| === n0ts is now known as n0ts_off | ||
| francisvgarcia | Hi everyone | 00:47 |
|---|---|---|
| francisvgarcia | I am having issues with ubuntu server 12.04 and this network card: intel Corporation 82562V-2 10/100 | 00:49 |
| francisvgarcia | It completely freezes after one or two hours working, and I have to reboot the server for the network card to work again. | 00:50 |
| === n0ts_off is now known as n0ts | ||
| === n0ts is now known as n0ts_off | ||
| === n0ts_off is now known as n0ts | ||
| === n0ts is now known as n0ts_off | ||
| === Lcawte is now known as Lcawte|Away | ||
| === LordOfTime is now known as TheLordOfTime | ||
| === n0ts_off is now known as n0ts | ||
| uvirtbot | New bug: #1003231 in vm-builder (universe) "vmbuilder generates many "method not found" errors" [High,Expired] https://launchpad.net/bugs/1003231 | 04:27 |
| lfactor | Hey guys, i need some help with ufw, i'm wondering if i should turn of stateful support, and if so how. | 04:43 |
| lfactor | The machine will have around 800k simultaneous connections, i'm assuming stateful will increase the memory requirements a lot, but not sure. | 04:44 |
| === MikaT_ is now known as MikaT | ||
| lfactor | i've used /sbin/sysctl -w net.ipv4.netfilter.ip_conntrack_max=1048576 to set my conntrack higher, but still unclear what the positives of a stateful firewall is and if i should have it set on or off. | 05:34 |
| === obelus|2 is now known as obelus | ||
| === jussio1 is now known as jussi | ||
| zaggynl | Why do I get a frozen virtual box vm every other time when I restart my ubuntu guest | 07:38 |
| uvirtbot | New bug: #1036093 in nova (main) "nova volume-attach with high device name keeps volume in state "attaching"" [Undecided,New] https://launchpad.net/bugs/1036093 | 08:51 |
| uvirtbot | New bug: #1030943 in python-swiftclient (universe) "[MIR] python-swiftclient" [Undecided,Fix released] https://launchpad.net/bugs/1030943 | 09:31 |
| === disposab1e is now known as disposable | ||
| JacKnife_ | hello, i'm having trouble with squirrelmail on 12.04, when i compose a message and click send it never browses away from the compose screen, even though the message does get sent | 10:38 |
| JacKnife_ | tried on ie and chrome. the system is all updated and there are no php errors in the apache logs | 10:39 |
| JacKnife_ | nada when i google "squirrelmail compose send" and similar | 10:40 |
| JacKnife_ | w00t, the dudes in #ubuntu got me a fix: http://comments.gmane.org/gmane.mail.squirrelmail.user/38887 | 10:48 |
| === n0ts is now known as n0ts_off | ||
| === Lcawte|Away is now known as Lcawte | ||
| === cpg is now known as cpg|away | ||
| Lachezar | Hello all... Is it normal to have fs.nr_open = 1048576 | 12:41 |
| Lachezar | 'sudo lsof | wc -l' shows 2390 open files/descriptors. And I am getting 'Too Many Open Files' crashes (I've raised limits to 65536 files). | 12:43 |
| zul | jdstrand: ping cinder should be good for main now | 13:10 |
| jdstrand | zul: yeah, it is on my list after I read email | 13:22 |
| zul | jdstrand: ok cool | 13:22 |
| Daviey | zul: can you confirm a pep8 backport fixes this FTBFS, https://launchpad.net/~ubuntu-cloud-archive/+archive/folsom-staging/+packages ? | 14:05 |
| zul | Daviey: sure just a sec | 14:06 |
| Daviey | (hint, zul - don't confirm by uploading :) | 14:08 |
| zul | Daviey: well duh :) | 14:08 |
| * zul quickly hits control-c | 14:09 | |
| Daviey | heh | 14:09 |
| uvirtbot | New bug: #1036206 in google-perftools (universe) "powerpc test suite execution fails" [Undecided,New] https://launchpad.net/bugs/1036206 | 14:16 |
| === fire_ is now known as codingenesis | ||
| === codingenesis is now known as fire_ | ||
| Daviey | zul: any news? | 14:21 |
| zul | Daviey: not yet still building | 14:21 |
| jamespage | utlemming, walinuxagent now in precise-proposed BTW | 14:26 |
| utlemming | jamepsage: awesome :) | 14:26 |
| jamespage | utlemming, do you have a handy way of testing it? we can then nudge it through to -updates ASAP | 14:30 |
| utlemming | jamespage: yeah, I can give that a test rather easily | 14:31 |
| jamespage | utlemming, marvellous! | 14:31 |
