*** GasbaKid has quit IRC | 00:03 | |
*** nelson has quit IRC | 00:06 | |
*** nelson has joined #openstack | 00:07 | |
*** chmouel has quit IRC | 00:07 | |
*** nelson has quit IRC | 00:07 | |
*** nelson has joined #openstack | 00:08 | |
*** chmouel has joined #openstack | 00:08 | |
*** _vinay has joined #openstack | 00:10 | |
_vinay | Hi vishy | 00:10 |
---|---|---|
_vinay | I want to know if there are any performance / stress tests written | 00:11 |
_vinay | We are evaluating openstack | 00:11 |
_vinay | and are trying to get hold of some perf tests which we can run | 00:12 |
*** jeffjapan has joined #openstack | 00:15 | |
*** adjohn has joined #openstack | 00:15 | |
*** jaypipes has quit IRC | 00:17 | |
*** deshantm has quit IRC | 00:17 | |
*** jaypipes has joined #openstack | 00:17 | |
hallyn | kbringard: no i'm not, using my own images | 00:18 |
hallyn | drat, missed him. by a long shot | 00:18 |
*** kashyap has quit IRC | 00:19 | |
*** deshantm has joined #openstack | 00:19 | |
*** kashyap has joined #openstack | 00:20 | |
*** dendrobates is now known as dendro-afk | 00:27 | |
*** kashyap has quit IRC | 00:33 | |
*** BK_man has joined #openstack | 00:37 | |
BK_man | jaypipes: around? | 00:38 |
*** gregp76 has quit IRC | 00:38 | |
*** thingee has joined #openstack | 00:40 | |
*** joearnold has quit IRC | 00:41 | |
*** dragondm has quit IRC | 00:43 | |
*** rchavik has quit IRC | 00:52 | |
*** kashyap has joined #openstack | 00:56 | |
*** DigitalFlux has quit IRC | 00:56 | |
*** mahadev has quit IRC | 00:57 | |
*** Dumfries has quit IRC | 00:59 | |
*** woleium has quit IRC | 00:59 | |
*** BK_man has quit IRC | 01:02 | |
*** clauden_ has quit IRC | 01:06 | |
*** agarwalla has left #openstack | 01:06 | |
*** winston-d has joined #openstack | 01:12 | |
*** mahadev has joined #openstack | 01:31 | |
*** rchavik has joined #openstack | 01:37 | |
*** mahadev has quit IRC | 01:41 | |
*** pquerna has quit IRC | 01:42 | |
*** pquerna has joined #openstack | 01:42 | |
*** thingee has quit IRC | 01:44 | |
*** CrazyGir has quit IRC | 01:44 | |
*** holoway has quit IRC | 01:45 | |
*** holoway has joined #openstack | 01:45 | |
*** zykes- has quit IRC | 01:45 | |
*** CrazyGir has joined #openstack | 01:45 | |
*** CrazyGir is now known as Guest74692 | 01:46 | |
*** Ryan_Lane has quit IRC | 01:46 | |
*** mahadev has joined #openstack | 01:48 | |
*** obino has quit IRC | 01:56 | |
*** jakedahn has joined #openstack | 01:56 | |
_nil | kpepple: i'm still not sure the best place for me to help | 01:57 |
_nil | i don't need hand holding | 01:57 |
_nil | just direction | 01:57 |
_nil | :) | 01:57 |
winston-d | kpepple : hi | 01:58 |
winston-d | kpepple : what would happen if certain instance_type was removed from database but there's an instance w/ that type still running? | 02:01 |
*** rchavik has quit IRC | 02:04 | |
kpepple | winston-d: hi | 02:04 |
kpepple | winston-d: that shouldn't be a problem ... we originally didn't have a foreign key in the instances table back to instance_types but we added one just before cactus release ... still i think that we just get an empty list if we query for the instance_type that was removed. also, it's hard to actually remove an instance_type from the table, you have to use the --purge option not just delete | 02:06 |
kpepple | _nil: hi | 02:06 |
winston-d | kpepple : yeah, i know. i tested create/delete instance_types and found i could not add new one with that flavor id which was deleted. then i go check your blog and find out --purge must be used to actually remove it from DB. | 02:08 |
kpepple | _nil: areas that can use help: document (see http://wiki.openstack.org/Documentation/HowTo), testing (learn python mox), answers (help at https://answers.launchpad.net/nova) and lastly coding (yes, you need to know Python but it's pretty easy to learn if you have some object oriented experience) | 02:09 |
kpepple | winston-d: in the new version of nova (trunk), you actually get a nice new error message which shows you which instance_types are duplicates :) | 02:09 |
_nil | kpepple: i know python | 02:11 |
*** jakedahn has quit IRC | 02:11 | |
_nil | but it doesn't seem like you need coders as you list it last. | 02:11 |
kpepple | _nil: oh, sorry. thought you said you were just doing c and go. if so, | 02:11 |
_nil | well those are my favorites | 02:11 |
kpepple | _nil: oh no, we need coders | 02:11 |
_nil | ok | 02:11 |
_nil | well i'm game to help when i can | 02:11 |
_nil | i want to start using this on duckduckgo | 02:12 |
_nil | s/on/for/g | 02:12 |
kpepple | _nil: i jumped in with starting bug fixes ... it's a way to learn your way around codebase in small chunks | 02:12 |
_nil | i just need to know where the hot spots are | 02:12 |
KnightHacker | _nil: Are you planning on using Nova for duckduckgo? | 02:12 |
_nil | considering it | 02:12 |
_nil | more fun to work on open source stuff first and then see if it's useful | 02:13 |
_nil | this project caught my eye and i need a break from the go compiler tbh | 02:13 |
KnightHacker | wow .. that sounds cool :). I was just checking out the architecture used for duckduckgo | 02:13 |
_nil | cool | 02:13 |
_nil | growth has been pretty crazy over there for some time. | 02:14 |
KnightHacker | Yeah, that's why I was interested to see how are you handling the scale with your infrastructure. | 02:14 |
_nil | ;) | 02:15 |
KnightHacker | I was surprised but glad to see you using *BSD systems there. | 02:15 |
_nil | both the founder and myself are pretty big FreeBSD nerds | 02:15 |
kpepple | _nil: i tend to work on the nova operational side (things like nova-manage, instance_types, etc that ops people like to use), but there is a ton of oppty everywhere. volume support is being worked on (nova-volume), there is a new load balancer project, networking is being expanded and reworked, authn/authz is becoming a new services ... i'd just look for something that interests you and dive in | 02:15 |
_nil | though i've been using ubuntu a lot recently | 02:15 |
_nil | ok | 02:16 |
_nil | kpepple: ^ | 02:16 |
*** obino has joined #openstack | 02:16 | |
creiht | _nil: A good starting point would be to do a dev install and kick the tires a bit | 02:16 |
* _nil fires up git-bzr-ng | 02:16 | |
_nil | creiht: *nod* | 02:16 |
_nil | getting after it | 02:16 |
_nil | see what strikes me :) | 02:16 |
kpepple | _nil: you'll be happy to hear that the codebase is moving to github ... soonish | 02:17 |
*** pquerna has quit IRC | 02:17 | |
*** pquerna has joined #openstack | 02:17 | |
winston-d | kpepple : does nova trunk now expose the '--purge' parameter for instance_type delete command, IMHO, it should be. | 02:18 |
_nil | kpepple: neat. thanks guys. | 02:18 |
kpepple | _nil: np | 02:19 |
kpepple | winston-d: what do you mean by 'expose' ? yes, it accept it | 02:20 |
winston-d | kpepple : i thought '--purge' was already in Cactus release but 'nova-mange instance delete' only suggest one argument, that's name of instance_type. | 02:21 |
kpepple | winston-d: oh, no it doesn't actually suggest that to you ... you have to read the man page to learn that sekrit :) | 02:24 |
kpepple | winston-d: and yes, it should be how nova-manage is written, it doesn't do it automagically | 02:24 |
winston-d | kpepple : so '--purge' was hidden secret agrument in Cactus :) | 02:25 |
winston-d | kpepple : hmm, no man entry for 'nova-mange' | 02:25 |
*** hadrian has quit IRC | 02:26 | |
kpepple | winston-d: there should be ... it's in the bzr repo (nova/doc/man/nova-manage) | 02:26 |
winston-d | kpepple : ok. thanks. i'll suggest BK_man to pack that into RHEL packages | 02:27 |
kpepple | winston-d: cool, i think i saw Andrey online earlier | 02:29 |
winston-d | kpepple : guess he's still in US. i haven't seen him during my timezone for days. | 02:30 |
uvirtbot | New bug: #777537 in nova "Multiple unit tests generate errors when run on a machine with nova services running" [Undecided,New] https://launchpad.net/bugs/777537 | 02:31 |
*** rchavik has joined #openstack | 02:35 | |
*** rchavik has quit IRC | 02:39 | |
*** Transformer has joined #openstack | 02:44 | |
*** Transformer has quit IRC | 02:46 | |
*** maplebed has quit IRC | 02:48 | |
hallyn | heh, well apparently i just suck. I finally gave up on doing it by hand and made a tarball for smoser's uec-publish-tarball to consume, and that worked :) | 02:54 |
hallyn | @smoser++ | 02:54 |
larissa | hallyn: Error: "smoser++" is not a valid command. | 02:54 |
hallyn | oops | 02:54 |
*** zykes- has joined #openstack | 02:57 | |
smoser | hallyn, just fyi, uec-publish-tarball sits atop uec-publish-image | 03:04 |
smoser | which may have been suitable for you. | 03:04 |
*** msivanes has joined #openstack | 03:04 | |
*** msivanes has left #openstack | 03:04 | |
*** msivanes has joined #openstack | 03:05 | |
*** msivanes has left #openstack | 03:09 | |
*** msivanes has joined #openstack | 03:10 | |
*** zns has joined #openstack | 03:15 | |
hallyn | smoser: looking at that, thanks | 03:18 |
hallyn | yeah that looks good | 03:19 |
hallyn | (save me some needless file mangling) | 03:19 |
hallyn | I just don't get what I was doing wrong by hand though | 03:19 |
