgraber> I think I'd prefer to rely on get_config_item to know whether something is a list or not and just use get_keys() to know what keys are set under a given path (lxc.network.0 for example)
[19:51]  * stgraber is starting to have the feeling we might actually get something that's pretty consistent and not too hackish ;)
[19:53] <elguapo99> how do I add tab autocomplete to my ssh account?
[19:53] <RoyK> elguapo99: it should be there by default, if you use a shell supporting it
[19:54] <RoyK> elguapo99: ps $$
[19:54] <elguapo99> I have tab autocomplete as su
[19:54] <stgraber> RoyK: doesn't that require HashKnownHosts to be set to no? otherwise I don't see how you could complete the hosts
[19:54] <elguapo99> but not as my ssh login
[19:54] <raubvogel> I have a machine here that is supposed to be dhcp and dns for a subnet. When I boot a machine to the subnet in dhcp mode, it is picking an IP (and I can see in the leases file it giving its hostname), but the dns is not being updated
[19:55] <RoyK> elguapo99: on which shell are you running that account?
[19:55] <elguapo99> I think its bash
[19:55] <elguapo99> $
[19:55] <RoyK> ps $$
[19:55] <RoyK> that shows the shell
[19:56] <elguapo99>  7701 pts/2    Ss     0:00 -sh
[19:56] <guntbert> raubvogel: does the dhcp server point the clients to the dns server?
[19:56] <RoyK> elguapo99: try chsh -s /bin/bash
[19:57] <RoyK> elguapo99: as that account
[19:57] <elguapo99> The problem was that I was not doing bash
[19:57] <RoyK> bingo
[19:57] <elguapo99> so I have to type bash when I first log in
[19:57] <raubvogel> guntbert: yep
[19:57] <RoyK> elguapo99: see the chsh command above
[19:57] <raubvogel> and I even checked the client's /etc/resolv.conf
[19:58] <raubvogel> But, even nslookup fqdn dns.ip.address does not work
[19:58] <guntbert> raubvogel: if you put your dhcpd.conf into a pastebin we could have a look at it
[20:01]  * RoyK goes to get some sleep - nite...
[20:01] <raubvogel> guntbert: http://pastie.org/private/ja7zx5gom4rtqtkijtidnw
[20:01] <raubvogel> (using external ntp server)
[20:02] <guntbert> raubvogel: looks good (aka I can't spot an error :-)) - do the clients pick up the ntp sevrer?
[20:04] <raubvogel> I honestly have not checked because I need to deal with my dns/dhcp issue and then why the firewall (ubuntu 12.04) iptables are inop
[20:05] <raubvogel> guntbert: My main issue with dns is the hostnames for the dhcp clients are not being added to bind
[20:05] <guntbert> raubvogel: my question was meant to help find the bug
[20:06] <guntbert> raubvogel: for the dhcp->bind wouldn't you need a key for authorization?
[20:07] <raubvogel> I understand; I am just a bit in a panic mode right now. I do not even can think of a good way to see if the clients see ntp. They do get the nameserver though
[20:07] <RoyK> raubvogel: what clients? dhcp?
[20:08] <raubvogel> RoyK: exactly
[20:08] <RoyK> what client OS?
[20:08] <guntbert> raubvogel: ahh - I misread - now I understand the issue - well for the updating of the DNS databse you will need some kind of authorization
[20:09] <raubvogel> ubuntu 12.04 just like the dns/dhcp server
[20:09] <RoyK> raubvogel: I'd think setting up an ntp server listening to broadcasts/multicasts on that network will do a lot - IIRC the clients will attempt to sync to such a machine first
[20:09] <guntbert> RoyK: only 7 minutes sleep? :-))
[20:10] <RoyK> raubvogel: I don't think the ubuntu dhcp client can change the client's ntp settings ;)
[20:10] <RoyK> guntbert: no, have to wait a while for some laundry...
[20:10] <raubvogel> RoyK: should I drop ntp for now and focus on the other issue?
[20:11] <RoyK> raubvogel: at least drop it in dhcp - rather setup ntp server on that server of yours and allow broadcast/multicast requests - it's rather easy, and I *beleive* the clients will try that first. if the clients connect directly to ubuntu's ntp server, it won't make much difference in the network traffic, though
[20:12] <guntbert> raubvogel: lets clarify the issue: as far as the clients are concerned all is well, what bothers you is that the DNS server is not updated when a client get an address?