| zul | Daviey: confirmed | 14:40 |
| Daviey | zul: confirmed it fixes it? | 14:42 |
| zul | Daviey: confirmed it fixes it | 14:42 |
| Daviey | zul: okay.. what version is it? | 14:43 |
| zul | Daviey: 1.2 from quantal | 14:43 |
| Daviey | zul: wait, i thought 1.2 was evil? | 14:44 |
| Daviey | for folsom? | 14:44 |
| zul | Daviey: it is...not for f1 though | 14:45 |
| Daviey | zul: wow, that much fail got introduced for >f1 ? | 14:48 |
| zul | Daviey: yeah | 14:48 |
| jamespage | xnox, I need todo something with dumbo for you today don't I | 14:48 |
| xnox | hmmm.... jamespage you could =) | 14:49 |
| jamespage | xnox, branch? | 14:49 |
| xnox | jamespage: i have packaging done, but debian/copyright | 14:49 |
| xnox | it's not done yet. | 14:49 |
| xnox | let me push it to lp.net | 14:49 |
| jamespage | xnox, as its PPA not to worried about d/copyright | 14:50 |
| * jamespage slaps himself | 14:50 | |
| Daviey | zul: well, there is a reasonable chance we might need to fall back to 1.1 for folsom | 14:50 |
| jamespage | well for the time being at least | 14:50 |
| Daviey | so doing the same for the cloud archive is reasonable | 14:50 |
| zul | Daviey: ack | 14:50 |
| Daviey | zul: can you upload a dsc and Friends somewhere? | 14:51 |
| zul | Daviey: for pep8? | 14:51 |
| Daviey | zul: yeah | 14:51 |
| zul | Daviey: hold on | 14:52 |
| xnox | jamespage: two branches: lp:~dmitrij.ledkov/+junk/typedbytes and lp:~dmitrij.ledkov/dumbo/packaging | 14:52 |
| xnox | it's two small python packages. | 14:52 |
| xnox | jamespage: feel free to repush to a more appropriate ~person | 14:53 |
| xnox | and if/when it's in the ppa, I can adjust juju charms to optionally include those | 14:53 |
| xnox | there is also pydumbo, but it's slower and I have no experience with it. And dumbo is sufficient so far. | 14:53 |
| xnox | although pydumbo has dfs bindings.... | 14:53 |
| Daviey | zul: so.. first line of the changelog for novas, i set to - nova (2012.2~f1-0ubuntu1~cloud0) precise-folsom; urgency=low .. .changes = "Distribution: precise" .. does that make sense? | 14:54 |
| zul | Daviey: yeah iirc thats what we agreed to | 14:54 |
| zul | Daviey: pep8 stuff is at: http://people.canonical.com/~chucks/tmp/ | 14:55 |
| zul | Daviey: because eventually you are going to have precise-grizzly, precise-h, precise-i, etc ,etc | 14:56 |
| Daviey | zul: right | 14:56 |
| jamespage | xnox, I've pushed them both to the dev PPA | 15:01 |
| jamespage | xnox, all of the hadoop related charms support use of dev|test|stable PPA's for that team | 15:02 |
| xnox | jamespage: cool, thanks =) | 15:02 |
| jamespage | xnox, I really like the idea of not having to write stuff in Java | 15:03 |
| xnox | jamespage: ideally i want to jujufy discoproject map-reduce | 15:05 |
| xnox | which uses tags instead of folders for dfs | 15:05 |
| xnox | and python instead of java for mapreduce | 15:06 |
| xnox | but server part is written in erlang and relies on DNS available for the nodes | 15:06 |
| xnox | but HPCloud doesn't support DNS at the moment | 15:06 |
| xnox | so I'm stuck with both discoproject and HPCloud lacking feature: dns-less setups or dns setup respectfully =) | 15:07 |
| * jamespage sighs | 15:08 | |
| Lachezar | Repeating after a few hours: Is it normal to have fs.nr_open = 1048576 | 15:11 |
| Lachezar | 'sudo lsof | wc -l' shows 2390 open files/descriptors. And I am getting 'Too Many Open Files' crashes (I've raised limits to 65536 files). | 15:11 |
| jdstrand | zul: re cinder> commented in the bug | 15:12 |
| utlemming | jamespage: it looks like walinuxagent hasn't landed in the archive yet...as soon as I see it, I'll test | 15:15 |
| RoyK | hm... seems when I reboot this machine, some drives in my raid come up as "missing" during initial bootup, and I get kicked into busybox. just exiting busybox works, and after that, I can mdadm --stop && mdadm --assemble and mount it - any idea how I can "slow down" this detection or increase the timeouts to avoid this problem? | 15:16 |
| jamespage | utlemming, should be - its in precise-proposed - https://launchpad.net/ubuntu/+source/walinuxagent/1.0~git20120606.c16f5e9-0ubuntu2~12.04.1 | 15:16 |
