=== dendrobates is now known as dendro-afk | ||
=== dendro-afk is now known as dendrobates | ||
=== dendrobates is now known as dendro-afk | ||
bobsomebody | im trying to figure out what would cause mysql "Can't connect to local MySQL server through socket" error to be produced | 01:33 |
---|---|---|
bobsomebody | im benchmarking a script that needs to be scalable and I initiaded 300 mysql_connects in php, i get about 30 of the above errors | 01:34 |
twb | The mysql client has this stupid thing where connecting to "localhost" is a special case that makes it use sockets. | 01:38 |
twb | That's the only time I've seen that error. | 01:38 |
kirkland | smoser: around? | 01:43 |
kirkland | smoser: does this look familiar? * Excellent, working condition, nothing wrong with it | 01:43 |
kirkland | smoser: sorry ... wrong paste | 01:43 |
kirkland | smoser: http://pastebin.ubuntu.com/383371/ | 01:43 |
bobsomebody | twb, what would you recommend as a 'fix'? | 01:43 |
twb | bobsomebody: you connect to 127.0.1.1 or something that's not hard-coded to be a stupid special case | 01:43 |
twb | Or better: use postgres or sqlite, depending on your use case. | 01:44 |
bobsomebody | twb, i cannot believe "that" cost me 6hrs of my life | 01:51 |
bobsomebody | ty none the less | 01:52 |
twb | bobsomebody: that was all it was? | 01:52 |
twb | Christ. F*cking mysql. | 01:53 |
bobsomebody | yes | 01:53 |
bobsomebody | my sentiments exactly... | 01:54 |
twb | So yeah, I reiterate my recommendation of postgres/sqlite | 01:57 |
kirkland | smoser: looks like https://bugs.edge.launchpad.net/ubuntu/+source/cloud-init/+bug/525675 again :-/ | 02:00 |
uvirtbot | Launchpad bug 525675 in python-boto "20100222 images fail to boot in UEC (HTTP error 500 retrieving ephemeral0 metadata)" [Wishlist,Confirmed] | 02:00 |
bobsomebody | twb, ive considered it many times | 02:04 |
bobsomebody | twb, this one requires mysql by the client | 02:06 |
twb | I try to educate my clients :-) | 02:07 |
bobsomebody | me 2, im just a little uneducated in this area | 02:08 |
bobsomebody | ok back to work it seems, thank you again twb, i will make it a point to read up on postgres | 02:16 |
Roxyhart0 | hi i am creating folder to new user and it is creating with owner dnslog | 02:23 |
Roxyhart0 | somebody know when it happen? | 02:24 |
axisys | uname -a does not tell me if i am running a server or desktop .. lsb_release -a does not disclose it either .. how do i find it ? | 02:49 |
zul | happy happy joy joy | 02:51 |
jasonmchristos | hi i already have installed ubuntu cloud, 1 controller and 1 node, my goal is to install elg on a virtual server running on top of the cloud how do i now install a virtual server to the cloud? | 02:55 |
jasonmchristos | im guessing somehow i need to create an image by installing elgg on a regular lamp server then duplicating the image and installing it to the cloud, am i correct?? | 02:56 |
Roxyhart0 | somebdy know why when i create a new account on ldap it doesn work..if i add permisions to some folder it take dnslog and not the user | 02:57 |
twb | Are the xtables kernel modules part of the -generic kernel packaging? | 02:59 |
twb | I can see xtables-addons-source, but no pre-rolled packages for the official kernels | 02:59 |
axisys | one of my ubuntu is not telling me in motd, when pkg is available .. rest tells me when new upgrade is available.. where do i look to fix it? | 03:05 |
kirkland | axisys: sudo apt-get install update-notifier-common | 03:06 |
axisys | kirkland: thanks a lot | 03:07 |
axisys | kirkland: any idea why other servers dont have that pkg but still motd is dynamic with new upgrade | 03:09 |
kirkland | axisys: what version of ubuntu server are you running? | 03:09 |
twb | WTF | 03:09 |
axisys | kirkland: never mind .. i am wrong about it | 03:09 |
twb | xtables-addons-common: /lib/xtables/libxt_CHAOS.so | 03:09 |
axisys | kirkland: i do have the -common | 03:09 |
kirkland | axisys: ls -alF /etc/update-motd.d/ | 03:09 |
kirkland | axisys: that should show you the scripts that generate your motd | 03:10 |
axisys | kirkland: yes i have that dir now.. thanks | 03:10 |
cef | doesn't the "packages to update" stuff belong to landscape-common? | 03:11 |
cef | hang on.. could be thinking about a different thing | 03:12 |
twb | "The actual locale definitions are not part of this package, these are shipped in the language packs and are installed and removed automatically." | 03:12 |
twb | Grumble -- that'd explain why my live image takes so flipping long to build en_US.UTF-8 during boot | 03:13 |
twb | It's probably doing an apt-get update and apt-get install in the background | 03:13 |
axisys | kirkland: according to uname -a this one server is not a server , but a desktop .. is there a easy way to convert a desktop to a server.. if it too much i dont care.. essentially they are same without the WM .. correct ? | 03:15 |
kirkland | axisys: more or less | 03:15 |
kirkland | axisys: you'd need to remove a ton of desktop packages, and install the linux-server kernel | 03:15 |
cef | axisys: about the only difference that comes to my mind is the kernel | 03:16 |
twb | kirkland: depends if you pick "desktop" in tasksel during the install | 03:16 |
twb | IIRC the server CD's default preseed also disables installation of localization packages. | 03:17 |
Graflan | Is there anyone here thats familiar with streaming VLC? | 03:19 |
jasonmchristos | Graflan: ive done it b4 | 03:28 |
jasonmchristos | using the gui | 03:28 |
Graflan | I can get it to stream without a problem but when I want to stop it I can't seem to figure it out | 03:28 |
Graflan | I'm doing it with the command line | 03:29 |
jasonmchristos | run an http controller | 03:34 |
jasonmchristos | and log in and it will look just like the gui and hit stop | 03:34 |
maxagaz | hi | 03:37 |
maxagaz | can I use cp with a list of files returned by a pipe ? | 03:38 |
maxagaz | ... | cp -r > my.directory/ | 03:38 |
persia | No. | 03:38 |
persia | But you can cp -r $(...) my.directory/ | 03:38 |
persia | (subshell vs. pipe) | 03:39 |
persia | Note that there exist (large) limits to the number of entries in a single command line. | 03:40 |
maxagaz | persia, what if the files are given one by line, and not separated with spaces ? | 03:40 |
persia | maxagaz: Doesn't matter. $() converts all classes of whitespace into spaces. | 03:41 |
persia | Note that this does break down in cases where you are dealing with filenames containing whitespace. | 03:41 |
maxagaz | persia, great... thanks a lot! | 03:41 |
Graflan | I figured out one thing I was doing wrong, I hit ctrl+z instead of ctrl+c ....z just suspends it | 03:45 |
jasonmchristos | does the karmic uec image have lamp like the standard iso? | 03:57 |