*** rchavik has joined #openstack | 03:21 | |
*** zns has joined #openstack | 03:23 | |
*** jero has joined #openstack | 03:26 | |
*** jero has left #openstack | 03:27 | |
*** mahadev has quit IRC | 03:29 | |
*** msivanes has quit IRC | 03:39 | |
*** mahadev has joined #openstack | 03:40 | |
*** msivanes has joined #openstack | 03:43 | |
*** jtimberman has quit IRC | 03:49 | |
*** paltman has quit IRC | 03:50 | |
*** paltman has joined #openstack | 03:51 | |
*** mahadev has joined #openstack | 03:57 | |
*** ewindisch has quit IRC | 04:11 | |
*** yamahata_lt has joined #openstack | 04:25 | |
*** mahadev has quit IRC | 04:34 | |
*** mahadev has joined #openstack | 04:44 | |
*** amadeus_ has quit IRC | 04:47 | |
*** jmckind has quit IRC | 04:49 | |
*** f4m8_ is now known as f4m8 | 04:51 | |
*** f4m8 is now known as f4m8_ | 04:51 | |
*** f4m8_ is now known as f4m8 | 04:51 | |
*** santhosh has joined #openstack | 04:55 | |
*** zns has quit IRC | 04:58 | |
*** dendro-afk is now known as dendrobates | 05:14 | |
*** cloudgroups has joined #openstack | 05:15 | |
*** mahadev has quit IRC | 05:18 | |
*** hagarth has joined #openstack | 05:19 | |
*** zenmatt has quit IRC | 05:25 | |
*** mahadev has joined #openstack | 05:25 | |
*** mahadev has quit IRC | 05:26 | |
*** psaikrishna_ has quit IRC | 05:28 | |
*** mahadev has joined #openstack | 05:29 | |
*** cloudgroups has left #openstack | 05:33 | |
*** cloudgroups has joined #openstack | 05:35 | |
*** mahadev has quit IRC | 05:35 | |
*** mahadev has joined #openstack | 05:38 | |
*** arun_ has joined #openstack | 05:39 | |
*** arun_ has joined #openstack | 05:39 | |
*** mahadev has quit IRC | 05:39 | |
*** cloudgroups has quit IRC | 05:42 | |
*** dendrobates is now known as dendro-afk | 05:43 | |
*** Ryan_Lane has joined #openstack | 05:46 | |
*** MarkAtwood has joined #openstack | 06:03 | |
*** jdurgin has joined #openstack | 06:06 | |
*** gregp76 has joined #openstack | 06:12 | |
*** rchavik has quit IRC | 06:12 | |
*** zaitcev has quit IRC | 06:13 | |
*** jamesurquhart has joined #openstack | 06:37 | |
*** toluene has joined #openstack | 06:39 | |
toluene | hi! I 'm a new user to openstack, I followed the installation instructions in openstack-wiki. However, I have problem with the "/bin/-manage db sync", the error is "Access denied for user 'root'@'localhost' (using passwoed: YES)", can someone helps me ? | 06:42 |
*** Ryan_Lane has quit IRC | 06:46 | |
*** lionel has quit IRC | 06:50 | |
*** lionel has joined #openstack | 06:51 | |
soren | toluene: Which instructions did you follow? | 06:51 |
toluene | soren, http://wiki.openstack.org/InstallFromSource | 06:51 |
soren | toluene: Can I see your nova.conf? | 06:52 |
soren | toluene: Put it on a pastebin. | 06:52 |
soren | toluene: Like http://paste.openstack.org/ | 06:52 |
toluene | soren http://pastebin.com/BiZVAtcy | 06:54 |
*** rcc has joined #openstack | 06:56 | |
soren | Oh, right, the instructions are wrong. | 06:56 |
soren | Err.. No, they're not. I'm an idiot. | 06:56 |
*** MarkAtwood has left #openstack | 06:58 | |
toluene | soren, any idea ? it seems like auth problem with the mysql . | 06:59 |
*** rchavik has joined #openstack | 06:59 | |
*** rchavik has joined #openstack | 06:59 | |
soren | toluene: Yeah, I know what the problem is. | 07:00 |
soren | toluene: You're connecting from localhost, but the password is set for root@%. | 07:01 |
soren | toluene: Why the instructions suggest that you use the mysql root user for nova is absolutely beyond me. | 07:01 |
toluene | soren, what should I do ? | 07:01 |
soren | toluene: Something like this: | 07:03 |
soren | toluene: Fire up the mysql client, logging in as root. | 07:03 |
soren | toluene: Then do this | 07:04 |
soren | toluene: grant all on nova.* to 'nova'@'%' identified by 'somepassword'; | 07:04 |
soren | toluene: ...and in your nova.conf, you change the sql_connection to: | 07:04 |
soren | toluene: --sql_connection=mysql://nova:somepassword@127.0.0.1/nova | 07:05 |
soren | That should do the trick. If not, I have one more thing that may need to be done, but I'm not sure it's necessary, so let's try this first. | 07:06 |
toluene | soren, thanks a lot ! | 07:06 |
soren | sure | 07:06 |
*** Zangetsue has joined #openstack | 07:07 | |
*** jfluhmann has quit IRC | 07:08 | |
*** zul has quit IRC | 07:11 | |
*** zul has joined #openstack | 07:12 | |
*** jfluhmann has joined #openstack | 07:12 | |
*** jamesurquhart has quit IRC | 07:14 | |
*** citral has joined #openstack | 07:27 | |
*** jdurgin has quit IRC | 07:34 | |
soren | toluene: Did it work? | 07:34 |
*** miclorb has quit IRC | 07:35 | |
*** fabiand__ has joined #openstack | 07:36 | |
pgregory | morning | 07:38 |
pgregory | could someone tell me if it's possible/reasonable to allow multiple users to create instances for a project? | 07:38 |
*** nacx has joined #openstack | 07:40 | |
HugoKuo__ | pqregory : yes | 07:40 |
HugoKuo__ | 1.add the user to project 2.allocate sysadmin role for both global and project environment | 07:41 |
pgregory | HugoKuo__: would that give the user systemwide access? | 07:42 |
HugoKuo__ | As I know , only sysadmin / projectmanager role has permission to lunch instance | 07:43 |
HugoKuo__ | about your question : no | 07:44 |
HugoKuo__ | user management depends on RBAC | 07:44 |
pgregory | that seems to have worked. | 07:45 |
pgregory | thanks | 07:45 |
HugoKuo__ | sure .... | 07:45 |
HugoKuo__ | see ya | 07:45 |
pgregory | is that 'normal' or is it preferred that only one user can create instances for a project? | 07:45 |
* pgregory is a bit confused about the role of the 'project' in the cloud scheme. | 07:46 | |
HugoKuo__ | sysadmin is a little different from project manager | 07:47 |
pgregory | I'm just wondering if it's better to have 1 project that multiple users can interact with, or give each user a project(s)? | 07:48 |
HugoKuo__ | in my environment , per project / per user... | 07:48 |
HugoKuo__ | I recommand each user a project | 07:49 |
pgregory | I see | 07:49 |
HugoKuo__ | btw a user can add to multi project | 07:49 |
pgregory | so a project is an organisational level 'below' user. | 07:49 |
pgregory | I was thinking of the project as being at a higher level. | 07:50 |
HugoKuo__ | the reason is...... all project sysadmins are free to run or "terminate" instance | 07:50 |
HugoKuo__ | it's quite danger | 07:50 |
pgregory | for example, in my usage case we would have 'productions' and that would relate to a 'project' but it makes more sense I think to just have users, and create a default project for each user, and then create other projects for users who need more fine grained control. | 07:51 |
HugoKuo__ | Actually , I made mistake before :\ , terminate my leader's intance | 07:51 |
HugoKuo__ | yes , you r right...I agree with you | 07:51 |
pgregory | thanks, that clears things up a bit. | 07:52 |
HugoKuo__ | RBAC cost me lots of time ....before :) | 07:52 |
toluene | hi! nova guys, I have problem running './bin/nova-manage project zipfile nova_proj nova' , the certificate db has not been created | 07:59 |
HugoKuo__ | nova-manage user list | 08:00 |
HugoKuo__ | nova-manage project list | 08:00 |
*** dirkx_ has joined #openstack | 08:01 | |
HugoKuo__ | if those exist , check nova-manage.log | 08:01 |
HugoKuo__ | or check database table | 08:01 |
toluene | HugoKuo__, user and project do exist | 08:03 |
HugoKuo__ | any error in nova-manage.log ? | 08:04 |
toluene | HugoKuo__, I think is the location problem for CA/ | 08:04 |
toluene | HugoKuo__, I can't find nova-manage.log in my system | 08:04 |
HugoKuo__ | no idea, bcz I installed nova from PPA ... | 08:05 |
toluene | HugoKuo__, :) | 08:05 |
HugoKuo__ | maybe you can check your nova.conf | 08:05 |
toluene | HugoKuo__, ok | 08:05 |
HugoKuo__ | the log path is written in nova.conf | 08:05 |
*** dirkx_ has quit IRC | 08:07 | |
*** miclorb_ has joined #openstack | 08:09 | |
pgregory | is there any way of assigning a more meaningful description to a published image (Dashboard) or is it always going to show the manifest name? | 08:10 |
soren | pgregory: If you use glance, yes. | 08:12 |
pgregory | oh, is that going to be a pain, or is it easy to switch? | 08:12 |
soren | pgregory: I've not really used glance much myself, I'm afraid. | 08:13 |
pgregory | things are all working quite nicely now, I'm worried about trying one thing too many and finding the house of cards all falls down. | 08:13 |
soren | pgregory: It should be a nice transition, though. | 08:13 |
pgregory | will it work with the Dashboard? | 08:13 |
soren | pgregory: The instructions I gave you yesterday was to get you going. To get something working, let you play with it and demonstrate that Nova actually works. There are a number of things I'd do differently for a production environment. | 08:14 |
soren | pgregory: Yes, that dashboard will work fine. | 08:14 |
soren | s/that/the/ | 08:14 |
pgregory | soren: can I trust in the manual for this? | 08:14 |
soren | For what, exactly? | 08:14 |
soren | Hmm... | 08:15 |
pgregory | setting up glance | 08:15 |
soren | Whatever your response, I probably don't know the answer. I've not personally proofread the manual. | 08:15 |
HugoKuo__ | pgregory : it's easy | 08:16 |
pgregory | HugoKuo__: should I just follow the manual, or is there a better reference. | 08:16 |
HugoKuo__ | just add three flags.... | 08:16 |
HugoKuo__ | yes | 08:16 |
HugoKuo__ | follow it | 08:16 |
* pgregory feel confident enough now that he can rebuild this if it all goes pear shaped. | 08:16 | |
pgregory | one quick question before I begin, should it be ppa:glance-core/trunk (manual) or ppa:glance-core/release? | 08:17 |
HugoKuo__ | soren : does it possible to bind public ip for instance at bootup ? | 08:18 |
HugoKuo__ | release I think | 08:18 |
soren | HugoKuo__: Sure. | 08:19 |
HugoKuo__ | soren : automatically ? | 08:19 |
HugoKuo__ | soren : is there any docs :) ? thanks | 08:19 |
*** jeffjapan has quit IRC | 08:20 | |
soren | HugoKuo__: Well, instead of using private IP's in your network, just use public ones? | 08:23 |
*** daveiw has joined #openstack | 08:23 | |
*** rcc has quit IRC | 08:25 | |
HugoKuo__ | soren : ok.....that's the way in my current openstack nova. but it still traffic from nova-network flat_interface to public interface... | 08:26 |
*** kashyap has quit IRC | 08:27 | |
toluene | soren, hi I face a new problem in "/bin/manage project zipfile nova_proj nova" | 08:27 |
HugoKuo__ | and I did a stupid thing this morning , I set nova-network dnsmasq listen same network with our corporate network , and corporate network crashed XD | 08:27 |
*** naehring has joined #openstack | 08:28 | |
*** zykes- has quit IRC | 08:29 | |
HugoKuo__ | pgregory : https://answers.launchpad.net/nova/+question/153430 for refference | 08:32 |
winston-d | HugoKuo__ : then you were the hero saving whole company. XD | 08:32 |
HugoKuo__ | winston-d : I'm the killer :\ | 08:33 |
HugoKuo__ | Our MIS want to kill me this morning :\ | 08:33 |