| jamespage | if its not there after 4 days we have a problem | 15:17 |
| xnox | RoyK: if you are using precise, please upgrade to mdadm from -precise, as I commented on your bug report? | 15:17 |
| xnox | from -proposed that is. | 15:17 |
| xnox | it has an extra timeout to wait for udev to finish processing events, before dropping into busybox, which helps most people. | 15:17 |
| utlemming | jamespage: duh, my apt sources.list was wrong | 15:17 |
| jamespage | lol | 15:18 |
| RoyK | xnox: how can I upgrade to that from -proposed? | 15:20 |
| zul | jdstrand: damn it...*grumble* *grumble* | 15:20 |
| RoyK | xnox: this is precise, btw | 15:21 |
| uvirtbot | New bug: #1036240 in cinder (universe) "cinder-common fails to install" [High,New] https://launchpad.net/bugs/1036240 | 15:21 |
| xnox | !proposed | 15:21 |
| xnox | RoyK: https://wiki.ubuntu.com/Testing/EnableProposed | 15:22 |
| RoyK | xnox: thanks | 15:22 |
| === dendrobates is now known as dendro-afk | ||
| RoyK | xnox: \o/ | 15:27 |
| === dendro-afk is now known as dendrobates | ||
| hallyn | zul: did a qa-regression-test run just fora sanity's sake, all still looks good. just lettin' you know cause i'm sure you're unable to sleep at nights worrying about it | 15:52 |
| zul | hallyn: libvirt? | 15:53 |
| hallyn | zul: yeah | 15:53 |
| zul | hallyn: coolness | 15:53 |
| hallyn | zul: do you know of anything we still need to do to libvirt during q? | 15:53 |
| zul | hallyn: nope just make sure it doesnt break | 15:54 |
| zul | hallyn: although i hope we can get the new libvirt-lxc stuff in for q | 15:54 |
| hallyn | which new stuff? | 15:54 |
| zul | hallyn: like the lxc reboot | 15:55 |
| hallyn | do you know where that went in? is it in 0.9.14? | 15:55 |
| zul | i think it is in trunk | 15:56 |
| hallyn | cause i assume that went in after those 500 'let's rename stuff for fun 'patches, so forget about backporting | 15:56 |
| utlemming | jamespage: confirmed | 15:56 |
| jamespage | utlemming, great - nice one | 15:56 |
| zul | hallyn: yeah thats why i want trunk :) | 15:56 |
| utlemming | jamespage: I fired up a couple of instances to be sure. | 15:56 |
| hallyn | by trunk you mean git head? | 15:56 |
| hallyn | (not trying to be pedantic, justnot sure what you mean) | 15:57 |
| zul | hallyn: ack | 15:59 |
| hallyn | k | 15:59 |
| souliaq | Hi, I have a little "legal" licensing problem in my company, so the lawyer are asking my for license of Ubuntu-Server, apache and subversion. What he need? GPLv3 and Apache License "texts" and tha'ts all? | 16:09 |
| === skaet_ is now known as skaet | ||
| xnox | Anyone has a spare intel matrix raid controller? | 17:18 |
| xnox | souliaq: tarball of /usr/share/common-licenses/ as well as /usr/share/doc/*/copyright | 17:19 |
| RoyK | xnox: is that real raid or fakeraid? | 17:27 |
| RoyK | looks like fakeraid to me | 17:28 |
| RoyK | better use software raid :) | 17:28 |
| xnox | RoyK: it's not real-real, but it's usually managed with dmraid but recent mdadm can store external metadata using intel matrix format | 17:28 |
| xnox | and i want to test that, cause I am about to update mdadm in precise | 17:29 |
| RoyK | ok | 17:29 |
| RoyK | xnox: will I have to update mdadm manually when you're done with the precise update, currently using the one in proposed? | 17:37 |
| xnox | RoyK: no you wont. The one in -proposed will be promoted into -updates pocket, such that everyone will get it and it will be included in the 12.04.1 | 17:38 |
| RoyK | thanks | 17:39 |
| xnox | RoyK: your welcome =) | 17:39 |
| * RoyK wonders slightly if bcache will make it into upstream kernel ;) | 17:41 | |
| smoser | hm.. i have this utility http://smoser.brickies.net/git/?p=tildabin.git;a=blob;f=make-seed-disk;hb=HEAD | 17:44 |
| smoser | that i'd like to have packaged. cloud-utils seems reasonablel place for it | 17:45 |
| smoser | but it would add a depends on genisoimage (and probably a 'Suggests:' for mtools) | 17:46 |
| smoser | i was going to name it "cloud-localds" (local datasource) | 17:46 |