jasonmchristos | in otherwords is the karmic uec identical? | 03:57 |
jasonmchristos | uec image i mean | 03:57 |
axisys | hmm so i should tasksel to pick ubuntu-server .. will that work.? | 04:08 |
axisys | kirkland, cef ^ | 04:08 |
axisys | twb: i jumpstarted and probably picked generic image | 04:09 |
twb | axisys: that would depend on what yo uwant | 04:09 |
axisys | twb: well.. it gets annoyed when motd says i need to upgrade openoffice .. stuff i would not get if i had installed server version.. so want to avoid any desktop app to take up server resource.. i am runnign it on x4120 .. so i have a power machine.. so i guess i could just leave it alone | 04:11 |
axisys | and dont mess with removing desktop image and put server image | 04:11 |
twb | axisys: that's because the default behaviour pulls in localization, which (stupidly) includes oo.org dictionaries. | 04:12 |
axisys | twb: gotcha | 04:12 |
twb | axisys: unless you seriously screwed up, you shouldn't have OO itself installed | 04:12 |
twb | axisys: just browse to the "localization" section in aptitude and purge the relevant -write packages. | 04:12 |
axisys | correction: it is a sun x4100 | 04:13 |
axisys | twb: ok | 04:13 |
Sakara | just performed a dist-upgrade on a server I have and now the system has become unbootable | 04:14 |
axisys | sudo aptitude search localization shows `p localization-config .. ' | 04:14 |
Sakara | Grub Error 15 is the error message I recieve | 04:14 |
axisys | Sakara: i can grub-install could fix that w/ a live cd | 04:15 |
Sakara | I tried to dpkg-reconfigure grub from a chroot with a live cd with no success | 04:15 |
axisys | s/i can/i think/ | 04:15 |
Sakara | axisys: grub-install is a program? | 04:15 |
axisys | Sakara: yes part of grub pkg | 04:16 |
Sakara | fdisk -l shows under the system heading that the partition is Linux LV | 04:17 |
Sakara | will that be a major problem/headache for grub? | 04:17 |
axisys | Sakara: grub | 04:22 |
axisys | Sakara: grub> find /boot/grub/stage1 | 04:23 |
twb | GRUB Legacy doesn't understand LVM. | 04:23 |
axisys | wait .. if you are using grub2 there is no stage1 i dont think in that folder | 04:24 |
Sakara | ok will try chroot and apt-get install grub2 | 04:24 |
axisys | Sakara: i found the steps to fix it | 04:28 |
axisys | Sakara: let me know if you need it | 04:28 |
Sakara | axisys: I would like that alot | 04:31 |
Sakara | just fired up a live cd session and chrooting atm | 04:32 |
axisys | Sakara: https://wiki.ubuntu.com/Grub2 go to section "Recover Grub 2 via LiveCD" | 04:32 |
axisys | then you may want to visit the "Error 15 - File not found" section under "Errors" | 04:33 |
Sakara | yup I did have a read of that | 04:36 |
Sakara | have done it a couple of time before on other broken systems | 04:36 |
axisys | Sakara: heh | 04:36 |
Sakara | rebooting now with grub2 install | 04:36 |
Sakara | looking good | 04:37 |
Sakara | its firing up now | 04:37 |
Sakara | yup fixed | 04:37 |
Sakara | chroot into the system and apt-get install grub2 fixed the error 15 problem | 04:38 |
axisys | sweet! | 04:40 |
jasonmchristos | on a lamp server where is the default www directory? | 05:58 |
pwnguin | /var/www | 05:58 |
pwnguin | you can also check the apache conf files for clues | 05:58 |
jasonmchristos | thanks | 05:59 |
jasonmchristos | experimenting trying to setup my own facebook type site | 06:01 |
jasonmchristos | man im rusty trying to use the cp command to copy the contents of a directory do i use cp folder/ -T /destfolder/ ? if i want to copy everything in one folder to another? | 06:09 |
jasonmchristos | im getting cp: extra operand `/home/zero/elgg1.6.1/index.php' | 06:10 |
pwnguin | what is -T for? | 06:10 |
jasonmchristos | specify destination? | 06:10 |
pwnguin | i dont have dosbox installed, but im pretty sure thats not even true for dos | 06:11 |
pwnguin | cp src dst | 06:11 |
pwnguin | but since it's a directory | 06:11 |
pwnguin | cp -R src dst | 06:11 |
jasonmchristos | sudo cp -R /home/zero/elgg1.6.1/ /var/www/ | 06:12 |
jasonmchristos | ? | 06:12 |
jasonmchristos | it worked except i didnt want the actual directory copied | 06:13 |
jasonmchristos | sudo cp -R /home/zero/elgg1.6.1/*.* /var/www/ <-- will that work? | 06:14 |
pwnguin | no | 06:14 |
pwnguin | but its close | 06:14 |
pwnguin | UNIX doesn't require files to have a . extension | 06:14 |
jasonmchristos | sudo cp -R /home/zero/elgg1.6.1/* /var/www/* <-- will that work? | 06:14 |
pwnguin | i think so | 06:14 |
jasonmchristos | thanks looks like this worked sudo cp -R /home/zero/elgg1.6.1/* /var/www/ | 06:16 |
pwnguin | oh yea, didnt notice the second * | 06:16 |
jasonmchristos | 3. Create a database Using your database administration tool of choice (if you're unsure about this, ask your system administrator), create a new database for Elgg. Make sure you know the username and password necessary to access this. | 06:41 |
jasonmchristos | which tool do i use for ubuntu server? | 06:41 |
twb | jasonmchristos: for toys, you want sqlite. | 06:52 |
jasonmchristos | twb: im using LAMP i think it has mqsql up and running out of box | 06:53 |
jasonmchristos | but im nott sure how to create a new database | 06:53 |
jasonmchristos | mysql> CREATE DATABASE database1; | 06:56 |
jasonmchristos | that made a file called CREATE | 06:56 |
jasonmchristos | i dont think it worked properly | 06:56 |
twb | jasonmchristos: I don't provide mysql support, sorry. | 06:57 |
jasonmchristos | ok thanks anyway | 06:57 |
Noble | What packages do I need to get php5 to speak with PostgreSQL? | 07:26 |
=== sabgenton_ is now known as sabgenton | ||
jmarsden | Noble: php5-pgsql | 07:53 |
soren | hggdh: stepmaker is part of the test system. To invoke it, change the test definition from "type = steps" to "type = stepmaker". | 07:57 |
soren | hggdh: This is likely a useful reference: http://www.linux-kvm.org/page/KVM-Autotest | 07:58 |
jiboumans | morning guys | 08:04 |
lool | soren: Mind dropping the preflight_check() locale-gen bits? I much prefer failing at the end of the vm build, even if that's a time/resource waste, rather than risking my hosts' locales ;-) | 08:30 |
jasonmchristos | where are the msql databases located on the LAMP karmic server? | 09:17 |
=== sabgenton_ is now known as sabgenton | ||
jussi01 | Just a reminder to ops in this channel, if you renewed your operatorship, you are now expected to idle in #ubuntu-ops :) See you there. | 09:28 |
lifeless | jussi01: hi | 09:40 |
lifeless | jussi01: speaking of operatorship; my motu irc ops membership is up for renewal | 09:40 |
jussi01 | ji lifeless | 09:40 |
jussi01 | hi even :) | 09:40 |
lifeless | :) | 09:40 |
jussi01 | lifeless: did you get the email this morning that allows you self renew? | 09:41 |
lifeless | ah yes,cool, doing so. | 09:41 |