winston-d | HugoKuo__ : kill company network, save all from having to work . XD | 08:34 |
soren | toluene: What's the problem? | 08:34 |
*** zykes- has joined #openstack | 08:35 | |
*** zykes- has quit IRC | 08:35 | |
toluene | soren, I pastebin for you, please wait | 08:35 |
RichiH | creiht: http://paste.debian.net/116062/ | 08:35 |
*** zykes- has joined #openstack | 08:36 | |
RichiH | creiht: those are the tcp streams of the two connections | 08:36 |
RichiH | exported as text :p | 08:36 |
pgregory | HugoKuo__: is it possible to get glance to use my existing objectstore, or do I need to re-publish my images? | 08:37 |
HugoKuo__ | pgregory : you need to convert the image from objectstore | 08:37 |
toluene | soren, here is the paste http://pastebin.com/hsaZfetp | 08:38 |
pgregory | HugoKuo__: so does glance not work with objectstore? does it have to use file based storage? | 08:38 |
HugoKuo__ | let me clear it, objectstore is nova-objectstore ? | 08:39 |
HugoKuo__ | or swift ? | 08:39 |
pgregory | yes | 08:39 |
pgregory | no, just nova-objectstore | 08:39 |
pgregory | it's what I've been using up to now, and seems to be ok for my requirements. | 08:39 |
HugoKuo__ | https://answers.launchpad.net/nova/+question/153430 take a look | 08:40 |
HugoKuo__ | you will know , I'm poor in engilsh.............. :\ | 08:41 |
pgregory | HugoKuo__: I'm not really following that, it seems to suggest that you need to use the file based store. | 08:44 |
HugoKuo__ | but nova-objectstore is file based store too | 08:45 |
HugoKuo__ | or you can choose swift for store image | 08:45 |
pgregory | hmm, I'm confused then. | 08:45 |
HugoKuo__ | just a name nova-objectstore | 08:46 |
pgregory | do I leave the glance.conf with 'file' as the default_store? if so, how do I get my existing images (in mybucket on objectstore) into glance, so that they show up in dashboard? | 08:46 |
HugoKuo__ | a path to store object , not means objectstorage ... | 08:47 |
HugoKuo__ | yes | 08:47 |
HugoKuo__ | sure ....the easiest way | 08:47 |
HugoKuo__ | convert !!! | 08:47 |
pgregory | so just get the .tar.gz again, and register with glance, or upload with uec-publish-tarball? (I'm a bit confused) | 08:48 |
HugoKuo__ | nova-manage image convert /path/to/images | 08:48 |
HugoKuo__ | no | 08:48 |
HugoKuo__ | your existing image is in /var/lib/nova/images | 08:48 |
HugoKuo__ | choose one for testing | 08:49 |
pgregory | so I can just do "sudo nova-manage image convert /var/lib/nova/images"? | 08:50 |
pgregory | 'sok, I sorted it, I just used uec-publish-tarball, and it seemed to work fine. | 09:07 |
soren | toluene: Is nova-api running? | 09:09 |
toluene | soren, I have no idea, my apt is messup , I 'm now reinstalling the ubuntu, what a bad day. I have recorded the tips you had gave me. | 09:12 |
soren | Err.. | 09:12 |
soren | I thought you were installing from soruce? | 09:12 |
soren | source, even . | 09:12 |
soren | Not from packages? | 09:12 |
toluene | soren, Yes, I doing thing using source. I'm a gentoo guy | 09:14 |
soren | Now you've lost me. Are you doing this on Gentoo or Ubuntu? | 09:14 |
toluene | soren, My team use ubuntu so I use ubuntu, at home I use gentoo. | 09:15 |
toluene | soren, I'm using ubuntu to install openstack | 09:16 |
soren | So why install from source? | 09:16 |
toluene | soren, I think using source is the best way to use up-to-date openstack | 09:18 |
soren | Why? | 09:18 |
toluene | soren, people committing codes to bzr daily, isn't it ? | 09:19 |
*** miclorb_ has quit IRC | 09:24 | |
soren | toluene: And we build packages correspondingly. | 09:24 |
soren | Admittedly, the last couple of days, we've been lagging a little bit (due to long build queues on Launchpad), but generally, less than half an hour from commit to trunk, there are UBuntu packages available. | 09:25 |
*** adjohn has quit IRC | 09:29 | |
*** gondoi has joined #openstack | 09:31 | |
*** kashyap has joined #openstack | 09:33 | |
alekibango | _nil: iam using duckduckgo... :) you are author? | 09:51 |
*** nijaba_afk has joined #openstack | 10:05 | |
*** nwww has quit IRC | 10:05 | |
*** widodh has quit IRC | 10:05 | |
*** widodh has joined #openstack | 10:05 | |
*** nwww has joined #openstack | 10:05 | |
*** cbeck1 has joined #openstack | 10:05 | |
*** nijaba has quit IRC | 10:05 | |
*** odyi has quit IRC | 10:05 | |
*** jamiec has quit IRC | 10:05 | |
*** cbeck has quit IRC | 10:05 | |
*** jamiec has joined #openstack | 10:06 | |
*** Bat_ has joined #openstack | 10:08 | |
Bat_ | Hi all | 10:08 |
Bat_ | Is there anyone here who could help me a little with OpenStack Swift? | 10:08 |
Bat_ | Really, really important. | 10:09 |
btorch | what's the issue ? i can try | 10:11 |
pgregory | anyone know how I get a description into a glance image metadata that will show up in Dashboard in place of the location? | 10:31 |
pgregory | I've looked into the Dashboard code, and it checks for the existence of 'description' and uses it if it's there, or uses 'location' if not. | 10:31 |
pgregory | so I jumped through hoops with the glance cli tool (why on earth does it wipe any properties not specified in the update command?) to get a description property in there, and it doesn't make any difference, still shows the location. | 10:32 |
soren | pgregory: It wipes the stuff not set in the update command, because that's the only useful way you can offer a way to delete properties. | 10:35 |
*** dirkx_ has joined #openstack | 10:35 | |
pgregory | soren: perhaps a separate command then, 'update' doesn't imply delete to me, perhaps that's just me. | 10:36 |
pgregory | anyway, I've now tried 'description', 'image_description' none of them show up, any ideas? | 10:36 |
soren | pgregory: Think of it as "update my set of properties to this". | 10:37 |
pgregory | soren: I see | 10:37 |
pgregory | fair enough. | 10:37 |
soren | pgregory: I could probably stare at the code and work it out, but I need food to fuel my brain. | 10:37 |
pgregory | I've looked at the code, Django is a bit of a maze, but still can't work it out. | 10:38 |
* soren goes to apply food | 10:41 | |
*** keds has joined #openstack | 10:42 | |
*** ovidwu has quit IRC | 10:43 | |
*** ovidwu has joined #openstack | 10:43 | |
Bat_ | Ok, so I was trying to use swfit with S3 interface | 10:46 |
*** Bat_ has quit IRC | 10:46 | |
*** Bat__ has joined #openstack | 10:47 | |
Bat__ | Ok, sorry | 10:47 |
Bat__ | So my problem is | 10:47 |
Bat__ | I'm using switf + S3 middleware written in python in order to get interface compatible to S3 for my application | 10:47 |
Bat__ | Unfortunatelly, when I try to send files via s3curl I'm getting Invalid URI message | 10:48 |
Bat__ | The problem only occurs if I try to upload a file, lets say /a/b/c.txt to the bucket, in which tere are no subdirecotires "a" and "b" | 10:49 |
Bat__ | This is a big problem for me, since S3 was accepting those requests without a problem | 10:49 |
Bat__ | My sys admin tried to upload the same file without s3 middleware and he told me that the request was processed OK. | 10:50 |
Bat__ | So it lookl like the python tier is somehow broken | 10:50 |
*** noisy_man has joined #openstack | 10:51 | |
HugoKuo__ | http://hugokuo-hugo.blogspot.com/2011/05/thinking-about-possibility-of-these-map.html I'm out of idea , need your help . | 10:54 |
*** toluene has quit IRC | 10:58 | |
*** Jordandev has joined #openstack | 10:59 | |
*** Jordandev has joined #openstack | 11:00 | |
*** adjohn has joined #openstack | 11:02 | |
Bat__ | Anyone? | 11:03 |
*** zns has joined #openstack | 11:03 | |
*** jmckind has joined #openstack | 11:03 | |
colinnich_ | Bat__: I've not used the S3 middleware - if you wait a few hours until the guys in the US start work you are more likely to get an answer. | 11:06 |
alekibango | Bat__: ius guys will be here in 2 hours | 11:12 |
alekibango | s/ius/us/ | 11:12 |
sandywalsh | Bat__, or post your question here: https://answers.launchpad.net/openstack/+addquestion (with code snippets here: http://paste.openstack.org/ ) and someone will pick it up | 11:13 |
*** dirkx_ has quit IRC | 11:17 | |
Bat__ | Ok | 11:23 |
Bat__ | I guess it could help | 11:23 |
*** Jordandev has quit IRC | 11:23 | |
*** dirkx_ has joined #openstack | 11:23 | |
*** mancdaz has quit IRC | 11:31 | |
*** markvoelker has joined #openstack | 11:33 | |
RichiH | generally speaking, is swift happier with faster cpus or with more cores? | 11:33 |
RichiH | i assume more cores, but i'd rather ask before buying stuff | 11:33 |
*** RJD22 is now known as RJD22|away | 11:35 | |
fabiand__ | pgregory: may i ask what guide you followed to setup swift and nova? | 11:36 |
*** rsaidan has joined #openstack | 11:40 | |
*** patcoll has joined #openstack | 11:41 | |
*** ctennis has quit IRC | 11:42 | |
*** rsaidan is now known as nerens | 11:44 | |
*** monaDeveloper has joined #openstack | 11:46 | |
*** nerens has quit IRC | 11:47 | |
*** nerens has joined #openstack | 11:47 | |
*** nerens has quit IRC | 11:50 | |
*** nerens has joined #openstack | 11:51 | |
*** ctennis has joined #openstack | 11:57 | |
*** ctennis has joined #openstack | 11:57 | |
heden | Anyone know why my uploaded AMI file from uec-publish-tarball gets registered with Glance but the state of the image is QUEUED? | 12:03 |
*** stewart has quit IRC | 12:08 | |
*** miclorb_ has joined #openstack | 12:11 | |
*** patcoll has quit IRC | 12:12 | |
*** stewart has joined #openstack | 12:20 | |
*** kakella has joined #openstack | 12:21 | |
*** kakella has left #openstack | 12:21 | |
*** heden has quit IRC | 12:29 | |
*** dendro-afk is now known as dendrobates | 12:38 | |
naehring | hi there! is there any auto scaling feature in nova? | 12:39 |
*** miclorb_ has quit IRC | 12:41 | |
*** deshantm_laptop has joined #openstack | 12:46 | |
*** denovo has quit IRC | 12:50 | |
*** foxtrotgulf has quit IRC | 12:51 | |
*** mray has joined #openstack | 12:57 | |
pgregory | fabiand__: I haven't setup swift, the guide I used for nova was... | 12:58 |
*** mray has quit IRC | 12:59 | |
pgregory | http://fnords.wordpress.com/2010/12/02/bleeding-edge-openstack-nova-on-maverick/ | 12:59 |