| smoser | anyone hav ea better idea than its own binary package of cloud-utils ? | 17:46 |
| smoser | utlemming, you want to do that^ ? | 17:51 |
| smoser | i cannot do it today for sure. | 17:51 |
| utlemming | smoser: yeah...I think I can give it a shot...are we thinking of a subpackage of "cloud-utils-localds" to the cloud-utils package? | 17:52 |
| utlemming | or just adding it in | 17:52 |
| smoser | https://bugs.launchpad.net/cloud-utils/+bug/1036312 | 17:55 |
| uvirtbot | Launchpad bug 1036312 in cloud-utils "please add cloud-localds from make-seed-disk" [Undecided,New] | 17:55 |
| smoser | utlemming, ^ i think a subpackage is best | 17:55 |
| utlemming | smoser: ack, we're on the same page | 17:56 |
| === tyhicks` is now known as tyhicks | ||
| === higgs_ is now known as Guest46041 | ||
| === cpg|away is now known as cpg | ||
| smw_ | rk | 18:47 |
| === smw_ is now known as smw_work | ||
| zul | jdstrand: so how you would you handle that cinder.conf bug? | 18:50 |
| === fire_ is now known as codingenesis | ||
| === codingenesis is now known as fire_ | ||
| jdstrand | zul: well, I don't know the issue intimately-- seems we should be shipping our own cinder.conf or patching the one in source before moving it into place. | 18:57 |
| zul | jdstrand: i was thinking something like ucf | 18:57 |
| jdstrand | well, it does say this: | 18:57 |
| jdstrand | The root_helper option (which lets you specify a root wrapper different from cinder-rootwrap, and defaults to using sudo) is now deprecated. You should use the rootwrap_config option instead. | 18:57 |
| jdstrand | zul: did you you root_helper instead of rootwrap_config? | 18:58 |
| jdstrand | s/you you/you use/ | 18:58 |
| zul | jdstrand: its in the cinder.conf for the new version of cinder | 18:58 |
| jdstrand | zul: you misunderstood | 19:01 |
| jdstrand | zul: your installed cinder.conf uses: | 19:02 |
| jdstrand | [DEFAULT] | 19:02 |
| jdstrand | root_helper = sudo /usr/sbin/cinder-rootwrap | 19:02 |
| zul | right | 19:02 |
| jdstrand | the error says that root_helper is deprecated. use rootwrap_config instead | 19:02 |
| zul | jdstrand: ahhhhhh....duh :) | 19:02 |
| jdstrand | so: s/root_helper/rootwrap_config/ in cinder.conf (doing whatever else you need use rootwrap_config) | 19:03 |
| jdstrand | heh, right :) | 19:03 |
| zul | jdstrand: okies fixed | 19:16 |
| jdstrand | cool | 19:18 |
| antihero | is there a way to have upstart run stuff as other users yet? | 19:21 |
| === dendrobates is now known as dendro-afk | ||
| === dendro-afk is now known as dendrobates | ||
| hdave | is Ubuntu JeOS and vmbuilder still actively developed? I ask because the vm-builder launchpad site has a 20 month old download link and there also doesn't seem to be a JeOS 12.04 image anywhere... Just curious | 19:49 |
| RoyK | jeos isn't a separate iso any longer | 19:51 |
| ssvss | Hi, I have a hardware related query. can anyone suggest a server hardware under USD $500 to run ubuntu server in home. | 19:54 |
| ssvss | I am looking for something that is portable too like the box shape of mac mini. | 19:54 |
| Psi-Jack | ssvss: ##hardware would be your channel. | 19:54 |
| ssvss | Thanks, I will ask in the ##hardware channel | 19:56 |
| hdave | RoyK: thanks | 19:57 |
| cheez0r | ssvss: raspberry pi. | 19:59 |
| RoyK | ssvss: a bit hard to use sata devices on a pi | 19:59 |
| wrapids | How much ram should lamp be using without any traffic? | 20:01 |
| RoyK | wrapids: lamp is apache, mysql, php, and may be using variable amounts of memory | 20:02 |
| wrapids | RoyK: Yes. | 20:02 |
| RoyK | wrapids: for a small database, mysql won't be using much. php may be using a lot, depending on the code | 20:02 |
| wrapids | RoyK: That's assuming I have traffic | 20:02 |
| wrapids | which I dont. | 20:02 |
| RoyK | apache isn't that heavy on memory | 20:02 |
| RoyK | say, 50 megs will go a long way without too much work | 20:03 |
| wrapids | php shouldn't be using much of anything as nothing is being exectued. There are no queries going on in the db either | 20:03 |
| wrapids | Would the database size affect the mysql services usage if it's not getting any queries? | 20:03 |