lifeless | I didn't notice the difference from the day before | 09:41 |
jussi01 | yeah, thats why they came 2x :) | 09:41 |
lifeless | thanks ;) | 09:41 |
jussi01 | (yes, I screwed it up the first time) | 09:42 |
lifeless | separate topic, for the floor. | 09:42 |
Steel__ | Hello to all, i have installed the linux source / headers and recompiled my kernel, then i copied ./arch/x86_64/boot/bzimage /boot/My-Kern, edited Grub and restartet my Machine. I see the Kernel starting but it doesnt load the modules/drivers it just hangs at the beginning. How do i get the information what i did wrong ? | 09:42 |
lifeless | Any opinions on the likelyhood of success running hardy with a lucid kernel ? | 09:42 |
twb | lifeless: why do you want to do that? | 09:42 |
persia | lifeless: Theoretically newer kernels with older userspace is safer than the opposite way about, but you'd want to test virtually first (because there's sometimes stuff that got deprecated that userspace needs). | 09:44 |
lifeless | twb: I have a new backup drive; I want to put ext4 or possibly something more sexy still on it. | 09:46 |
twb | Backups should not be sexy | 09:47 |
lifeless | twb: Its my home mail/web box. So it runs LTS; I will upgrade to lucid, just not this week. | 09:47 |
lifeless | twb: You asked why :) | 09:47 |
twb | I'm just saying that your rationale is bogus | 09:47 |
lifeless | twb: thats fine; I want to do something bogus. | 09:47 |
twb | Offhand I would expect it to explode due to udev, but not much else. | 09:47 |
persia | Does udev fail to ignore unknown events? | 09:48 |
twb | Dunno. | 09:48 |
persia | I thought it did, but if not, yeah, that would be a big source of issues. | 09:48 |
twb | But I know that upgrading udev and the kernel out-of-step has caused me lots of troubles in the past | 09:48 |
pts | what is the status of the clusterstack in lucid? A few days ago I would install pacemaker-openais, but now that packages is non-existent when I'm trying a full reinstall with daily build from 23. feb. The clusterstack testing wiki says to run apt-get install pacemaker, but the dependency there is openhpi not openais. Anyone know what's going on? | 09:48 |
pts | After some googling I guess I can rephrase my question :) where did pacemaker-openais go? | 10:00 |
twb | According to rmadison, it's currently only available in karmic/universe | 10:01 |
twb | If this were debian, I'd look at the debian-release mailing list for a removal request from its last Debian maintainer. | 10:01 |
persia | The process is precisely the same for Ubuntu (check for removal requests in Debian). | 10:02 |
twb | persia: haha | 10:02 |
persia | The exceptions show up in the sync blacklist (http://people.canonical.com/~ubuntu-archive/sync-blacklist.txt ) | 10:03 |
persia | The exceptions to that are the few Ubuntu-local packages that exist. There used to be a removals page for those, but I haven't been able to find it recently. | 10:03 |
twb | persia: if I was some peon, how would I have discovered that file? | 10:03 |
twb | I guess it's linked from some wiki page | 10:04 |
persia | wiki.ubuntu.com/ArchiveAdministration used to have it. | 10:04 |
persia | The reason it's in my browser history cache is because someone mentioned it to me when blacklisting something in #ubuntu-devel a couple weeks ago. | 10:04 |
twb | pts: you'll have to google harder than I did; I couldn't spot it on gmane in thirty seconds. | 10:07 |
pts | trying to learn how pacemaker/openais works so we can Implement it in our network, but when trying to install pacemaker it depends on heartbeat, and for what I can understand that is more or less obsolete | 10:11 |
pts | So I am a bit confused :D | 10:11 |
twb | Yay for HA computing on a student's minix knockoff | 10:12 |
=== sabgenton_ is now known as sabgenton | ||
_ruben | heartbeat as a full HA stack is obsoleted, heartbeat as just a messaging layer (together with pacemaker for instance) is very much alive/active | 10:15 |
pts | I see, but isnt openais a better choice? As I've understood Heartbeat is maintained by Linbit, and they will only maintain it for stability. But then again I see that corosync is pulled as a dep of pacemaker, but I havent fully understood the relationship between corosync and openais | 10:21 |
lifeless | wish me luck | 10:30 |
kaushal | hi | 10:35 |
kaushal | Is there a tool available to deploy same war to multiple tomcat server ? | 10:35 |
lifeless | so, lvm doesn't come up; digging :P | 10:49 |
sabgenton | anyone famillar with grub2? | 11:25 |
sabgenton | (grub karmic) | 11:25 |
sabgenton | i set grub to bot i diffren menu entry rebooted and nothing changed I'm guessing grub failed to boot that entry and reverted to the last working entry? | 11:35 |
sabgenton | is there a log of what happened someware? | 11:35 |
sabgenton | grub2 | 11:36 |
sabgenton | in karmic | 11:36 |
kaushal | hi | 12:08 |
kaushal | is there a way to start some specific services during boot up in a sequential order ? | 12:08 |
kaushal | for example 1)tomcat 2)mysql 3)dns server | 12:09 |
Cromulent | hmm I have a strange problem - I can ping www.mydomain.com, mydomain.com and mail.mydomain.com properly but ns.mydomain.com and ns2.mydomain.com don't work - my zone file for BIND is set up the same way for both - is there any reason that some of my DNS stuff works but the nameservers don't? | 12:20 |
jiboumans | Cromulent: your dns servers may simply not respond to ping? (not all machines do) | 12:26 |
Cromulent | jiboummans: its all the same machine which is what is perplexing :) | 12:27 |
Cromulent | I think I'm going to leave it for a bit and see if it is a DNS propagation issue | 12:28 |
jiboumans | Cromulent: and all the same interface? what does 'host ns.mydomain.com' tell you? | 12:28 |
Cromulent | it gives me the correct IP address | 12:29 |
Cromulent | but ping says unknown host | 12:29 |
jiboumans | Cromulent: that's somewhat strange. i trust that a ping of that ip works just fine | 12:34 |
jiboumans | Cromulent: at which point your 'ping' is acting strange by not resolving teh same way 'host' does, but the system runs fine | 12:34 |
Cromulent | yep ping to the IP address directly works fine - everything else is fine including as I say pinging other parts of the same nameserver - but the nameserver records themselves seem to be playing up | 12:38 |
Jeeves_ | compliments on the Lucid server installer. To whom it may concern :) | 12:53 |
cjwatson | thanks :) | 12:57 |
Jeeves_ | It installed almost without issues. | 12:57 |
ogra | almost ? | 12:57 |
Jeeves_ | grub(2) didn't want to install | 12:57 |
pmatulis | that's pretty serious! | 13:10 |
Jeeves_ | Lilo did, and i'm doing stuff with raid1 and have been fiddling around with lvm | 13:15 |