*** mray has joined #openstack | 13:00 | |
soren | naehring: No. | 13:01 |
*** adjohn has quit IRC | 13:08 | |
*** hadrian has joined #openstack | 13:08 | |
*** alex-meade has joined #openstack | 13:14 | |
naehring | soren, thanks. | 13:14 |
*** nijaba_afk is now known as nijaba | 13:15 | |
*** headoff has joined #openstack | 13:15 | |
*** nijaba has quit IRC | 13:15 | |
*** nijaba has joined #openstack | 13:15 | |
*** headoff has quit IRC | 13:16 | |
nerens | Hey guys, I've got nova running on a single host and want to move nova-api to another host, as far as I can see the only 2 flags in nova.conf I need to change is --ec2_host & --ec2_url on both hosts? | 13:16 |
nerens | And obviously stop the nova-api service and start up on the new host. | 13:18 |
*** jmckind has quit IRC | 13:18 | |
*** lborda has joined #openstack | 13:21 | |
*** ctennis has quit IRC | 13:23 | |
*** bcwaldon has joined #openstack | 13:24 | |
*** hagarth has quit IRC | 13:25 | |
*** patcoll has joined #openstack | 13:27 | |
*** Zangetsue has quit IRC | 13:27 | |
*** Zangetsue has joined #openstack | 13:28 | |
*** santhosh_ has joined #openstack | 13:29 | |
*** santhosh has quit IRC | 13:30 | |
*** santhosh_ is now known as santhosh | 13:30 | |
dabo | naehring: There are auto-scaling products that will work with OpenStack. Cloudscaling is one (http://cloudscaling.com/), and Scalr (https://scalr.net/) was very active at the recent summit about their plans to work with OpenStack. | 13:30 |
*** hagarth has joined #openstack | 13:33 | |
*** Bat__ has quit IRC | 13:34 | |
naehring | dabo, thanks. I´ ll take a closer look | 13:34 |
*** jmeredit has joined #openstack | 13:35 | |
*** hagarth has quit IRC | 13:36 | |
*** mahadev has joined #openstack | 13:40 | |
*** m_3 has joined #openstack | 13:41 | |
*** arun__ has joined #openstack | 13:41 | |
*** arun_ has quit IRC | 13:43 | |
RichiH | are people using 10.04 or 11.04 for actual deployments? | 13:46 |
btorch | we use LTS only | 13:46 |
*** ChameleonSys has quit IRC | 13:48 | |
*** ChameleonSys has joined #openstack | 13:48 | |
*** f4m8 is now known as f4m8_ | 13:48 | |
*** ctennis has joined #openstack | 13:50 | |
*** amccabe has joined #openstack | 13:50 | |
*** zenmatt has joined #openstack | 13:51 | |
*** ianp100 has joined #openstack | 13:52 | |
ianp100 | hi, openstack novice here | 13:53 |
* soren is breaking a personal record for number of e-mails sent in an hour right now. | 13:54 | |
soren | RichiH: Either should be fine. | 13:54 |
soren | RichiH: If not, it's a bug that I want to fix. | 13:54 |
ianp100 | was interested to know if there are best practise details for modifying the rings once you cluster is in production | 13:54 |
*** mgoldmann has joined #openstack | 13:54 | |
*** mgoldmann has joined #openstack | 13:54 | |
*** dirkx_ has quit IRC | 13:54 | |
*** jmckind has joined #openstack | 13:55 | |
*** guynaor has joined #openstack | 13:57 | |
colinnich_ | ianp100: I'm not sure there is any documentation for that - not that I can find anyway. Did you have any specific questions? | 13:58 |
*** naehring has quit IRC | 13:59 | |
*** guynaor has left #openstack | 13:59 | |
nhm | Rich: I'm using LTS here. | 14:00 |
*** jtimberman has joined #openstack | 14:00 | |
btorch | RichiH: btw I was referring to swift not nova | 14:01 |
nhm | Any of you guys looked at Jeff Darcy's cloudfs project? | 14:01 |
*** rattlesnake has joined #openstack | 14:02 | |
ianp100 | colinnich: I would like to know if the process for modifying the rings is the same as it is for creating (by modifying I mean adding more storage etc). I'd also be interested in undestanding the potential performance hit (and subsequent best practise to minimize) of rebalancing when in production (assuming I'd have to do that if Ive added more drives/nodes), thanks | 14:02 |
*** Zangetsue has quit IRC | 14:04 | |
rattlesnake | A question about Nova: Does the /var/lib/nova/images folder need to be shared across all the compute nodes with NFS or similar? i'm using nova-objectstore with local storage running in a different node, with nova-api and nova-scheduler | 14:05 |
*** nanderso has joined #openstack | 14:06 | |
*** santhosh has quit IRC | 14:07 | |
jaypipes | vishy: see rattlesnake's question above... I'm not sure? | 14:07 |
jaypipes | nhm: jdarcy showed us a few months ago and I had a quick look back then, but not since. any major updates? | 14:08 |
nhm | jaypipes: not sure, I'm trying to get a feel for it. | 14:09 |
jdarcy | Still in progress, unfortunately. | 14:09 |
nhm | jaypipes: ie what the relationship with gluster will look like, and what gluster's relationship with openstack will look like. | 14:09 |
jaypipes | nhm: hmm, yeah, I can't help you with that, sorry :) jdarcy does hang out here, though, so he might reply../. | 14:09 |
jaypipes | hehe, speak of the devil :) | 14:09 |
jdarcy | More like an imp, actually. | 14:09 |
jaypipes | hehe | 14:09 |
*** jkoelker has joined #openstack | 14:10 | |
jdarcy | I honestly haven't quite figured out what the Gluster/OpenStack connection is, other than a sort of general "let's all work together in the cloud" thing. | 14:11 |
*** viirya has quit IRC | 14:11 | |
*** viirya has joined #openstack | 14:11 | |
*** viirya has quit IRC | 14:11 | |
*** MarkAtwood has joined #openstack | 14:12 | |
jdarcy | CloudFS will mostly just follow GlusterFS in that regard, as a consequence of being based on it. I guess there might be some ways that CloudFS's specific features could be integrated into OpenStack (e.g. coordinating users/tenants) but I haven't thought about it much. | 14:12 |
*** naehring has joined #openstack | 14:14 | |
*** viirya has joined #openstack | 14:16 | |
*** imsplitbit has joined #openstack | 14:19 | |
*** mahadev has quit IRC | 14:21 | |
*** knorthcut has joined #openstack | 14:22 | |
fabiand__ | pgregory: thanks. are you from europe, btw? :) | 14:25 |
*** diegoparrilla has joined #openstack | 14:27 | |
*** fabiand has joined #openstack | 14:27 | |
*** fabiand__ has quit IRC | 14:27 | |
diegoparrilla | Regarding GlusterFS I think there are a clear use case: KVM live migration. I asked the Gluster guys in Santa Clara about a reference architecture of GlusterFS for KVM LiveMigration with Nova and they told me they would work on it soon | 14:29 |
*** kbringard has joined #openstack | 14:31 | |
*** naehring has quit IRC | 14:31 | |
diegoparrilla | ... but the most popular sentence in Santa Calra was: 'We will work on it soon' ;-) | 14:32 |
pgregory | fabiand: yes, UK. | 14:33 |
creiht | RichiH: re: cpus, I would say more cores is better than faster cpus | 14:34 |
jdarcy | diegoparrilla: I think we have folks here (Red Hat) working on something like that too, not specific to GlusterFS though. The shrinking-delta approach should basically work with any shared FS. | 14:34 |
*** hazmat has quit IRC | 14:35 | |
*** hazmat has joined #openstack | 14:35 | |
*** thingee has joined #openstack | 14:35 | |
creiht | man lots of swift questions today | 14:35 |
*** nanderso has left #openstack | 14:36 | |
*** mancdaz has joined #openstack | 14:36 | |
*** mancdaz1203 has joined #openstack | 14:36 | |
creiht | ianp100: re: rings: yes you use the same process. You add new devices to the .builder files from the previous ring, rebalance then push the rings out to all the nodes | 14:36 |
*** mancdaz1203 has quit IRC | 14:36 | |
creiht | if it is a lot of storage, then we recommend starting with a smaller weight, let replication settle, bump up the weight, repeat until the new devices are at full weight | 14:37 |
diegoparrilla | jdarcy: we use Nexenta, but obviously cannot scale how Gluster does. | 14:38 |
ianp100 | creiht: thanks, any performance considerations during the rebalancing to users? | 14:41 |
*** zenmatt has quit IRC | 14:41 | |
*** keds has quit IRC | 14:41 | |
creiht | ianp100: for the most parts, the end users may only notice some extra latency until everything is rebalanced | 14:42 |
creiht | ianp100: you might want to look at: https://answers.launchpad.net/swift/+question/152024 | 14:44 |
ianp100 | is there any documentation on billing and metering for swift? | 14:44 |
creiht | ianp100: http://swift.openstack.org/overview_stats.html | 14:44 |
creiht | that will dump .csv files with all the stats which you can dump into your favorite db or analytics engine | 14:45 |
creiht | and use that for billing, etc. | 14:45 |
ianp100 | creiht: thanks, one final question :) there's plenty of info on resilence around the proxy and storage nodes, is there any best practise for the auth node? | 14:47 |
*** koolhead17 has joined #openstack | 14:48 | |
creiht | RichiH: I looked at the paste, but I'm not sure I can tell much from that | 14:48 |
creiht | ianp100: which auth are you using? swauth? | 14:49 |
ianp100 | creiht: assume for the moment yes | 14:49 |
creiht | you can run swauth on any number of proxies, and load balance between those | 14:51 |
creiht | you can run it standalone if you need to, or on top of the normal proxies | 14:51 |
*** rattlesnake has quit IRC | 14:52 | |
* soren calls it a day | 14:52 | |
kbringard | hey guys, quick ? | 14:52 |
kbringard | http://nova.openstack.org/devref/network.html#component-diagram | 14:52 |
kbringard | I'm looking at this, and it's confusing me a bit | 14:52 |
RichiH | creiht: you can see that it's sending practically the same in both bases and swift3 just dies and does not respond | 14:52 |
kbringard | this seems to be a logical diagram... but physically, if I'm using VLAN mode, the Routing Node and Addressing Node are both the network controller, correct? | 14:53 |
RichiH | creiht: if you suspect the colon to be the cause, can you just edit line 403 to accept _ as seperator in the access key? | 14:53 |
creiht | RichiH: well I saw that part, but it doesn't tell me if the client library used is dieing, or something is wrong with the server | 14:53 |
creiht | _ could be in a username for swift | 14:54 |
*** zns has quit IRC | 14:54 | |
*** troytoman-away is now known as troytoman | 14:56 | |
*** mahadev has joined #openstack | 14:57 | |
*** zenmatt has joined #openstack | 14:57 | |
*** mahadev has quit IRC | 15:03 | |
koolhead17 | hi all | 15:05 |