| RoyK | once php starts running things, and mysql starts buffering things, say, 512MB should normally do well | 20:03 |
| RoyK | but then, you can't say unless you know the database size and the php code | 20:04 |
| wrapids | RoyK: Wouldn't running/buffering require traffic? | 20:04 |
| ssvss | Yes Rasberry Pi is not what I am looking for, I was thinking something close to the size fo mac mini in which I can have 2 sata disks | 20:04 |
| RoyK | nothing is buffered unless it is accessed | 20:04 |
| wrapids | I'm trying to figure out why I'm using nearly 512mb with 0 traffic, nothing being executed, no queries. | 20:04 |
| RoyK | ssvss: you can get some mini itx boards quite cheap with SATA | 20:04 |
| wrapids | hrm, after a reboot it's doing better | 20:06 |
| RoyK | ssvss: or pico itx or pc/104 or ... | 20:06 |
| wrapids | 188mb with no traffic? | 20:06 |
| RoyK | wrapids: is that RSS or DRS? | 20:06 |
| wrapids | RoyK: I'm not sure how to determine | 20:07 |
| RoyK | ps axfv | 20:07 |
| RoyK | top also tells that | 20:07 |
| raubvogel | RoyK: there is always cubox | 20:07 |
| wrapids | I've been using top | 20:07 |
| RoyK | top shows VIRT and RES and SHR | 20:08 |
| wrapids | mysql is using about 50mb idling, apache using about 25mb idling | 20:08 |
| RoyK | what you want to look for is RES | 20:08 |
| wrapids | Sorry, about 40 for apache | 20:08 |
| RoyK | should be fine | 20:09 |
| wrapids | It was running in the several hundreds before I rebooted | 20:09 |
| wrapids | free -m was giving me 11 free with only apache/mysql using above .5% | 20:09 |
| RoyK | resident or virtual? | 20:09 |
| wrapids | res | 20:09 |
| RoyK | free usually shows very low "free" memory | 20:09 |
| RoyK | most of the memory is spent on caching | 20:09 |
| wrapids | It was fairly accurate compared to the top results | 20:09 |
| RoyK | Mem: 8178284 8056176 122108 0 469068 6906568 | 20:10 |
| wrapids | apache had 6-10 processes using about 5-10% each | 20:10 |
| RoyK | that's close to zero free | 20:10 |
| RoyK | which is fine | 20:10 |
| wrapids | according to top | 20:10 |
| RoyK | because you want linux to spend its memory on caching | 20:10 |
| RoyK | wrapids: seriously - if you don't have a performance issue, don't care about how much memory is spent | 20:11 |
| wrapids | RoyK: I do have a performance issue when it starts doing that | 20:11 |
| RoyK | does it start swapping? | 20:11 |
| wrapids | I have no idea | 20:11 |
| RoyK | how much memory do you have? | 20:11 |
| wrapids | 512 | 20:11 |
| wrapids | It's just a dev server | 20:11 |
| RoyK | not a whole lot | 20:11 |
| wrapids | the problem is that it starts eating ram like that with nothing going on, I get very delayed response from the ssh interface | 20:11 |
| RoyK | no idea why | 20:12 |
| wrapids | hrm | 20:12 |
| RoyK | check the logs | 20:12 |
| RoyK | and check swap use | 20:12 |
| RoyK | wrapids: apache uses prefork with php, so each of te processes weren't using 5-10% each, they were probably sharing most of that | 20:17 |
| wrapids | ah | 20:17 |
| === bitmonk_ is now known as bitmonk | ||
| jkyle | I have some interfaces configured for bonding. I have to bring up each of the slave interfaces before bringing up the bond0 interface or it times out waiting for slaves to be available | 20:45 |
| jkyle | https://gist.github.com/3343969 | 20:46 |
| jkyle | e.g. Waiting for a slave to join bond0 (will timeout after 60s) | 20:46 |
| jkyle | if I do, ifup eth2; ifup eth3;ifup bond0; it works | 20:46 |
| arrrghhh | hey all. can anyone help me shrink an LVM partition? | 21:03 |
| arrrghhh | i booted a livecd and tried to shrink it thru gparted, but i guess gparted doesn't support LVM | 21:03 |
| arrrghhh | so then i figured i had to remove it from LVM to get gparted to see it... that hasn't worked out so far. | 21:04 |
| arrrghhh | i removed and added back a logical volume... and i have a feeling it's FUBAR now. i can't seem to get the system to boot. | 21:05 |
| arrrghhh | is there a way to recover it, or should i just reinstall? | 21:05 |
| xnox | arrrghhh: what do you actually want to resize and what is it stacked on top of? | 21:07 |