Jeeves_ | so that might be the issue | 13:15 |
lenios | anyone knows where i should put 'aide --check' to have it to check on each boot? | 13:17 |
lenios | grub2 is supposed to be able to boot on raid1 | 13:18 |
Jeeves_ | lenios: I know | 13:21 |
Jeeves_ | But i've seen more installers having issues with it | 13:22 |
Jeeves_ | btw, there allready was a bug about lucid server staying in TTY7 and not showing TTY0, right? | 13:22 |
acalvo | easiest way to deploy a load-balanced with failover web/mysql server? | 13:28 |
ttx | kirkland: ping | 13:35 |
ttx | I already filed a bug about the network/metadata issue with CLC separated from CC | 13:35 |
ttx | https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/527648 | 13:35 |
uvirtbot | Launchpad bug 527648 in eucalyptus "Running instances can't be contacted in CLC+Walrus / CC+SC / NC topology" [High,Confirmed] | 13:35 |
ttx | so no need for you to file one | 13:36 |
Jeeves_ | acalvo: Hmm, I would say keepalived and mysql ring-replication | 13:36 |
acalvo | thanks! (I've just googled and found pound and mod_proxy_balancer for apache) | 13:37 |
ttx | kirkland: I'm currently trying to reproduce using plain ISOs, just to make sure it's not something we introduce in preseeds. It definitely used to work, so I wonder when it regressed :) | 13:37 |
Jeeves_ | acalvo: That would still introduce a single point of failure | 13:38 |
acalvo | well, I'll take a closer look at your suggestions | 13:38 |
lenios | you can read that too https://help.ubuntu.com/community/HighlyAvailableLAMP | 13:38 |
kaushal | hi | 13:40 |
kaushal | can someone guide me about http://tomcat.apache.org/tomcat-5.5-doc/deployer-howto.html ? | 13:40 |
kaushal | i did downloaded the TCD | 13:40 |
kaushal | After extracting i see only build.xml docs images lib LICENSE | 13:41 |
kaushal | NOTICE RELEASE-NOTES | 13:41 |
kaushal | got stuck and not able to proceed | 13:41 |
acalvo | lenios: thanks | 13:41 |
lenios | kaushal, well it says you have to install it | 13:45 |
kirkland | ttx: okay | 13:45 |
kaushal | lenios: i did not understand | 13:46 |
kirkland | ttx: i saw your second email before i filed one | 13:46 |
kirkland | ttx: agreed, this is a recent regression | 13:46 |
ttx | kirkland: I suspect that some of the networking setup is no longer applied, or something like that | 13:46 |
lenios | kaushal, "An installation is however required if you wish to use the Tomcat Client Deployer (TCD)." | 13:47 |
kaushal | lenios: really confused | 13:47 |
kaushal | I have already TC server running | 13:48 |
lenios | i guess you have to install it like any other webapp on tomcat | 13:48 |
kaushal | ok | 13:48 |
kaushal | lenios: i reread http://tomcat.apache.org/tomcat-5.5-doc/deployer-howto.html | 13:53 |
kaushal | still not clear | 13:53 |
ttx | kirkland: reproduced issue with plain ISOs, fwiw | 13:53 |
kaushal | please guide | 13:53 |
kirkland | ttx: right; i saw it across the board; all 3 methods i tested installation | 13:53 |
kirkland | ttx: i wonder if our ephemeral "fix" regressed this | 13:54 |
kirkland | ttx: that's the only thing that's changed related the metadata in a while | 13:54 |
kirkland | ttx: unless it's a problem at another layer | 13:54 |
ttx | kirkland: It's not metadata that is failing | 13:54 |
ttx | it's plain networking | 13:55 |
ttx | you can ping karmic instance from CC/SC | 13:55 |
ttx | but not from CLC/Walrus | 13:55 |
kirkland | hmm | 13:55 |
ttx | so it's just the network between instance and CLC/Walrus that's broken. A side effect being... metadata can't be accessed from the instance | 13:56 |
smoser | can i get at somewhere that shows this ? | 13:56 |
smoser | i can hack an image and register it to put some ssh keys inside already | 13:56 |
smoser | so we don't have to worry about the metadata service not being there, you'll be able to get in and see what is going wrong | 13:56 |
kirkland | smoser: the datacenter setup | 13:58 |
smoser | o. | 13:58 |
smoser | ok | 13:58 |
ttx | looks like the bridge device is set up correctly... | 13:59 |
uvirtbot | New bug: #527648 in eucalyptus (main) "Running instances can't be contacted in CLC+Walrus / CC+SC / NC topology" [High,Confirmed] https://launchpad.net/bugs/527648 | 14:01 |
ttx | kirkland: is VNET_PUBINTERFACE / VNET_PRIVINTERFACE supposed to be set on CCs ? | 14:03 |
kirkland | ttx: http://manpages.ubuntu.com/manpages/lucid/en/man5/eucalyptus.conf.5.html | 14:05 |
kirkland | ttx: according to that, it's in the " Options affecting CC, NC only:" section | 14:06 |
kirkland | ttx: i wrote that manpage based on an email i got from nurmi | 14:06 |
ttx | kirkland: ok, thx | 14:06 |
kirkland | ttx: if you haven't seen that manpage yet, it's really good; writing it really demystified eucalyptus.conf for me | 14:06 |
ttx | kirkland: I'll have to read it | 14:07 |
ttx | kirkland: VNET_*INTERFACE is set correctly, so it's not coming from there... | 14:07 |
ttx | kirkland: as far as I can tell, network is set up like it's supposed to... | 14:07 |
ttx | kirkland: I'm running out of ideas and need to spend time on some other things... i suspect we'll need Dan's help on this one. | 14:11 |
ttx | smoser: having an image that can boot and be sshed into without requiring metadata service would probably help in debugging | 14:12 |
smoser | its being uploaded now | 14:12 |
ttx | smoser: cool, thanks | 14:12 |
smoser | i *think* that emi-B1F6186B should come up, and be reachable by ubuntu@cempedak | 14:12 |
smoser | testing that assertion now | 14:12 |
ttx | i'm pretty sure you'll be able to ssh into that one from the CC, but not from the CLC | 14:12 |
smoser | i'm on cempedak | 14:13 |
ttx | and that you won't be able to query the metadata service at all. | 14:13 |
smoser | shoot. | 14:13 |
smoser | it went immediately to shutting_down | 14:14 |
ttx | smoser: the karmic images boot alright | 14:14 |
ttx | you can ping them from CC / it's just that you can't SSh into them because no authorized_keys is queried from metadata | 14:15 |
ttx | but... you can't ping them from CLC | 14:15 |
ttx | that kinda proves that there is an issue in networking between the components. | 14:15 |
smoser | so it didn't boot because i forgot --instance-type c1.medium | 14:18 |
smoser | i think small fails because root filesystem is 2G and small disk is 2G, leaving no room for space | 14:19 |
smoser | s/space/swap/ | 14:19 |
ttx | kirkland: ping me when you've 5 minutes | 14:24 |
smoser | ttx, ok, where should i be able to reach this instance from ? | 14:26 |
ttx | CC | 14:26 |
kirkland | ttx: i have all the time you need | 14:28 |
ttx | kirkland: wanted to quickly discuss which bugs we should escalate to eucalyptus on todays call, trying to prioritize | 14:29 |
kirkland | ttx: okay, you want to call me? | 14:29 |