*** larzy has joined #openstack | 15:07 | |
*** Jeng-Hsueh has joined #openstack | 15:07 | |
ccooke | so every now and then, openstack suddenly starts to chew through CPU | 15:10 |
ccooke | nova-network, nova-compute and nova-scheduler all end up taking about 25% of a CPU, with beam.smp (the rabbitmq main process) taking another 20% | 15:11 |
*** markwash has joined #openstack | 15:11 | |
ccooke | I see no issues in the logs | 15:11 |
ccooke | and load drops to normal once I've restarted *all* the openstack daemons | 15:11 |
ccooke | (Load drops quite a bit when I restart all the three I mentioned, but it stays above normal until nova-api and nova-objectstore are restarted, too) | 15:12 |
*** dirkx_ has joined #openstack | 15:19 | |
*** arun__ is now known as arun_ | 15:19 | |
*** troytoman is now known as troytoman-away | 15:19 | |
*** troytoman-away is now known as troytoman | 15:19 | |
*** blamar has joined #openstack | 15:21 | |
*** orangeacme has left #openstack | 15:22 | |
*** Jeng-Hsueh has quit IRC | 15:24 | |
*** rchavik has quit IRC | 15:26 | |
*** maplebed has joined #openstack | 15:31 | |
*** daveiw has quit IRC | 15:34 | |
*** noisy_man has quit IRC | 15:35 | |
*** dirkx_ has quit IRC | 15:36 | |
*** nerens has quit IRC | 15:40 | |
*** ianp100 has quit IRC | 15:44 | |
*** Ryan_Lane has joined #openstack | 15:44 | |
*** koolhead17 has quit IRC | 15:46 | |
*** koolhead17 has joined #openstack | 15:46 | |
*** ianp100 has joined #openstack | 15:46 | |
*** ianp100 has quit IRC | 15:48 | |
*** Ryan_Lane has quit IRC | 15:52 | |
*** jpuchala has quit IRC | 15:53 | |
*** throughnothing has joined #openstack | 15:54 | |
*** med_out is now known as medberry | 15:57 | |
*** obino has quit IRC | 16:00 | |
*** Shentonfreude has joined #openstack | 16:01 | |
*** dodeeric has joined #openstack | 16:08 | |
*** jpuchala has joined #openstack | 16:10 | |
*** blamar has quit IRC | 16:12 | |
*** mgoldmann has quit IRC | 16:13 | |
*** mgoldmann has joined #openstack | 16:13 | |
*** BK_man has joined #openstack | 16:18 | |
*** dendrobates is now known as dendro-afk | 16:19 | |
*** mgoldmann has quit IRC | 16:19 | |
RichiH | load balancing swift... we are pondering DNS RR or pushing everytihng through a central IP and abusing inbound nat. obviously, dns rr scales linearly while inbound nat can't scale indefinitely unless you do "geo" ip in your backbone or similar. (or source-based routing, which sucks) that is, unless it's possible to seperate singnalling and data transfer (i.e. you have one or two dispatchers which tell the clients which host to connct to to ... | 16:20 |
RichiH | ... actually get the data) | 16:20 |
RichiH | tl;dr: how are people load balancing swift? | 16:20 |
kpepple | RichiH: interesting post to read from the Rackspace guys http://pandemicsyn.posterous.com/load-balancing-i-can-haz-lots-of-ssl-throughp | 16:21 |
*** johnpur has joined #openstack | 16:22 | |
*** ChanServ sets mode: +v johnpur | 16:22 | |
RichiH | kpepple: i am going through that as we speak, but it's always good to get more input | 16:22 |
RichiH | and yes, we are in full control of our own backbone | 16:22 |
kpepple | RichiH: it's an interesting way to do it without resorting to big iron LB ... but ospf can be a bear to configure/troubleshoot. depends on how much traffic, i suppose | 16:23 |
RichiH | and what i don't get is how ospf equal cost routing helps at all | 16:23 |
RichiH | as this will distribute packets, not flows/streams | 16:24 |
jaypipes | diegoparrilla: hey there :) safe trip back to Spain? | 16:24 |
*** citral1 has joined #openstack | 16:24 | |
*** alex__ has joined #openstack | 16:24 | |
RichiH | and even if distribute handshake information in the background, you are still left with the tcp stack not having the correct sequence numbers | 16:24 |
*** alex__ is now known as romain42 | 16:24 | |
*** romain42 is now known as katkee | 16:25 | |
*** obino has joined #openstack | 16:25 | |
katkee | hello | 16:25 |
diegoparrilla | hey Jay! Yes, a long trip but safe at home! | 16:25 |
RichiH | so unless you simply patch your kernel, disable NIC offloading and accept _everything_, this can not work | 16:25 |
RichiH | unless i am not getting how it's done | 16:25 |
RichiH | kpepple: and yes, ospf sucks donkeys through straws | 16:25 |
diegoparrilla | jaypipes: hey Jay! Yes, a long trip but safe at home! | 16:25 |
katkee | is it possible to run nova with xen without using XCP? using libvirt for example? | 16:26 |
RichiH | "here is an interface/subnet. have fun trusting anything and everything you will ever see there and decide your routing based on this with (almost) no manual control over route selection" | 16:26 |
jaypipes | diegoparrilla: good to hear :) | 16:27 |
*** citral has quit IRC | 16:27 | |
RichiH | ospf is fine to make your backbone aware of other boxes' loopback addresses so they can talk ibgp, but that's about it | 16:27 |
kpepple | RichiH: is he saying that you use DNS RR and then maybe per destination load balancing ? | 16:28 |
RichiH | pandemicsyn: see above, please | 16:28 |
*** zenmatt has quit IRC | 16:28 | |
*** citral1 has quit IRC | 16:29 | |
diegoparrilla | jaypipes: it was a great summit. I learnt a lot | 16:31 |
*** photron has joined #openstack | 16:31 | |
vishy | rattlesnake: if you want to use local image service on multiple hosts, then yes you have to share that folder via nfs, but the recommended solution for multiple hosts is to use glance | 16:31 |
*** lvaughn has quit IRC | 16:32 | |
katkee | is it possible to run nova with xen without using XCP? using libvirt for example? | 16:33 |
RichiH | katkee: we saw it the first time, but no one seems to have an answer | 16:33 |
vishy | katkee: yes | 16:33 |
jaypipes | diegoparrilla: me too! :) | 16:33 |
vishy | katkee: although i don't know of anyone actually using that method | 16:33 |
*** kashyap has quit IRC | 16:34 | |
jaypipes | diegoparrilla: good discussions (arguments?!), good coding, overall an excellent time. | 16:34 |
vishy | --libvirt_type=xen | 16:34 |
*** dragondm has joined #openstack | 16:34 | |
katkee | vishy: thanks very much. is KVM the most widely used hypervisor with OpenStack? | 16:35 |
vishy | katkee: KVM through libvirt or xen through xen driver | 16:36 |
kbringard | I'd actually be very interested people's thought's on KVM vs Xen | 16:36 |
diegoparrilla | vishy: mmm... I guess rattlesnake is somebody from my team that just left the channel... we are testing NFS and KVM live migration and we were wondering what is a better solution: NFS sharing or Glance copy-on-deploy approach | 16:36 |
vishy | seems to be roughly split, kvm is way easier to install | 16:36 |
jaypipes | vishy: hey, if you've got a sec, could you re-check https://code.launchpad.net/~jaypipes/glance/bug773388/+merge/59858? thx in advance! | 16:37 |
vishy | diegoparrilla: I would suggest glance for multi-machine | 16:37 |
jaypipes | diegoparrilla: glance with a swift backend. | 16:37 |
kbringard | glance == good | 16:37 |
*** kashyap has joined #openstack | 16:37 | |
diegoparrilla | vishy: yes, I agree | 16:38 |
* jaypipes <3 kbringard. | 16:38 | |
kbringard | jaypipes: lol, I only speak the truth | 16:38 |
vishy | jaypipes: approved | 16:38 |
jaypipes | vishy: cheers mate :) | 16:38 |
*** keds has joined #openstack | 16:38 | |
kbringard | vishy: quick ? about HA on the network controller | 16:39 |
kbringard | well, a couple | 16:39 |
jaypipes | vishy: now, if you have 15 minutes... ;) https://code.launchpad.net/~jaypipes/glance/bug713154/+merge/59110 | 16:39 |
katkee | we are studying OpenStack in order to migrate our current xen installation on it but it seems that OpenStack and XCP are overlapping themselves... We would prefer OpenStack to XCP but use Xen as a hypervisor. | 16:39 |
kbringard | do we think the NaaS stuff is going to address that? | 16:39 |
kbringard | secondly, if I heartbeat IPs and replicate IPtables and the dnsmasq files (leases, pids, etc... probably via NFS) | 16:39 |
diegoparrilla | jaypipes: we were considering NFS shared approach for images enabling dedup and thin provisioing in the NexentaStor system | 16:40 |
kbringard | any reason you can think of that wouldn't work for failover? | 16:40 |
*** zenmatt has joined #openstack | 16:40 | |
diegoparrilla | jaypipes: But glance is a better approach. And more resilient, probably | 16:40 |
jaypipes | diegoparrilla: I'm not too familiar with the NexentaStor system, but you should still be able to use NFS with Glance's filesystem backend. Just set filesystem_store_datadir to point to the NFS mount... | 16:41 |
kbringard | that's what I do, on a netapp | 16:41 |
jaypipes | diegoparrilla: and ensure write permissions and all that yada :) | 16:41 |
jaypipes | kbringard: w00t. | 16:42 |
*** aliguori has quit IRC | 16:42 | |
diegoparrilla | jaypipes: write permissions and root attributes, my favourite part in NFS ;-) | 16:42 |
jaypipes | diegoparrilla: indeed ;) | 16:42 |
*** katkee has quit IRC | 16:42 | |
RichiH | pandemicsyn: unless you are using something that, once it has the routes in its FIB, does flow-based routing | 16:42 |
RichiH | that would work | 16:42 |
diegoparrilla | jaypipes & vishy: Thank you again, see u later | 16:43 |
RichiH | well, not even flow-based routing, but flow-based equal cost distribution | 16:43 |
*** diegoparrilla has left #openstack | 16:43 | |
*** keds has quit IRC | 16:44 | |
*** dodeeric has quit IRC | 16:44 | |
vishy | XCP is going to use openstack afaik | 16:45 |
vishy | kbringard: there are instructions by ntt on network node ha | 16:46 |
kbringard | oh, nice | 16:46 |
kbringard | that is mucho good | 16:46 |
vishy | they managed to get a 4 second failover | 16:46 |
kbringard | I'll go find 'em | 16:46 |
*** Ryan_Lane has joined #openstack | 16:46 | |
vishy | kbringard: this was the message that had the instructions: https://lists.launchpad.net/openstack/msg01994.html | 16:50 |
kbringard | hehe, i was just reading that, thanks | 16:51 |
vishy | it was an attachment, i can forward to you if you didn't get it | 16:51 |