[20:12] <raubvogel> guntbert: exactly
[20:12] <RoyK> raubvogel: that's dynamic dns
[20:12] <RoyK> not enabled by default
[20:13] <raubvogel> RoyK: exactly. How do I enable it? I thought I have done but I probably overlook something
[20:13] <guntbert> raubvogel: if I recall correctly you have to configure both servers to talk to each other and the DNS to trust the DHCP - therefor you need keys
[20:13] <RoyK> raubvogel: google it
[20:13] <raubvogel> guntbert: they are actually the same machine
[20:14] <guntbert> raubvogel: doesn't matter, two different programs
[20:14] <maplesoft>  have not been getting any help since a long time on this .  my tomcat is pointing towards /var/www instead of /var/lib/tomcat7/webapps/ and if i place a .jsp in /var/www   it starts downloading it instead of parsing it.  i had apache2 installed on my ubuntu but its now turned off.    what is happening?
[20:15] <guntbert> raubvogel: as of now you only told dhcpd what kind of updates to send
[20:16] <chiggins>  Hey I'm using likewise-open to bind my desktop to my domain, but my user doesn't show up in unity-greeter. How can I have it show up?
[20:20] <RoyK> hm...
[20:20] <RoyK> [108984.234696] ata4.00: failed command: READ FPDMA QUEUED
[20:20] <RoyK> any idea how to map ata4.00 to a device name?
[20:21] <_ruben> RoyK: i posted that question quite some time ago on askubuntu and got some nice "oneliners" for that
[20:23] <_ruben> RoyK: http://askubuntu.com/questions/64351/how-to-match-ata4-00-to-the-apropriate-dev-sdx-or-actual-physical-disk
[20:25] <raubvogel> guntbert, RoyK: I thought when I defined the ips I have in allow-update http://pastie.org/private/kddg4xci8uihvaq3fzeokq, in essence I am defining who machines can do the updating.
[20:26] <raubvogel> Of course, since this machine has 3 ips (localhost, static internal, dhcp external) I wonder if that sufficed
[20:45] <Eitan> where would be the place to check a log to see if ubuntu server was denying connections on port 3306 for mysql because there are to many active
[20:46] <Eitan> im getting connection timeouts on my clients, but i cant find these connections anywhere on my db server
[20:47] <raubvogel> Ok, dhcp-dns issue solved. Onwards to firewall!
[21:00] <beandog> Eitan: was it previously listening remotely?
[21:04] <RoyK> _ruben: thanks - guess I'll let this conversion to raid-6 go finish and then start testing...
[21:06] <raubvogel> Shouldn't something like 'iptables -A INPUT -p tcp -m tcp --dport 443 -m state --state NEW,ESTABLISHED -j ACCEPT' in a firewall allow a machine behind it to reach a website?
[21:19] <hallyn> all right i'll be back later to work on some apitodos
[21:29] <maplesoft> jamespage  i have installed tomcate by tar. and still having same problem. its pointing to /var/www when i use mysite.com/8080
[22:31] <raubvogel> Wouldn't line 21 in http://pastie.org/private/zlkonugv9eyij3km5jb8a allow me to connect to that site on port 80?
[22:38] <bonez2046> after I upgraded from 10.04 to 12.04 when I boot my machine I get..."Waiting for Network Configuration"..and another advising Waiting 60 seconds, etc...yet, I have connectivity.. how do I prevent that message at boot up?
[22:38] <bonez2046> I installed ubuntu 12.04 LTS, not the server version.... but i get that message
[22:40] <peterrus> so there should be some way to disable that waiting
[22:40] <bonez2046> that's what I'd like to find
[22:40] <bonez2046> in fairness though, before I upgraded, if I blew the password, the way my system was configured I'd have to wait at least 2 minutes before I could reenter the password
[22:42] <peterrus> this might be something
[22:42] <peterrus> http://tech.pedersen-live.com/2012/05/disable-waiting-for-network-configuration-messages-on-ubuntu-boot/
[22:43] <peterrus> what you should look for is a way to disable that waiting
[22:43] <peterrus> still strange how it found its way to your regular ubuntu installation
[22:45] <\sh> grmpf
[22:45] <\sh> do we have a libldap package which is not build against libgssapi3-heimdal?
[22:45] <\sh> but against libgssapi-krb5?
[22:46] <bonez2046> tell me about it.
[22:46] <bonez2046> its' frustrating
[22:49] <bonez2046> thanks, I'll try that, very much obliged
[23:29] <koolhead11> jcastro, ping