ttx | kirkland: no, we can discuss it here, should be quick | 14:29 |
ttx | I think we should mention bug 525989, bug 527648 and bug 526506 | 14:30 |
uvirtbot | Launchpad bug 525989 in eucalyptus "Eucalyptus runs images without ramdisk with a default ramdisk" [High,Confirmed] https://launchpad.net/bugs/525989 | 14:30 |
uvirtbot | Launchpad bug 527648 in eucalyptus "Running instances can't be contacted from CLC in CLC+Walrus / CC+SC / NC topology" [High,Confirmed] https://launchpad.net/bugs/527648 | 14:30 |
uvirtbot | Launchpad bug 526506 in eucalyptus "[karmic]: attaching volume to /dev/sda results in non operational node" [High,Confirmed] https://launchpad.net/bugs/526506 | 14:30 |
kirkland | ttx: definitely | 14:31 |
ttx | the last one seems of lesser importance, but I'm scared by users taking out nodes just by running the wrong command | 14:31 |
kirkland | ttx: agreed | 14:31 |
ttx | so if we can reproduce in lucid... | 14:31 |
ttx | anything else I missed ? | 14:31 |
smoser | kirkland, what host is the cc in the datacenter ? | 14:32 |
ttx | bug 522204 is also an upstream fix, but it's less important, and we can suggest a patch, so i'd leave it out for the moment | 14:32 |
uvirtbot | Launchpad bug 522204 in eucalyptus "euca_conf --[de]register-nodes is brittle" [Medium,Confirmed] https://launchpad.net/bugs/522204 | 14:32 |
kirkland | ttx: i'm looking | 14:32 |
kirkland | smoser: are there any standing euca2ools/boto issues? | 14:33 |
kirkland | smoser: seems like we've been staying on top of working around them as they come up | 14:34 |
smoser | not that i'm aware of. | 14:34 |
ttx | kirkland: also if you agree with my comment on https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/519870, feel free to move it to cloud-init | 14:34 |
uvirtbot | ttx: Error: This bug is private | 14:34 |
kirkland | smoser: santol is the cc | 14:34 |
ttx | smoser: you'll have to copy keys over to connect from the CC | 14:35 |
kirkland | ttx: i do agree with your comment; i was wondering how to mark a bug as affecting the image itself | 14:35 |
kirkland | ttx: i suppose cloud-init is about as close as we can get | 14:35 |
ttx | kirkland: yes, it's basically a ec2-init SRU for karmic and a respin | 14:35 |
ttx | smoser: you had a karmic refresh planned in the next weeks, iirc ? | 14:36 |
smoser | ttx, yes | 14:36 |
kirkland | ttx: done | 14:36 |
ttx | smoser: you think it's doable to do the hostname trick in the same run ? | 14:36 |
smoser | hold on | 14:37 |
kirkland | ttx: that looks like the most critical short list | 14:38 |
ttx | kirkland: Ok, thanks. I prefer us to sync before the call so that we come up with a sane argument :) | 14:39 |
kirkland | ttx: agreed | 14:39 |
ttx | (the "importance" was correctly reflecting what's important, which is good) :) | 14:39 |
smoser | ttx, ok. so above, the hostname trick yeah, we can do that back into karmic. | 14:46 |
kirkland | smoser: okay, cool | 14:47 |
smoser | ttx, kirkland in data center, emi-B1E0185E is runnable. | 14:50 |
smoser | it is mostly a static instance, and the ubuntu user on santol can go straight in | 14:51 |
kirkland | smoser: and ssh-able? | 14:51 |
smoser | yeah | 14:51 |
kirkland | smoser: from santol | 14:51 |
kirkland | smoser: what about from cempedak? (the clc) | 14:51 |
ttx | smoser; right | 14:51 |
smoser | right. go to santol, then ssh 172.19.1.3 | 14:51 |
smoser | i can't get there from cempedak, as ttx expected, no ping either. | 14:51 |
ttx | shoudl fail from clc .. and you should fail to query metadata from the instance as well | 14:51 |
ttx | that confirms my findings | 14:51 |
smoser | yeah | 14:54 |
smoser | from inside the guest | 14:54 |
smoser | $ telnet 169.254.169.254 80 | 14:54 |
smoser | Trying 169.254.169.254... | 14:54 |
smoser | telnet: Unable to connect to remote host: Connection refused | 14:54 |
smoser | ttx, kirkland i'll try to write a script to "fix" an instance | 14:54 |
smoser | err, i mean an image | 14:54 |
smoser | so it really does nothing on boot but has your keys embedded already | 14:55 |
gzur | Hi there, I'm running Ubuntu Server 9.10 and I'm gettins No space left on device errors all of a sudden, running "df- Th" shows me that I have 8.5 Gigs left on my ext4 primary filesystem- there are some other tmpfs with 200-500 MB free, so I don't understand why I'm getting this message | 14:58 |
netrat | gzur, are you using quotas? are you out of inodes? | 14:59 |
soren | gzur: What does "df -ih" say? | 14:59 |
smoser | also, as non-root, there may be 5% reserved for root | 15:00 |
Omahn | Hi all. I just tried upgrading from 8.04 to 10.04 on a cloned VM and it failed immediately with: | 15:03 |
Omahn | Exception during pm.DoInstall(): E:Couldn't configure pre-depend mountall for upstart, probably a dependency cycle. | 15:04 |
Omahn | Should I report a bug on mountall or upstart? :) | 15:04 |
zul | is there a way to cache ssh keys on the CLI | 15:05 |
netrat | Omahn, Exception during pm.DoInstall(): E:Couldn't configure pre-depend mountall for upstart, probably a dependency cycle | 15:05 |
netrat | Omahn, sorry about that https://bugs.launchpad.net/ubuntu/+source/update-manager/+bug/527605 | 15:05 |
uvirtbot | Launchpad bug 527605 in update-manager "kubuntu 8.04 -> 10.04 upgrade fails, due to likely dependency cycle" [Undecided,New] | 15:05 |
Omahn | ta. | 15:05 |
Omahn | netrat: Does update-manager provide do-release-upgrade? | 15:05 |
smoser | zul, what does cache ssh keys mean ? | 15:05 |
Omahn | Yes it does. :-) | 15:06 |
zul | smoser: nm figured it out | 15:06 |
Pici | ssh-agent | 15:06 |
netrat | Omahn, sorry i'm not for sure.. I'm sticking with 8.04 for awhile until all the bugs get sorted out | 15:08 |
Omahn | netrat: We're trying to do as many test upgrades as possible to make sure all the bugs are caught as early as possible :-) | 15:09 |
netrat | Omahn, thanks for sorting all the bugs out for me! | 15:13 |
Omahn | netrat: You need to understand, I'm not sorting them, I'm just finding them ;-) | 15:14 |
gzur | netrat and soren: df -oh says Inodes: 2.3 - IUsed: 2.3 IFree: 1 IUse%: 100%, so yes I', out of inodes :) | 15:21 |
gzur | thanks | 15:21 |
=== JaixBly_ZzZ is now known as JaixBly | ||
gzur | What does that mean? | 15:23 |
gzur | can I increase the number of inodes? | 15:23 |
netrat | gzur, i think the number of inodes is set when the filesystem is created... what are you using the filesystem for if you don't mind me asking? | 15:24 |
gzur | Storing images used in a tiled web mapping service | 15:24 |
gzur | It's an obscene number of files/directories | 15:25 |
netrat | gzur, what type of filesystem? be back in a few | 15:25 |
gzur | netrat: ext4 I think | 15:26 |
gzur | netrat: ext4 for certain :) | 15:27 |
=== Guest26987 is now known as fahadsadah | ||