kbringard | I see it attached on the page there | 16:51 |
kbringard | lemme see if I can download it | 16:51 |
*** tblamer has joined #openstack | 16:51 | |
kbringard | I got it, thanks again | 16:53 |
*** keds has joined #openstack | 16:54 | |
*** dendro-afk is now known as dendrobates | 16:55 | |
*** MotoMilind has quit IRC | 16:55 | |
*** ryker has quit IRC | 16:58 | |
*** jdurgin has joined #openstack | 16:58 | |
kbringard | vishy: so, if I'm reading this right, nova-network brings up all the correct IP tables rules, dnsmasq leases, etc automagically when it comes up on the new node... so we don't have to manage that? | 16:58 |
kbringard | if so, that makes this infinitely easier... I was thinking I was going to need to sync all that stuff manually | 16:58 |
exlt | creiht: might want to think about using 'apt-get build-dep swift' in that section of installing build dependencies - if something is missing at build, then the build dependencies need adjusting | 17:01 |
vishy | correct | 17:01 |
kbringard | vishy: that is ephing awesome | 17:01 |
kbringard | good work sir :-) | 17:01 |
*** ryker has joined #openstack | 17:04 | |
*** Dweezahr has joined #openstack | 17:08 | |
creiht | exlt: good point | 17:09 |
*** dragondm has quit IRC | 17:11 | |
*** jamesurquhart has joined #openstack | 17:12 | |
*** kbringard has quit IRC | 17:12 | |
*** h0cin has joined #openstack | 17:13 | |
*** h0cin has joined #openstack | 17:13 | |
*** kbringard has joined #openstack | 17:13 | |
*** omidhdl has joined #openstack | 17:15 | |
*** troytoman is now known as troytoman-away | 17:18 | |
*** Jordandev has joined #openstack | 17:18 | |
*** yamahata_lt has quit IRC | 17:18 | |
*** krish has joined #openstack | 17:26 | |
*** gondoi has quit IRC | 17:27 | |
*** cbeck1 is now known as cbeck | 17:27 | |
BK_man | hi all. Is euca-upload-bundle is supposed to work with Nova / Glance backend with filesystem at the moment? Mine got an error http://paste.openstack.org/show/1288/ | 17:31 |
*** denovo has joined #openstack | 17:31 | |
vishy | Bk_man: yes. Is nova-objectstore running? | 17:32 |
BK_man | vishy: nope. Should be? | 17:32 |
*** deshantm_laptop_ has joined #openstack | 17:32 | |
vishy | yup, can't use euca-upload-bundle without it | 17:33 |
*** dendrobates is now known as dendro-afk | 17:33 | |
BK_man | vishy: does that mean that objectstore is not obsoleted by Glance? | 17:33 |
vishy | euca-upload uses s3 | 17:34 |
vishy | so you need an s3 server of some sort | 17:34 |
*** deshantm_laptop has quit IRC | 17:34 | |
*** deshantm_laptop_ is now known as deshantm_laptop | 17:35 | |
*** omidhdl has quit IRC | 17:36 | |
*** maplebed has quit IRC | 17:36 | |
*** dragondm has joined #openstack | 17:36 | |
*** omidhdl has joined #openstack | 17:36 | |
*** denovo has quit IRC | 17:37 | |
*** medberry is now known as med_out | 17:37 | |
*** maplebed has joined #openstack | 17:38 | |
*** nacx has quit IRC | 17:38 | |
*** zaitcev has joined #openstack | 17:39 | |
*** maplebed has quit IRC | 17:40 | |
*** omidhdl has quit IRC | 17:40 | |
*** omidhdl has joined #openstack | 17:40 | |
*** maplebed has joined #openstack | 17:41 | |
*** dragondm has quit IRC | 17:43 | |
BK_man | vishy: started objectstore, but the error isn't gone | 17:45 |
BK_man | vishy: telnet test passed on IP:3333 | 17:45 |
vishy | are you sure your novarc has the right ip? | 17:46 |
creiht | RichiH: so I'm looking again at s3curl.pl, trying to understand the code | 17:46 |
creiht | I'm starting to think that it isn't the : messing it up | 17:47 |
BK_man | vishy: yep, it's fresh. let me begin from scratch again | 17:49 |
creiht | on my test setup, the s3-curl script still doesn't sign the request the same as the middleware is expecting | 17:49 |
creiht | using the boto libs works though | 17:49 |
creiht | anyone know of another commandline s3 tool that I could test? | 17:50 |
*** Jordandev has quit IRC | 17:51 | |
*** aliguori has joined #openstack | 17:53 | |
* BK_man finally found error in /etc/nova/nova.conf | 17:56 | |
*** MarkAtwood has quit IRC | 17:56 | |
BK_man | why nova-manage project zipfile command does not honor --s3_host flag in /etc/nova/nova.conf? | 17:56 |
*** photron has quit IRC | 17:57 | |
*** mgoldmann has joined #openstack | 17:57 | |
imsplitbit | BK_man I have found that you should always explicitly pass the --flagfile=/etc/nova/nova.conf argument to nova-manager | 17:58 |
imsplitbit | nova-manage | 17:58 |
vishy | not, sure | 17:58 |
imsplitbit | woohoo off to a 2 hour meeting... I <3 meetings | 17:58 |
*** pguth66 has joined #openstack | 18:02 | |
*** maplebed has joined #openstack | 18:03 | |
creiht | RichiH: so it looks like the s3curl script is for some reason adding localhost to the string of stuff to hash | 18:06 |
*** Jordandev has joined #openstack | 18:06 | |
* creiht digs deeper | 18:06 | |
*** dendro-afk is now known as dendrobates | 18:08 | |
*** fabiand__ has joined #openstack | 18:08 | |
*** deshantm_laptop has quit IRC | 18:10 | |
*** MotoMilind has joined #openstack | 18:16 | |
pandemicsyn | RichiH: sorry fixing to run off to a meeting but, aggr/pub switch to quagga/ospf on a linux box running your lb software. Its 1 flow to 1 quagga instance , not per packet. | 18:16 |
*** dragondm has joined #openstack | 18:18 | |
*** MarkAtwood has joined #openstack | 18:18 | |
*** mgoldmann has quit IRC | 18:18 | |
*** MotoMilind has quit IRC | 18:20 | |
*** Jordandev has quit IRC | 18:21 | |
RichiH | creiht: oh? | 18:23 |
creiht | RichiH: getting closer | 18:24 |
RichiH | pandemicsyn: quagga does not route, it only computes routes and injects those into the kernel | 18:24 |
RichiH | (also, even though we are using quagga on linux exclusively, bird is better) | 18:24 |
*** MotoMilind has joined #openstack | 18:25 | |
pandemicsyn | RichiH: yea, bird is cool | 18:25 |
RichiH | pandemicsyn: i will need to try this once i have more than scavenged hardware, i.e. next week | 18:25 |
pandemicsyn | pretty sure theres actually bird example for what your looking for up on the bird site | 18:25 |
*** daveiw has joined #openstack | 18:25 | |
pandemicsyn | its a pretty common way to deal with a ton of traffic, think most of the big swift deployments are doing this | 18:26 |
creiht | RichiH: --debug on s3curl has turned out to be useful :) | 18:26 |
RichiH | pandemicsyn: you should update your blog to reflect that you need something that does flow-based routing | 18:26 |
*** troytoman-away is now known as troytoman | 18:26 | |
RichiH | you had three of us staring at the monitor going "wtf?" :p | 18:26 |
creiht | RichiH: first step so far is to add your host to the @endpoints in the script | 18:27 |
creiht | so in my test setup, I added 'localhost' to the list | 18:27 |
creiht | now I am getting a 404, just trying to figure out why that is | 18:27 |
RichiH | creiht: i clobbered it with print Dumper foo, but i didn't look at what goes into the base64'ed part, yet | 18:28 |
*** clauden has joined #openstack | 18:28 | |
creiht | if you add your storage host to the @endpoints, it fixes the signing issue | 18:28 |
*** omidhdl has left #openstack | 18:29 | |
*** MarkAtwood has quit IRC | 18:30 | |
RichiH | creiht: i am at home atm. let me finish my spaghetti real quick and i'll grab the source | 18:30 |
creiht | hehe | 18:30 |
creiht | no rush | 18:30 |
*** photron_ has joined #openstack | 18:33 | |
*** daveiw has left #openstack | 18:34 | |
*** troytoman is now known as troytoman-away | 18:34 | |
*** jbryce has joined #openstack | 18:34 | |
BK_man | jaypipes: around? | 18:36 |
*** ewanmellor has joined #openstack | 18:36 | |
ewanmellor | o/ | 18:36 |
jaypipes | BK_man: yep! | 18:36 |
jaypipes | ewanmellor: heya | 18:37 |
BK_man | jaypipes: look here: http://paste.openstack.org/show/1290/ | 18:37 |
_vinay | Hi vishy | 18:37 |
jaypipes | bcwaldon: http://etherpad.openstack.org/GlanceApi | 18:37 |
ewanmellor | Does anyone know what it means when the swift unit test test_invalidate_hash fails? I presume that I have got an incorrect dependency, because it doesn't do it on another machine, but I can't figure out which one it would be. | 18:37 |
jaypipes | BK_man: heh. | 18:38 |
_vinay | can you point me to any stress tests written specifically for nova? | 18:38 |
BK_man | jaypipes: is that expected? | 18:38 |
_vinay | I am looking at smoketests but they are not stress | 18:38 |
jaypipes | BK_man: update to the latest rev :) bug was *just* sent into trunk for that (bug 773388) | 18:38 |
BK_man | jaypipes: I uploaded my images to Nova using euca2ools | 18:38 |
uvirtbot | Launchpad bug 773388 in glance "glance index does not show queued or saving images" [High,Fix committed] https://launchpad.net/bugs/773388 | 18:38 |
BK_man | jaypipes: ah, ok | 18:38 |
*** MarkAtwood has joined #openstack | 18:38 | |
jaypipes | BK_man: sorry for the confusion | 18:38 |
_vinay | We are evaluating openstack/nova for IaaS layer.. and we want to run some load and perf tests to get some statistics | 18:39 |
*** dendrobates is now known as dendro-afk | 18:39 | |
*** smoser has quit IRC | 18:40 | |
*** smoser has joined #openstack | 18:41 | |
*** keds has quit IRC | 18:42 | |
kpepple | RichiH: can you document what you come up re: ospf/bird/quagga ? i may be embarking on a similar quest soon ... | 18:46 |
*** keds has joined #openstack | 18:51 | |
*** jkoelker has quit IRC | 18:55 | |
*** fabiand__ has left #openstack | 18:55 | |
Dweezahr | Is hardware acceleration essential for running openstack? | 18:55 |
*** jkoelker has joined #openstack | 18:55 | |
Dweezahr | I am running a cloud controller from one vm and a node from another vm | 18:55 |
Dweezahr | And I get: libvirtError: internal error no supported architecture for os type 'hvm' | 18:56 |
*** mahadev has joined #openstack | 18:57 | |
_vinay | Hi kpepple, | 18:59 |
*** KnuckleSangwich has joined #openstack | 18:59 | |
_vinay | do you have any idea about the stresstests question I asked above? | 18:59 |
*** anotherjesse has joined #openstack | 19:01 | |
vishy | _vinay: there aren't any that I am aware of | 19:02 |