| xnox | the whole chain | 21:07 |
| arrrghhh | so there's a set physical SAS disks | 21:07 |
| arrrghhh | then i have physical volumes setup, and logical volumes underneath it | 21:08 |
| arrrghhh | i'd like to shrink one logical AND physical volume | 21:08 |
| arrrghhh | then shrink the actual amount provided to the OS | 21:08 |
| arrrghhh | this is in an ESXi environment, and i'd like to reclaim a bit i've allocated | 21:08 |
| arrrghhh | i fear i've already done too much. i removed the LV, and readded a smaller one - now the OS won't boot, and I'm not sure if it can be recovered. | 21:08 |
| arrrghhh | makes me wish i had snapshotted it before doing all this... oy | 21:09 |
| xnox | arrrghhh: you are doing it wrong way around | 21:09 |
| arrrghhh | OK | 21:09 |
| xnox | first you shrink the OS filesystem. | 21:09 |
| xnox | then you shrink logical volume | 21:09 |
| xnox | then you shrink physical volume | 21:09 |
| xnox | then you can shrink the partition | 21:09 |
| arrrghhh | my issue was #1 - i couldn't shrink the OS filesystem when it's mounted | 21:09 |
| arrrghhh | so i went to a liveCD, and that didn't support LVM | 21:10 |
| xnox | ok. | 21:10 |
| arrrghhh | (gparted doesn't support LVM rather) | 21:10 |
| xnox | in livecd you install lvm2 package | 21:10 |
| arrrghhh | ok, done :) | 21:10 |
| xnox | then you scan lvm groups | 21:10 |
| arrrghhh | yes | 21:10 |
| xnox | then you mount the logical volume you want to shrink | 21:10 |
| arrrghhh | ok let me try | 21:10 |
| xnox | then you start shrinking that filesystem | 21:10 |
| xnox | then lvresize the logical volume | 21:11 |
| xnox | or lvreduce | 21:11 |
| xnox | and etc. downwords | 21:11 |
| xnox | good night | 21:11 |
| arrrghhh | thanks | 21:11 |
| arrrghhh | hrm. xnox are you leaving? | 21:11 |
| arrrghhh | i'll take that as a yes. can anyone else lend a hand with LVM? | 21:12 |
| arrrghhh | i am trying to learn about it the hard way, as usual. | 21:12 |
| arrrghhh | i have a LV name of /dev/ubuntu/root in lvs, but i can't seem to mount it... | 21:13 |
| arrrghhh | anyone? is the data still on the physical volume perhaps? can i just remove LVM and use all the data on the disk? | 21:14 |
| === dendrobates is now known as dendro-afk | ||
| arrrghhh | perhaps someone can help me with that? | 21:25 |
| arrrghhh | remove LVM, preserve data | 21:25 |
| arrrghhh | perhaps restore LVM down the road once i get more comfy with it ;) | 21:26 |
| SpamapS | smoser: hey, have we ever considered putting the cloud images into the archive as packages? | 21:29 |
| SpamapS | utlemming: ^^ | 21:29 |
| smoser | SpamapS, no. | 21:30 |
| smoser | ther ewas a thread once on debian-devel (or maybe ubuntu-devel) | 21:30 |
| smoser | about "appliance" packages | 21:30 |
| SpamapS | We're trying to solve the "how to have everything cached for LXC on install" problem | 21:30 |
| smoser | i forget how started it | 21:30 |
| smoser | its just hucky | 21:30 |
| SpamapS | well the thinking is that users are used to downloading things with the package manager | 21:30 |
| smoser | well, we certainly want ot make downloding of those simple and "cached" | 21:30 |
| smoser | there is a plan for that. | 21:31 |
| m_3 | oh do tell | 21:31 |
| SpamapS | and things like update-manager is pretty good at downloading in the background and stuff... | 21:31 |
| m_3 | smoser utlemming: so we've been throwing around a couple of ideas... | 21:34 |
| m_3 | doing nothing will result in the juju local provider being confusing to use (due to the initial "stealth" download of the lxc image on the first deploy) | 21:35 |
| m_3 | one idea was to bust the juju package up into 'juju' and 'juju-local-provider'... the latter downloads the lxc image during postinst | 21:36 |
| m_3 | with a couple of variations on that theme | 21:36 |
| m_3 | smoser utlemming: these all suck... I want an easy (or at least idiomatic to packaging) way to download images | 21:38 |
| smoser | m_3, i'm sorry. i really ahve to run right now. | 21:38 |
| m_3 | smoser: no prob... lemme know if you think of anything pls | 21:39 |