netrat | gzur, the number of inodes is set at filesystem creation. i don't think you can increase that number afterwards | 15:41 |
netrat | gzur, normally the number of inodes is set with a bytes-per-inode ratio. check the man page for mkfs.ext4, the i, I, and N flags should be of interest | 15:43 |
gzur | is it possible that the there might be junk inodes? | 15:45 |
gzur | netrat: and thanks for the man page ref - it's exactly what I need it seems. | 15:48 |
=== robbiew_ is now known as robbiew | ||
panter | hi .. i have problem with my own deb repository ... i have few own packages (mostly with scripts) .. and need to install and update this packages on several servers .... i followed this tutorial ... http://mediakey.dk/~cc/howto-create-your-own-debian-or-ubuntu-package-repository/ ... i can add repository to /etc/apt/sources.list but i can't see packages from my repo ... is there anybody who can help me? | 16:20 |
zul | why not just use ppa? | 16:40 |
uvirtbot | New bug: #527875 in libvirt (main) "libvirt0 unable to operate due to missing essential files" [Undecided,New] https://launchpad.net/bugs/527875 | 16:56 |
=== toabctl_ is now known as toabctl | ||
oru_work | in ubuntu with postfix/dovecot, i need to forward everything from one email address to the other, how can I accoplish this ? | 18:21 |
uvirtbot | New bug: #526944 in squid (main) "Unknown $NAME variable in squid upstart script" [Undecided,Fix released] https://launchpad.net/bugs/526944 | 18:46 |
Dittohead | I have recently setup a couple of 9.10 x86 servers for some LAMP and zenoss duties. I've added a new source for zenoss, but it kicks back a 407 Proxy Authentication Required for the zenoss source. Strangely, the others work just fine. It doesn't seem to be making an attempt to authenticate on the zenoss source. Any ideas? | 18:47 |
shang | hi all, I am having some issue with likewise joining the domain. After the system reboot, it will take about 1 minute for the system to get the ticket from the windows server | 19:00 |
shang | any direction that where I should look will be much appreciated! | 19:00 |
exobuzz | any smartd experts? drive reported a bad sector (Current_Pending_Sector of 1).. i did a long selftest that failed on LBA 976768629. i tried reading from that block (dd if=/dev/sdv skip=976768629 bs=512) and it was fine! did a raid check/resync and no recoverable error. did another long selftest and the disk reported no errors but the Current_Pending_Sector and Offline_Uncorrectable are still not 0 | 19:07 |
exobuzz | will the pending sector not go until the block is rewritten ? despite the fact it seems to be able to read it now (or it seems it was a intermittent issue?( | 19:08 |
ivoks | kees: i'm here to answer any of your questions regarding heartbeat/openais/corosync/pacemaker | 19:15 |
kees | ivoks: ah-ha, excellent. | 19:17 |
kees | ivoks: so, what's the issue with 2.99 vs 3.0 ? | 19:17 |
ivoks | kees: 2.99 was mutant | 19:17 |
ivoks | kees: it was a development snapshot that worked at one point | 19:17 |
ivoks | kees: but since then heartbeat isn't the same | 19:18 |
ivoks | kees: in 3.0 it contains only portions of 2.99, while portions were split into cluster-glue and cluster-agents | 19:18 |
kees | ivoks: how soon do you think Debian will have the 3.0 stuff? | 19:18 |
ivoks | kees: and some parts were removed | 19:18 |
ivoks | kees: very | 19:19 |
ivoks | kees: actually, i'll be working with debian-ha-maintainers on adopting new stuff | 19:19 |
kees | ivoks: I would feel much better about it in an LTS if it were reasonably in sync with Debian. | 19:19 |
kees | overall, I'm find with heartbeat being in main from both an MIR and security perspective. | 19:19 |
kees | s/find/fine | 19:19 |
kees | but I would rather the discontinuity with Debian was solved first. | 19:20 |
ivoks | kees: i doubt new version will get in before beta1 | 19:20 |
ivoks | (in debian) | 19:20 |
kees | ivoks: I'll comment on the bug and conditionally approve it. | 19:22 |
ivoks | ok | 19:22 |
kees | ivoks: on a totally separate topic, do you have any idea where ACL/xattr support stands with tar? I know you'd looked into it briefly at one point. | 19:23 |
ivoks | kees: haven't look at it | 19:23 |
ivoks | kees: there were some issues with the patch | 19:23 |
kees | ivoks: yeah, the email thread seems to have died out a bit. | 19:24 |
ivoks | kees: i'm not sure there was any change after that | 19:24 |
ivoks | well, that's just bad. | 19:24 |
kees | ivoks: what confused me is that RH seems to carry that flawed patch; I'd figure it would cause a lot of pain for them. | 19:24 |
ivoks | but it looks it's working | 19:24 |
kees | yeah, weird. | 19:26 |
ivoks | kees: if you have any other questions about cluster stuff, ping me | 19:28 |
kees | ivoks: sure thing. thanks! | 19:28 |
areay | hi all... does anyone know of a good guide to getting ldap/kerberos working? all the guides i've found either don't work, don't make sense, or are outdated... | 19:40 |
unit3 | Ok. It looks like nfs mounts in fstab attempt to happen before the system brings up my main bridge device on boot. | 19:41 |
ivoks | unit3: karmic? | 19:41 |
unit3 | yep. | 19:41 |
unit3 | is there an easy way to reorder this? get it to activate the bridge device so I have networking before it tries to mount nfs stuff? | 19:42 |
ivoks | unit3: you have /home on nfs? | 19:42 |
unit3 | yeah | 19:42 |
unit3 | exactly. :) | 19:42 |
ivoks | unit3: workaround i did was mounting nfs share to /srv/home | 19:42 |
ivoks | unit3: and then in rc.local | 19:42 |
unit3 | oh, bind it to /home. | 19:42 |
unit3 | gotcha. | 19:42 |
ivoks | unit3: mount -a followd by mount bind /srv/home to /home | 19:42 |
unit3 | I guess that'd work in the short term. do you know if there's a launchpad bug I can follow about this? | 19:43 |
ivoks | this was fixed in lucid | 19:43 |
unit3 | oh, is there a package I can backport to fix it? | 19:43 |
ivoks | hopefully, there'll be a backport at some point | 19:43 |
ivoks | fix is in 'upstart' package | 19:43 |
ivoks | and i wouldn't recomend backporting it :) | 19:44 |
unit3 | oh, I guess because there's tons of other changes that'd need to accompany it? | 19:44 |
ivoks | correct | 19:44 |
ivoks | lucid will be out soon | 19:44 |
unit3 | I could still use a link to the launchpad bug then, just to try and track down the changes and maybe make my own patch. | 19:44 |
unit3 | Yeah, but I have broken production servers now. :P | 19:44 |
ivoks | i know | 19:44 |
ivoks | same problem i had | 19:44 |
ivoks | but i'll be using workaround till lucid is released | 19:45 |
unit3 | fair enough. I guess I'll put that in place for the time being. | 19:45 |
unit3 | thanks. :) | 19:45 |
ivoks | np | 19:45 |
ivoks | well, 'night all | 19:47 |