*** pguth66 has quit IRC | 19:02 | |
vishy | aside from maybe http://github.com/vishvananda/novatools.git | 19:03 |
_vinay | okay.. so if we want stats on performance, best way is to write ourselves ? | 19:03 |
_vinay | ok will check novatools | 19:03 |
vishy | but that is for one host at a time | 19:03 |
kpepple | _vinay: you'll probably have to write something yourself with fog or something like that | 19:03 |
vishy | qa at nasa did some stress tests using grinder | 19:04 |
vishy | but I don't think they released them publicly | 19:04 |
*** obino has quit IRC | 19:05 | |
*** obino has joined #openstack | 19:05 | |
*** namaqua has joined #openstack | 19:05 | |
nhm | vinay: what kind of performance? We might do some HPL runs at some point. | 19:05 |
creiht | RichiH: success! | 19:06 |
_vinay | nhm we are evaluating openstack at this point | 19:07 |
_vinay | so we need perf stats like how many instances can openstack support... | 19:07 |
_vinay | at what point does it choke.. does it fail cleanly? etc. etc. | 19:08 |
creiht | RichiH: so in the slcurl.pl script, you have to add your host to @endpoints | 19:09 |
creiht | I had forgotten that the requests are mapped from amazon style requests to swift requests | 19:09 |
_vinay | nhm what is HPL ?? | 19:09 |
creiht | so running the following on my test system gave me the bucket list: | 19:09 |
creiht | ./s3curl.pl --debug --id test:tester --key testing get -- -s -v http://localhost:8080/ | 19:09 |
*** mahadev has quit IRC | 19:10 | |
Dweezahr | _vinay, High performance linux, it is Linpack for clusters | 19:10 |
*** mahadev has joined #openstack | 19:10 | |
nhm | I've heard some rumors that virtualized IB is getting better. Not nearly as good as native IB, but is beating native ethernet. | 19:11 |
*** keds has quit IRC | 19:12 | |
_vinay | ok Dweezahr, nhm | 19:14 |
throughnothing | Hey, I'm following this tutorial http://wiki.openstack.org/InstallFromSource and everything seems to be working finally, but when I try to create an instance using 'euca-run-instances ami-tiny --kernel aki-lucid --ramdisk ari-lucid -k nova_key' I get an ImageNotFound error. I have the images/ directory with the images in them in the project checkout root as per the tutorial, but it seems to not be finding them. Anyone have the | 19:15 |
_vinay | can you share your approach on how to go about it | 19:15 |
_vinay | do you have any test written / test plan ? | 19:15 |
*** brd_from_italy has joined #openstack | 19:15 | |
throughnothing | also, all the tests pass for me using the virtualenv | 19:16 |
*** markvoelker has quit IRC | 19:16 | |
*** ctennis has quit IRC | 19:16 | |
Dweezahr | throughnothing, why from source? | 19:16 |
kpepple | throughnothing: can you euca-describe-images and those images ? | 19:16 |
throughnothing | kpepple: I don't have any images running, but when i do euca-describe-images i dont get any errors | 19:17 |
*** dendro-afk is now known as dendrobates | 19:17 | |
*** ctennis has joined #openstack | 19:17 | |
*** ctennis has joined #openstack | 19:17 | |
throughnothing | Dweezahr: because I want to develop | 19:17 |
kpepple | throughnothing: euca-describe-IMAGES just lists your available images. i think you are thinking of euca-describe-INSTANCES | 19:17 |
throughnothing | kpepple: ah let me check | 19:18 |
throughnothing | ok, so you are correct :) euca-describe-images shows me nothing (no errors though) | 19:18 |
throughnothing | this is what I figured the problem was, just not sure how to solve...its not seeing my available images | 19:18 |
kpepple | throughnothing: there's an error on http://wiki.openstack.org/InstallFromSource ... it has you pull down the image tar ball but then you never upload it into nova | 19:20 |
throughnothing | ahh | 19:22 |
throughnothing | i was wondering :) | 19:22 |
throughnothing | is there somewhere on the wiki that describes this process | 19:22 |
* kpepple thumbs through wiki looking for when those instructions got deleted | 19:22 | |
throughnothing | hehe | 19:24 |
kpepple | throughnothing: go to http://wiki.openstack.org/RunningNova and jump down to the "Register an image" section. use the ttylinux test image for starters ... i think this still works ... once you uec-publish, make sure it is up there with euca-describe-images | 19:24 |
throughnothing | kpepple: thanks a bunch! Will let you know when its working :) | 19:24 |
kpepple | throughnothing: np | 19:24 |
Dweezahr | libvirterror: no supported architecture for os type 'hvm' | 19:26 |
Dweezahr | But I am using qemu I think | 19:26 |
Dweezahr | sudo qemu-nbd -d /dev/nbd15 from (pid=848) execute /usr/lib/pymodules/python2.6/nova/utils.py:150 | 19:27 |
throughnothing | kpepple: awesome, worked | 19:30 |
kpepple | throughnothing: cool | 19:30 |
throughnothing | kpepple: are you a dev? | 19:30 |
kpepple | throughnothing: i've contributed some code to nova | 19:31 |
throughnothing | cool | 19:33 |
photron_ | Dweezahr: is you host running as 64 or 32bit linux? | 19:35 |
Dweezahr | photron_, both hosts are 64 bit | 19:35 |
RichiH | creiht: ugh sorry, friends came over before i could join in | 19:35 |
creiht | RichiH: np | 19:35 |
RichiH | creiht: can you paste a patch? | 19:35 |
RichiH | or just submit to cpan | 19:35 |
creiht | well | 19:35 |
Dweezahr | photron_, But, euca-describe-images show i386 | 19:35 |
*** galthaus has joined #openstack | 19:36 | |
* RichiH still has an account, somewhere if need be | 19:36 | |
Dweezahr | photron_, IMAGEami-05c95c5edub-bucket/maverick-server-uec-amd64.img.manifest.xmlavailablepublici386machineaki-396602b2 | 19:36 |
creiht | Dweezahr: http://paste.openstack.org/show/1294/ | 19:36 |
creiht | It is a customization of that script | 19:36 |
creiht | not sure there is anything really to submit | 19:36 |
creiht | just anyone who wants to use it with a swift cluster will have to modify that part of the script | 19:37 |
photron_ | Dweezahr: i dont mean the images or virtual machines, i really mean the physical host running nova | 19:37 |
*** lcfseth has joined #openstack | 19:37 | |
Dweezahr | photron_, cloud controller is a vm in vmware on windows 7, cloud node is a vm on vmware on osx on a mac pro | 19:37 |
*** MarkAtwood has quit IRC | 19:38 | |
Dweezahr | creith, what script is this, nova-install? | 19:38 |
photron_ | Dweezahr: and those vmware vms are running 64bit linux? | 19:38 |
*** jmeredit has quit IRC | 19:39 | |
creiht | Dweezahr: that script is s3curl.pl, a command line utility for interacting with s3 | 19:39 |
creiht | trying to help someone get it to work with swift | 19:39 |
Dweezahr | photron_, cloud controller: x86_64 node: x86_64 | 19:39 |
RichiH | creiht: that's fine; i'll look at adding code to expose the switch to the outside | 19:39 |
creiht | RichiH: cool, and thanks | 19:40 |
creiht | there still seems to be another bug with it that I am tracking down | 19:40 |
* RichiH suspect that he will rely on this library _a lot_ | 19:40 | |
photron_ | Dweezahr: okay then the "libvirterror: no supported architecture for os type 'hvm'" error is not caused by the problem i had a while ago | 19:40 |
creiht | I was able to use it to list buckets and objects, create buckets, and upload an object | 19:40 |
RichiH | s/supect/knows/ | 19:40 |
creiht | though if you try to upload an object with a path in the object name, it fails | 19:41 |
creiht | RichiH: why not just use st? :) | 19:41 |
RichiH | creiht: i am a perl guy and i will end up writing a lot of custom code to stuff data into swift | 19:41 |
Dweezahr | creiht, locate s3curl.pl gives nothing on node and cloud controller, is it only included in the source version? | 19:41 |
creiht | RichiH: ah gotcha | 19:42 |
RichiH | i don't really care about s3ls etc other than as a codebase to steal code from | 19:42 |
*** Ryan_Lane has quit IRC | 19:42 | |
vishy | people still use perl? | 19:42 |
*** lcfseth has left #openstack | 19:42 | |
vishy | :p | 19:42 |
creiht | Dweezahr: nah, this is an external script that people would normally use with s3 | 19:42 |
RichiH | but i need the library :) | 19:42 |
*** Ryan_Lane has joined #openstack | 19:42 | |
creiht | gotcha | 19:43 |
*** lcfseth has joined #openstack | 19:43 | |
RichiH | vishy: the ones who know that syntax and formatting should not be mixed do ;) | 19:43 |
* RichiH hides from the pythons | 19:43 | |
*** nid0 has joined #openstack | 19:43 | |
*** troytoman-away is now known as troytoman | 19:44 | |
vishy | I thought all of you graduated to ruby! :) | 19:44 |
RichiH | heh | 19:44 |
RichiH | let's put it this way | 19:44 |
RichiH | i wrote a custom http proxy and a sms sending daemon, in perl, in 2004 | 19:45 |
RichiH | or 2005 | 19:45 |
RichiH | it still works today | 19:45 |
RichiH | try that with ruby, or python, and we'll talk | 19:45 |
*** joearnold has joined #openstack | 19:46 | |
* creiht stays out of religious battles :) | 19:46 | |
deepy | RichiH: I wrote an application using Twisted, it still works | 19:46 |
RichiH | creiht: i am just teasing :p | 19:46 |
alekibango | RichiH: i wrote perl script 1200 lines long in 199x and now i really do now know what it does | 19:47 |
alekibango | lol | 19:47 |
RichiH | alekibango: but it still works | 19:48 |
alekibango | RichiH: not sure about that... as it was working with netscape browser internals somehow | 19:49 |
RichiH | alekibango: k; external applications can change | 19:50 |
alekibango | you know, good old times of netscape 3, when browser was lean and quick and even javascript worked :) | 19:50 |
alekibango | oh, i remember mosaic and gopher and stuff... darn i am getting old | 19:50 |
kpepple | RichiH: <snark>it only works because Perl hasn't released a new version since 1992</snark> | 19:50 |
Dweezahr | c++ is even older | 19:51 |
alekibango | dont worry, skynet is now online and it will end those wars | 19:51 |
Dweezahr | skynet? stuxnet | 19:52 |
*** joearnold has quit IRC | 19:53 | |
*** jamesurquhart has quit IRC | 19:54 | |
*** joearnold has joined #openstack | 19:54 | |
*** MarkAtwood has joined #openstack | 19:54 | |
*** thingee has joined #openstack | 19:55 | |
*** krish has quit IRC | 19:55 | |
*** lcfseth has left #openstack | 19:55 | |
RichiH | kpepple: debian sid is just migrating to 5.12 | 20:00 |