| utlemming | m_3: when you say "package" are you looking for a pacakge that does the download? | 21:39 |
| m_3 | utlemming: sure.. or even a package that _is_ the download 'juju-local-provider-data' would work | 21:39 |
| SpamapS | I'd prefer the package to *contain the images* | 21:40 |
| SpamapS | my reasoning being that postinsts doing downloads is counter-intuitive when we have *package managers* to do downloads. | 21:40 |
| utlemming | SpamapS: yikes....that would mean SRU'ing each and every new spin of the images. | 21:40 |
| m_3 | good with the least sucky variation at this point :) | 21:41 |
| SpamapS | utlemming: MRE would be pretty easy to get given the contents are just the same packages already SRU'd ;) | 21:41 |
| utlemming | SpamapS: right now we spin up new images as needed, and next cycle we are heavily considering a 3-week new release cadiance | 21:41 |
| m_3 | it might be able to live in a ppa | 21:41 |
| utlemming | SpamapS: what does a packaging of the image offer over a post-install that downloads and verifies? | 21:42 |
| === cpg is now known as cpg|away | ||
| SpamapS | utlemming: uniformity and discoverability | 21:42 |
| SpamapS | utlemming: its useful as a Suggests: in many cases (glance.. virt-manager) | 21:43 |
| SpamapS | utlemming: the download behind the scenes is very mysterious. Downloading a new version of the cloud images as a package is obvious. | 21:43 |
| utlemming | SpamapS: I not opposed to this, I'm just unsure about how to make this atomic. | 21:44 |
| SpamapS | utlemming: after you publish the cloud image, you run a package build including get-orig-source which downloads, dch's, and uploads the updated -data package. | 21:44 |
| utlemming | Spamaps: So if you update the package to a new spin of the images, then all of the sudden you have people that were wanting YYYYMMDD are now getting a different YYYYMMDD | 21:45 |
| utlemming | right, I understand that bit | 21:45 |
| utlemming | its want users are going to expect, and what they get | 21:45 |
| SpamapS | utlemming: its really more like the kernel than regular packages. I could see ubuntu-cloud-images-current depending on the latest cloud image.. but each one would get its own package (ubuntu-cloud-images-20120813) | 21:47 |
| utlemming | it would be worse than that....it would have to be ubuntu-cloud-images-<release>-<build serial> | 21:48 |
| utlemming | since not all images are build on the same day and the build serial is almost never the same between releases. | 21:49 |
| * m_3 nods | 21:49 | |
| SpamapS | utlemming: I don't see that as "worse" | 21:51 |
| SpamapS | utlemming: just different, so each release would have its own meta | 21:51 |
| utlemming | well, its worse because this a manual process | 21:51 |
| SpamapS | ubuntu-cloud-image-precise would -> ubuntu-cloud-image-precise-YYYYMMDD | 21:51 |
| SpamapS | manual would be out of the question | 21:51 |
| utlemming | we are looking at automating all aspects of the builds next cycle | 21:51 |
| SpamapS | It should be an idempotent thing that gets run after images are published | 21:52 |
| utlemming | I'm showimg my packaging ignorance here, but is there a way to automate this? | 21:52 |
| SpamapS | utlemming: totally! | 21:52 |
| utlemming | docs? | 21:52 |
| * SpamapS points to the packaging guide | 21:52 | |
| SpamapS | utlemming: all packaging is automatable. | 21:53 |
| SpamapS | you're just used to doing it the most manual way | 21:53 |
| SpamapS | because tht is less error prone | 21:53 |
| SpamapS | but if all you are doing is bumping upstream version.. very simple | 21:53 |
| utlemming | right, that bit makes sense. Do we give upload rights to bots? | 21:54 |
| SpamapS | debian/rules get-orig-source && dch -v 12.04.1-20120813 'New Upstream Release' && dpkg-buildpackage | 21:55 |
| SpamapS | something like that | 21:55 |
| SpamapS | utlemming: who signs your cloud images? | 21:55 |
| utlemming | that is an automated process | 21:56 |
| SpamapS | that key is pretty much entrusted with all cloud users' safety.. :) | 21:56 |
| utlemming | this is true | 21:56 |
| SpamapS | so yes, that would be trustable | 21:56 |