=== lenios is now known as amelie | ||
JamesCard | Where should I place my web files? Using IIS I usually have them in separate directories away from the webroot. I don't think I want them all in /var/www. What makes sense for a public webserver built on Ubuntu? | 19:55 |
RoAkSoAx | JamesCard, you can place them where ever you want. You *don't* have to place them under /var/www | 19:56 |
RoAkSoAx | JamesCard, however, you will need to edit the root in the webserver's config file | 19:57 |
JamesCard | Right, I know I can create aliases. Just wondering what is suggested as best practice. | 19:57 |
JamesCard | This server will have a single user, but we'd rather not have to deal with sudo every time we want to touch any of our web files. | 19:58 |
da65 | is it easy to set up irc on my new ubuntu server, | 19:59 |
JamesCard | Keeping them in the admin user's home directory seems most convenient, but are reasons we might not want to do that? | 19:59 |
RoAkSoAx | JamesCard, well, if you are using apache, and you create a public_html file in $HOME of each user, you can do that | 19:59 |
RoAkSoAx | or either, give different permissions to whereever your files are | 19:59 |
=== amelie is now known as lenios | ||
JamesCard | On Windows servers we usually create a C:\Projects\Web directory to contain everything the webserver should see. Perhaps these files make sense in /opt or /srv -- they are, after all, not user-specific files and don't seem to belong in /home. | 20:06 |
RoAkSoAx | JamesCard, whatever makes your live easier :). I mean /var/www is just a default as many other things windows has for default directories... | 20:09 |
JamesCard | RoAkSoAx: thanks. | 20:11 |
RoAkSoAx | JamesCard, no prob. | 20:12 |
JamesCard | According to <http://www.pathname.com/fhs/pub/fhs-2.3.html#SRVDATAFORSERVICESPROVIDEDBYSYSTEM> it looks like /srv is the "proper" place. Makes sense to me, at least. Thanks. | 20:16 |
alvin | unit3: The bug you're looking for is bug #470776 | 20:47 |
uvirtbot | Launchpad bug 470776 in mountall "retry remote devices when parent is ready after SIGUSR1" [Medium,Fix released] https://launchpad.net/bugs/470776 | 20:47 |
unit3 | oh, thanks. :) | 20:51 |
mdeslaur | soren: is this a known issue with the latest vmbuilder? http://paste.ubuntu.com/383941/ | 20:55 |
mdeslaur | soren: also, vmbuilder --help seems to be broken...didn't that work before? | 20:56 |
soren | mdeslaur: Not entirely a known problem. Not entirely surprising either, now that I think about it, though. | 21:09 |
soren | Gah. | 21:10 |
soren | 22:09:53 < soren> mdeslaur: Not entirely a known problem. Not entirely surprising either, now that I think about it, though. | 21:12 |
soren | mdeslaur: If you file a bug about it, I'll bet it'll be fixed within the next 12-14 hours. | 21:12 |
mdeslaur | soren: hehe, will do then :) | 21:13 |
* soren ponders | 21:13 | |
mdeslaur | soren: bug #528027 | 21:17 |
uvirtbot | Launchpad bug 528027 in vm-builder "vmbuilder crashed with AttributeError in set_defaults()" [Undecided,New] https://launchpad.net/bugs/528027 | 21:17 |
uvirtbot | New bug: #528027 in vm-builder (universe) "vmbuilder crashed with AttributeError in set_defaults()" [Undecided,New] https://launchpad.net/bugs/528027 | 21:26 |
decembre | hello | 21:47 |
decembre | I have some probleme with usb ! | 21:47 |
decembre | my lsusb does return anything except the Linux root hub ! | 21:47 |
unit3 | decembre: what kind of devices do you have plugged in via USB, and what release of Ubuntu are you using? | 22:23 |
udha | echo request | 22:50 |
* lifeless requests | 22:50 | |
lifeless | kirkland: I've upgraded my home server to lucid; so I will be able to try those UEC bugs | 22:51 |
kirkland | lifeless: well done! | 22:51 |
lifeless | kirkland: but also, I'm going to stab at getting all services on one machine: DHCP, DNSd, UEC* | 22:51 |
kirkland | lifeless <---- ambitious man | 22:51 |
lifeless | kirkland: well, it was a bit of a disaster: 4 dependency bugs found and reported; up to 3:30ish am. | 22:51 |
kirkland | lifeless: btw, Karmic UEC -> Lucid UEC upgrade is known broken | 22:52 |
lifeless | kirkland: it will be an experience | 22:52 |
kirkland | lifeless: i'm tackling that on Monday | 22:52 |
lifeless | kirkland: thats ok, this machine was running hardy. | 22:52 |
kirkland | lifeless: ah | 22:52 |
lifeless | which is why I found all these bugs ;) | 22:52 |
udha | can I ask a noob questions without being flammed here? haven't been in this channel before... | 22:53 |
udha | *question | 22:53 |
lifeless | udha: yep, we're nice. We might point you at a better forum, but we can't tell until you ask the question. | 22:53 |
Roxyhart0 | sombody know why when i create a new user (ldap) and i do permision in a folder for this user is just take permision as dnslog and not as the "user" | 22:54 |
udha | lifeless, thanks, I'm just downloading ubuntu 9.10 server (64bit) for a headless box at home, and noticed for the first time the enterprise cloud features for multiple systems, but was wonding if someone can explain the basics of it? | 22:55 |
lifeless | Ubuntu Enterprise Cloud? | 22:55 |
lifeless | udha: ^ is that what you're referring to? | 22:55 |
udha | lifeless, yes that's right | 22:56 |
lifeless | http://www.ubuntu.com/cloud/private has some high level info | 22:57 |
lifeless | what sort of basics are you interested in: capabilities? how it works? what it lets you do? | 22:57 |
udha | thanks lifeless, haven't seen this page before, reading now... | 22:57 |
Pirate_Hunter | hi can anyone provide me a linux alternative to what this link from m$ proposes http://networkedblogs.com/p28446238 thanks | 22:58 |
udha | So with the cloud options, could I say, run a LAMP website and have it scale automatically across machines in the cloud as it's usage expands? | 22:59 |
udha | Or use it for a high-availability DNS cluster etc? | 22:59 |
udha | What would happen if a machine failed? | 23:00 |
Pirate_Hunter | anyone? even knowing udha might be talking about what I am looking for | 23:01 |
udha | Pirate_Hunger: I'm having a look at that link now... | 23:01 |
Pirate_Hunter | oh ok I found it quite interesting specially the whole application and cloud working together | 23:02 |
udha | Having MS and Government tied together even closer than they are already makes me very afraid | 23:02 |
udha | Pirate_Hunter: that video and article are talking about cloud services online right? | 23:04 |
udha | I'm still watching the vid, 4 mins.. | 23:04 |
Pirate_Hunter | udha, not exactly it talks about the cloud service as well as having tailored software on the server. However the main benefit is both cloud and the server (software) working together | 23:05 |