RichiH | though admittedly, 5.8 -> 5.10 took ages | 20:00 |
uvirtbot | New bug: #778089 in swift "swift3 middleware doesn't allow / in object names" [Undecided,New] https://launchpad.net/bugs/778089 | 20:06 |
*** jamesurquhart has joined #openstack | 20:07 | |
*** mdomsch has joined #openstack | 20:09 | |
*** jonkelly has joined #openstack | 20:10 | |
*** clauden has quit IRC | 20:11 | |
*** bcwaldon has quit IRC | 20:17 | |
*** ctennis has quit IRC | 20:18 | |
Dweezahr | Allright, I have no hardware acceleration on both the cloud controller and the node, how can I get past this error code: libvirterror: internal error no supported architecture for os type 'hvm'? Qemu is running fine I when kvm module is disabled. /dev/kvm is absent as well. This used to work with the Bexar release, I got it running back then without this error | 20:18 |
*** rnirmal has joined #openstack | 20:18 | |
Dweezahr | Both vms are running in vmware, and it seems to be impossible to get hardware acceleration working in vmware | 20:19 |
kpepple | Dweezahr: use uml by changing your libvirt type to "--libvirt_type=kvm" in nova.conf | 20:19 |
kpepple | Dweezahr: err, i mean "--libvirt_type=uml" | 20:19 |
*** bcwaldon has joined #openstack | 20:19 | |
*** gondoi has joined #openstack | 20:21 | |
vishy | --libvirt_type=qemu? | 20:21 |
vishy | uml requires special images | 20:21 |
Dweezahr | okay, qemu first | 20:21 |
kpepple | vishy: ooops forgot about that | 20:21 |
*** mahadev_ has joined #openstack | 20:22 | |
* vishy devs on vms all the time | 20:23 | |
* vishy and uses qemu | 20:23 | |
Dweezahr | 2 computers and 8 vms here | 20:23 |
Dweezahr | Wow, it is running now | 20:24 |
*** photron_ has quit IRC | 20:24 | |
vishy | \o/ | 20:25 |
*** mahadev has quit IRC | 20:25 | |
*** pguth66 has joined #openstack | 20:26 | |
*** smoser has quit IRC | 20:29 | |
*** ewanmellor has quit IRC | 20:29 | |
*** alex-meade has quit IRC | 20:31 | |
*** smoser has joined #openstack | 20:31 | |
*** anotherjesse has quit IRC | 20:32 | |
*** ctennis has joined #openstack | 20:32 | |
*** ctennis has joined #openstack | 20:32 | |
*** dendrobates is now known as dendro-afk | 20:33 | |
*** gregp76 has joined #openstack | 20:34 | |
*** troytoman is now known as troytoman-away | 20:34 | |
*** vernhart has quit IRC | 20:34 | |
*** imsplitbit has quit IRC | 20:35 | |
*** odyi has joined #openstack | 20:37 | |
*** odyi has joined #openstack | 20:37 | |
*** bcwaldon has quit IRC | 20:39 | |
vishy | mtaylor: ping | 20:46 |
*** Shentonfreude has quit IRC | 20:57 | |
*** Shentonfreude has joined #openstack | 21:00 | |
*** Shentonfreude has quit IRC | 21:03 | |
*** gondoi has quit IRC | 21:03 | |
*** gondoi has joined #openstack | 21:03 | |
*** mahadev_ has quit IRC | 21:05 | |
*** mahadev has joined #openstack | 21:06 | |
*** MarkAtwood has quit IRC | 21:07 | |
*** smoser has quit IRC | 21:08 | |
*** keds has joined #openstack | 21:08 | |
*** jamesurquhart has quit IRC | 21:09 | |
*** anotherjesse has joined #openstack | 21:10 | |
*** anotherjesse has quit IRC | 21:10 | |
*** smoser has joined #openstack | 21:12 | |
*** keds has quit IRC | 21:13 | |
*** Jordandev has joined #openstack | 21:14 | |
*** joearnold has quit IRC | 21:16 | |
openstackjenkins | Project swift build #255: SUCCESS in 32 sec: http://jenkins.openstack.org/job/swift/255/ | 21:17 |
openstackjenkins | Tarmac: Fixes issue of swift3 middleware not working correctly with '/' in object names | 21:17 |
*** gregp76 has quit IRC | 21:18 | |
*** vernhart has joined #openstack | 21:23 | |
*** woleium has joined #openstack | 21:23 | |
*** h0cin has quit IRC | 21:27 | |
*** icarus901 has quit IRC | 21:29 | |
*** mdomsch_ has joined #openstack | 21:29 | |
*** mdomsch has quit IRC | 21:30 | |
*** knorthcut has quit IRC | 21:35 | |
*** miclorb_ has joined #openstack | 21:36 | |
*** smoser has quit IRC | 21:39 | |
*** smoser has joined #openstack | 21:41 | |
*** gondoi has quit IRC | 21:42 | |
*** icarus901 has joined #openstack | 21:43 | |
*** namaqua has quit IRC | 21:45 | |
*** londo_ is now known as londo | 21:46 | |
*** grapex has joined #openstack | 21:48 | |
markwash | any pointers on how clean.sqlite is generated? | 21:52 |
*** lborda has quit IRC | 21:54 | |
*** lborda has joined #openstack | 21:54 | |
*** rnirmal has quit IRC | 21:56 | |
*** aliguori has quit IRC | 21:57 | |
markwash | nm I got it: nova.db.migration.db_sync() | 21:57 |
*** icarus901 has quit IRC | 22:00 | |
*** brd_from_italy has quit IRC | 22:03 | |
*** nerens has joined #openstack | 22:06 | |
*** Jordandev has quit IRC | 22:08 | |
*** mdomsch_ has quit IRC | 22:11 | |
*** kbringard has quit IRC | 22:13 | |
*** mray has quit IRC | 22:14 | |
*** kashyap has quit IRC | 22:14 | |
*** kashyap has joined #openstack | 22:15 | |
*** smoser has quit IRC | 22:15 | |
vishy | markwash: look in __init__.py in the tests dir | 22:15 |
*** jmckind has quit IRC | 22:16 | |
*** amccabe has quit IRC | 22:16 | |
*** dendro-afk is now known as dendrobates | 22:17 | |
*** smoser has joined #openstack | 22:25 | |
*** maplebed has quit IRC | 22:34 | |
Ryan_Lane | I'm not sure if I should have created this as a nova blueprint, or if I should have proposed it for one of the networking services: https://blueprints.launchpad.net/nova/+spec/public-and-private-dns | 22:36 |
vishy | Ryan_Lane: +1 to the spec for sure | 22:37 |
*** galthaus has left #openstack | 22:37 | |
Ryan_Lane | grapex was asking me about it, and I definitely want to move this code out of my extension | 22:37 |
vishy | as to where it goes, it can probably go into nova for now | 22:37 |
*** dendrobates is now known as dendro-afk | 22:37 | |
Ryan_Lane | I'm not totally sure I can actually get this in for diablo :) | 22:37 |
Ryan_Lane | I have what? 5 months? | 22:38 |
Ryan_Lane | (I'm writing up a spec right now, btw.) | 22:38 |
*** aliguori has joined #openstack | 22:40 | |
*** jmckind has joined #openstack | 22:41 | |
*** jkoelker has quit IRC | 22:45 | |
grapex | So I'm wanting to use Rackspace's internal DNS services which are very simple. I could do this from RedDwarf dbcontainer create call, but I need to store the association between an instance and DNS entry somewhere (heck, it might even be "extra" data for now). | 22:45 |
Ryan_Lane | so nova wouldn't update DNS? | 22:46 |
*** jonkelly has quit IRC | 22:46 | |
Ryan_Lane | just store data in the database about what is in DNS? | 22:46 |
grapex | It would be easier for us if it did, but we could work around it in the short term by putting code to do that in the dbcontainer portion of our API that piggybacks off the Nova compute stuff. We'd still need the association to persist somewhere though. | 22:48 |
*** joearnold has joined #openstack | 22:48 | |
Ryan_Lane | how is DNS itself not the association? | 22:48 |
Ryan_Lane | you know the private IP address of the instance | 22:48 |
Ryan_Lane | by doing a reverse lookup you have the entry | 22:48 |
Ryan_Lane | or were you saying it would be easier if nova updated DNS? | 22:49 |
*** icarus901 has joined #openstack | 22:51 | |
Ryan_Lane | are instance names unique? or only the instance id? | 22:51 |
Ryan_Lane | I guess I could test that really quick :) | 22:51 |
grapex | True, but if the instance got blown away the IP address would be gone. Then again if the DNS association was in the DB it'd probably disappear too- I think you're right, it may not be needed. If Nova updated DNS it would still be nice. Maybe even that isn't needed. | 22:51 |
Ryan_Lane | I'm proposing nova manage the DNS entries | 22:51 |
*** patcoll has quit IRC | 22:52 | |
Ryan_Lane | that way if you create an instance, a DNS entry is added, and if you delete an instance, the DNS entry is also removed | 22:52 |
grapex | I think thats a great idea. | 22:53 |
Ryan_Lane | the way I'm going to propose won't require schema changes | 22:53 |
*** tblamer has quit IRC | 22:53 | |
Ryan_Lane | which is a plus, because I don't understand how that part of the code works :D | 22:54 |
* Ryan_Lane never uses ORMs | 22:54 | |
*** joearnold has quit IRC | 22:55 | |
grapex | Yeah, now that I think about it putting another association in the db between an instance and DNS doesn't help anything at all as long as you have a way of knowing when the DNS stuff needs to change. | 22:56 |
Ryan_Lane | well, the good thing is that you can always use the IP address to find the entry | 22:57 |
Ryan_Lane | so deleting the entries is easy | 22:57 |
Ryan_Lane | no database needed | 22:57 |
Ryan_Lane | public DNS is more difficult, because an IP address can be in multiple entries ;) | 22:57 |
Ryan_Lane | but nothing requires association in a database | 22:58 |
Ryan_Lane | hmm. public DNS should likely allow multiple DNS zones too | 22:59 |
*** blamar has joined #openstack | 23:03 | |
*** miclorb_ has quit IRC | 23:06 | |
*** thingee has quit IRC | 23:10 | |
*** NelsonN has joined #openstack | 23:11 | |
*** miclorb_ has joined #openstack | 23:19 | |
*** aliguori has quit IRC | 23:24 | |
*** timrc has quit IRC | 23:25 | |
*** timrc has joined #openstack | 23:26 | |
*** mgoldmann has joined #openstack | 23:26 | |
*** nerens has quit IRC | 23:27 | |
grapex | So I was wrong- internally we can't depend on looking up the entry via IP for whatever reason so I think our application needs the association. We could get by just stashing it as meta data somewhere though, even if we had to just decorate the DNS driver based system you're proposing to do it. | 23:34 |
grapex | Anyway, thanks for starting the blueprint. I look forward to talking more about it tomorrow. | 23:34 |
*** adjohn has joined #openstack | 23:35 | |
*** nerens has joined #openstack | 23:35 | |
*** grapex has left #openstack | 23:39 | |
*** dendro-afk is now known as dendrobates | 23:39 | |
*** dragondm has quit IRC | 23:42 | |
*** rchavik has joined #openstack | 23:51 | |
*** zaitcev has quit IRC | 23:58 | |
*** adjohn has quit IRC | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!