| utlemming | okay, so I am not opposed to this plan | 21:56 |
| utlemming | lets ping Mr. Rosales and talk about this next team meeting | 21:56 |
| SpamapS | utlemming: aye | 21:57 |
| Daviey | utlemming: uscan & uupdate also make new upstream versions pretty easy | 21:59 |
| === Ursinha` is now known as Ursinha | ||
| Vamps-AFK | has anyone setup a PXE Install server with FOG? i'm got a question for those who have | 22:08 |
| zastern | What does it take for Ubuntu to automatically "discover" it's correct fqdn? | 22:19 |
| zastern | rather than setting it in /etc/hosts | 22:20 |
| arrrghhh | zastern, dnsmasq? | 22:25 |
| zastern | arrrghhh: so I have to use some sort of private dns server? | 22:25 |
| arrrghhh | zastern, just trying to think of how to solve that issue | 22:25 |
| arrrghhh | you'd need some sort of a DNS server in order for the FQDN to be automatically propagated | 22:26 |
| three18ti | what's the default username for cobbler on an Ubuntu install? The communitydocs say the install prompts for the password, but it does not. | 22:28 |
| three18ti | I found a way to "define" the password, http://openskill.info/topic.php?ID=201 | 22:28 |
| arrrghhh | three18ti, cobbler/cobbler? | 22:30 |
| three18ti | no love. that's what all the docs for fedora say though... | 22:30 |
| three18ti | `htdigest /etc/cobbler/users.digest "Cobbler" cobbler` | 22:30 |
| arrrghhh | hrm | 22:31 |
| three18ti | allowed me to "reset" the password though. | 22:31 |
| three18ti | maybe there's a way to update the docs? | 22:31 |
| three18ti | (actually, it's all coming back to me, this is not the first time I've chased myself in circles following the Ubuntu cobbler docs) | 22:32 |
| arrrghhh | lol | 22:32 |
| three18ti | ;) | 22:32 |
| arrrghhh | never heard of cobbler before, just googled hoping i could help ;) | 22:32 |
| three18ti | from what I've read cobbler is pretty cool. I've looked at a number of bare-metal provisioning systems and so far it looks the best. | 22:33 |
| arrrghhh | fancy | 22:33 |
| arrrghhh | just looked it up myself | 22:34 |
| three18ti | FAI, cobbler, linmin, OpenQRM (though openqrm is a whole 'nuther ball game) | 22:34 |
| arrrghhh | so based on the install guide, you set the password on install | 22:34 |
| arrrghhh | https://help.ubuntu.com/community/Cobbler/Installation | 22:34 |
| three18ti | yea, that's what I'm saying, the install guide is wrong. | 22:34 |
| arrrghhh | okie :) | 22:34 |
| arrrghhh | ah, i missed that line. haha. | 22:35 |
| three18ti | lol. :) | 22:35 |
| === zyga is now known as zyga-afk | ||
| arrrghhh | so can anyone help me save my LVM setup? | 22:55 |
| arrrghhh | *crickets* | 23:03 |
| zastern | arrrghhh: yes but I am using a public DNS already | 23:05 |
| zastern | with nsrecords etc | 23:05 |
| zastern | i wonder if theres a way to do it with public dns | 23:06 |
| arrrghhh | zastern, not that i know of, how would that work? | 23:06 |
| zastern | arrrghhh: no idea :) | 23:06 |
| arrrghhh | everyone populating their local DNS friendly names to the wide internet? | 23:06 |
| zastern | no | 23:06 |
| zastern | im setting it manually of course | 23:06 |
| zastern | i was just hoping there was a way for ubuntu to pull thast in | 23:06 |
| zastern | that in* from the dns server where i set it | 23:06 |
| arrrghhh | hrm | 23:07 |
| arrrghhh | crap. looks like i should just start over. | 23:13 |
| arrrghhh | can anyone help me start over? lol. i just want to make sure my other untouched and fine LVM is restored | 23:16 |
| arrrghhh | http://tldp.org/HOWTO/LVM-HOWTO/recipemovevgtonewsys.html | 23:19 |
| arrrghhh | looks like that covers it. | 23:19 |
| grendal | ok i just need to be able to send an email from this server... | 23:40 |
| grendal | i have an account on an email server and an smtp server address to us | 23:40 |
| grendal | i would like to set up a smarthost... | 23:40 |
| grendal | i got to tell you the step by step configureation verbage in dpkg-reconfigure exim4 makes no sence | 23:40 |
| === Lcawte is now known as Lcawte|Away | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!