Pirate_Hunter | there is quite a lot of benefits to that I just want to know if linux has something similar as an alternative even if it hasn't progressed to that level | 23:06 |
udha | :Pirate_Hunter: I only just came here to ask about Ubuntu's cloud services myself, so I'm no authority on the subject, but I'll need to see more than a paper-cut-out video to form any kind of opinion about the Software+ idea | 23:07 |
udha | I've just started reading this if it's of any help: http://www.ubuntu.com/system/files/introduction-to-cloud-computing_server_cloud.pdf | 23:07 |
udha | Pirate_Hunder: Is that video/article talking about anything beyond SaaS? | 23:08 |
Pirate_Hunter | fair enough I understand what you mean, I got quite hyped about this piece of news it is on the lines of what I was already thinking. I am currently designing a server so looking at the options available and will read that link might contain something that will help me out | 23:09 |
udha | Pirate_Hunder: lifeless has answered a few Qs for me already, and would probably be better suited to your question, I'd stick around to see if you get a reply | 23:09 |
Pirate_Hunter | will do, however it is late evening/night here so not sure how long I can be here for, you could just pastebin your part of the convo for me to read | 23:11 |
udha | not enough to require pastbin: < lifeless> http://www.ubuntu.com/cloud/private has some high level info | 23:12 |
lifeless | udha: UEC provides VM images on-demand using a programming API | 23:12 |
udha | I haven't followed up with anything else yet | 23:12 |
lifeless | udha: so yes, you can use it to do a LAMP based website with automatic scaling; there are various folk writing tools to use the programming API to do just that. | 23:12 |
udha | lifeless: Ok, so it is a Virtual Hyper Visor sortof thing, like VMWare ESX or XEN Server etc? | 23:12 |
lifeless | it builds on top of those | 23:13 |
lifeless | on top of kvm specifically | 23:13 |
lifeless | so you have as 'node controller' machines as you want | 23:13 |
udha | lifeless: thanks, I'd have expected it to use kvm, thanks for clarifying | 23:13 |
lifeless | and when you use the API to start a vm it picks a machine that has enough capacity and starts the vm there, gives it an IP address etc | 23:13 |
udha | lifeless: so I'd need two physical machines to start playing with it right? | 23:14 |
lifeless | if you have an existing network with DHCP and DNS working, you can add a single machine to that network to play with. | 23:14 |
udha | lifeless: I know I started with home use, but for the cloud I'm thinking of datacenter use, no DHCP, just static IPv4/6 address assignments | 23:15 |
lifeless | it works fine in that configuration, though there are bugs that can affect getting started. The Topologies page in the ubuntu help wiki documents the config for a single machine setup | 23:15 |
lifeless | udha: you want dhcp | 23:15 |
lifeless | for this, in the datacentre | 23:15 |
udha | DHCP across 100+ switches and routers? | 23:15 |
lifeless | thats what its designed for | 23:15 |
lifeless | its not really intended for networks the admin can walk around to see each machine :) | 23:16 |
udha | the routers aren't going to forward that traffic without reconfiguration | 23:16 |
udha | hang on, are we talking about the physical machine or the VM deployments? | 23:17 |
lifeless | for the VMs | 23:17 |
Roxyhart0 | hi there somebody know any irc tool for free? | 23:17 |
udha | Roxyhart0: irc 'tool'? like a client? | 23:18 |
udha | Roxyhart0: and for cli or gui? | 23:18 |
lifeless | udha: but you can [and I would argue should] be using it for your machines too | 23:18 |
lifeless | udha: you don't need to use it for your machines; you do for the VM's | 23:18 |
lifeless | otherwise they can't act as templates | 23:18 |
udha | lifeless: DHCP for customer equipment isn't really an option, it would actually complicate management and billing enourmously. | 23:19 |
lifeless | udha: you can statically assign /via/ DHCP | 23:19 |
lifeless | udha: have a reserved range for unknown machines - they would get firewalled off, can use that for bringing up new hardware | 23:19 |
udha | lifeless: each assignment is allocated to a specific port and vlan, unless we go for cisco's Unified fabric I don't know an easy way to automate this | 23:20 |
lifeless | udha: anyhow, point is, you don't need to change your environment to do this; but you will need /a/ DHCP server to serve IP's to the VM instances brought up on node controllers | 23:20 |
udha | lifeless: how about the primary could server? | 23:20 |
udha | run DHCP3 or somethign on it? | 23:20 |
lifeless | quite possibly | 23:21 |
udha | lifeless: well it won't be hard to accomodate the dhcp requirement at any rate | 23:21 |
Roxyhart0 | hi udha icr as client with gui | 23:23 |
Roxyhart0 | urc* | 23:23 |
udha | Roxyhart0: mIRC is a popular one for windows, I've only use cli clients in linux though, I'm using irssi atm | 23:24 |
Roxyhart0 | thanks udha | 23:25 |
hggdh | there are xchat, pidgin, empathy/telepathy (graphical); also irssi and weechat (curses/terminal) | 23:26 |
Roxyhart0 | im looking for windows | 23:26 |
udha | mIRC | 23:26 |
Roxyhart0 | thanks | 23:26 |
udha | unless times have changed since 10 years ago ;P | 23:26 |
cef | there is also a version of xchat for windows.. which is useful if you use xchat on linux having the same thing in both places' | 23:27 |
Pirate_Hunter | no they havent it is still popular just annoying now they've done major changes on the design... well to say I can't use it and not frown | 23:27 |
Pirate_Hunter | udha, thanks for that will continue reading it in 2-3hrs from now need to have some shut eyes before than take care | 23:28 |
udha | Pirate_Hunter: take it easy | 23:29 |
ruben23 | hi anyone used automysqlbackup..? | 23:31 |
udha | ruben32: I havn't, no. What does it do? Anything like MySQL replication sort of thing? | 23:33 |
udha | ruben32: or is it like a mysqladmin database dump wiht cron? | 23:34 |
ruben23 | udha: auto dump and have schedules | 23:35 |
udha | is automysqlbackup an actual program? Or something you want to do? | 23:35 |
ruben23 | its a script, you can run | 23:36 |
ruben23 | on cron | 23:36 |
Roxyhart0 | hi someone know how i can unistall djbdns (tinydns) from a installation followed from http://www.howtoforge.com/perfect-djbdns-setup-on-ubuntu8.04-amd64 | 23:39 |
soren | mdeslaur: Fix committed. | 23:42 |
udha | Roxyhart0: I haven't used that tutorial or djbdns before, but should it be: sudo aptitude remove <packagename> | 23:42 |
soren | mdeslaur: Took a bit longer than I'd hoped, but it should be pretty shiny now. | 23:42 |
* soren heads bedwards | 23:42 | |
mdeslaur | soren: thanks! | 23:56 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!