| nelson | bwalk7125: easiest if you paste the URL: http://paste.openstack.org/show/531/ | 00:07 |
|---|---|---|
| *** pvo is now known as pvo_away | 00:08 | |
| *** greenisus has joined #openstack | 00:08 | |
| nelson | bwalk7125: what are you doing with the value of c? and what are you doing with response_headers? printing them out? | 00:10 |
| bwalk7125 | nelson: c is probably not necessary, but yes, I print out the response headers | 00:11 |
| nelson | I'm not an expert on pycurl (at all) , but I'm wondering where you're telling it what to write as part of the PUT ? | 00:11 |
| *** adiantum has quit IRC | 00:13 | |
| bwalk7125 | let me check the response with curl again | 00:13 |
| nelson | pycurl isn't very pythonic. | 00:13 |
| nelson | or are you using it because you have working curl calls and are trying to put them into Python? | 00:13 |
| *** Tushar has joined #openstack | 00:14 | |
| anticw | how do i update a merge proposal | 00:14 |
| anticw | or will a push do that automagically? | 00:15 |
| anticw | i can never tell with lp as it seems to be quite laggy from place to place | 00:15 |
| *** Tushar has quit IRC | 00:15 | |
| *** Tushar has joined #openstack | 00:15 | |
| *** kashyapc has quit IRC | 00:16 | |
| bwalk7125 | just using pycurl as it was recommended by a friend | 00:17 |
| bwalk7125 | but you're right about how it's not very pythonic | 00:17 |
| *** adiantum has joined #openstack | 00:17 | |
| *** kashyapc has joined #openstack | 00:18 | |
| Tushar | I have push a branch on LP, but I see Author as root. How do change from root to my name? | 00:18 |
| *** dubsquared has quit IRC | 00:27 | |
| *** adiantum has quit IRC | 00:27 | |
| Ryan_Lane | Tushar: "bzr whoami", I believe | 00:29 |
| Tushar | Yes, I tried bzr whoami and Do I need to update Authors file also? | 00:30 |
| *** jimbaker has quit IRC | 00:31 | |
| Ryan_Lane | I'm pretty sure I needed to | 00:31 |
| *** adiantum has joined #openstack | 00:31 | |
| Ryan_Lane | I think you need to add yourself to the contributor's page on the wiki too: http://wiki.openstack.org/Contributors | 00:31 |
| Tushar | Ryan: I did that couple of hours back.. | 00:32 |
| *** kashyapc has quit IRC | 00:34 | |
| *** troytoman has quit IRC | 00:36 | |
| termie | soren: db version should tell you the version | 00:40 |
| termie | soren: btw, related to my packaging merge, i am trying to make sure which branch i should be trying to merge to | 00:40 |
| termie | soren: what is the main packaging branch we are using? | 00:40 |
| *** littleidea has quit IRC | 00:43 | |
| termie | soren: db version doesn't tell you you need to sync, but we could add that probabl | 00:44 |
| termie | soren: and db sync is idempotent | 00:44 |
| *** kashyapc has joined #openstack | 00:46 | |
| *** adiantum has quit IRC | 00:47 | |
| eday | termie: any better idea on if you guys will be in town on the 31? | 00:47 |
| termie | eday: yup, i think we leave in the evening | 00:49 |
| termie | eday: but we hsould be there during the day | 00:49 |
| termie | eday: if you want to chill, hang or hack | 00:50 |
| termie | s/there/here/ | 00:50 |
| *** littleidea has joined #openstack | 00:51 | |
| *** adiantum has joined #openstack | 00:52 | |
| *** dragondm has quit IRC | 00:55 | |
| *** pharkmillups has quit IRC | 00:55 | |
| *** Tushar has quit IRC | 00:57 | |
| eday | termie: ok, cool. I'll come down in the morning then | 00:58 |
| *** koji-iida has joined #openstack | 00:58 | |
| *** adiantum has quit IRC | 01:01 | |
| *** adiantum has joined #openstack | 01:02 | |
| vishy | creiht: are you here? | 01:08 |
| *** greenisus has quit IRC | 01:14 | |
| *** Pentheus has joined #openstack | 01:19 | |
| *** saicologic has joined #openstack | 01:19 | |
| *** ccustine has quit IRC | 01:23 | |
| *** dendro-afk is now known as dendrobates | 01:30 | |
| *** gaveen_ has quit IRC | 01:36 | |
| *** adiantum has quit IRC | 01:41 | |
| *** dirakx has joined #openstack | 01:42 | |
| *** saicologic has quit IRC | 01:44 | |
| *** adiantum has joined #openstack | 01:46 | |
| *** reldan has joined #openstack | 01:51 | |
| *** Ryan_Lane has quit IRC | 01:51 | |
| *** sophiap has quit IRC | 01:53 | |
| *** aimon has quit IRC | 01:55 | |
| *** maplebed has quit IRC | 01:55 | |
| *** joearnold has quit IRC | 01:59 | |
| *** jimbaker has joined #openstack | 02:00 | |
| *** greenisus has joined #openstack | 02:02 | |
| *** bwalk7125 has quit IRC | 02:04 | |
| *** ChrisAM1 is now known as ChrisAM | 02:05 | |
| *** hansode has joined #openstack | 02:05 | |
| *** adiantum has quit IRC | 02:10 | |
| *** aimon has joined #openstack | 02:12 | |
| *** littleidea has quit IRC | 02:14 | |
| *** adiantum has joined #openstack | 02:15 | |
| anticw | any swift rackers about right now? | 02:16 |
| *** westmaas has joined #openstack | 02:21 | |
| *** burris has joined #openstack | 02:22 | |
| *** greenisus has quit IRC | 02:23 | |
| *** reldan has quit IRC | 02:26 | |
| *** dendrobates is now known as dendro-afk | 02:27 | |
| *** jdurgin has quit IRC | 02:27 | |
| *** adiantum has quit IRC | 02:31 | |
| *** MarkAt2od has quit IRC | 02:33 | |
| *** MarkAt2od has joined #openstack | 02:38 | |
| *** aimon has quit IRC | 02:40 | |
| notmyname | anticw: what's up? | 02:46 |
| *** littleidea has joined #openstack | 02:47 | |
| anticw | notmyname: x-auth-groups ... that's new in austin | 02:50 |
| anticw | what does that typically look like? | 02:50 |
| anticw | act:usr ? | 02:50 |
| notmyname | good question | 02:51 |
| notmyname | let me check | 02:51 |
| *** zul has joined #openstack | 02:52 | |
| notmyname | anticw: seems to be details on http://swift.openstack.org/development_auth.html | 02:53 |
| *** retr0h has quit IRC | 02:54 | |
| notmyname | it's for ACLs and allows users to grant read/write permissions to other users | 02:54 |
| anticw | yeah, i get that much ... it's newish in austin | 02:54 |
| anticw | and something we tinkered with. both good & bad | 02:55 |
| *** aimon_ has joined #openstack | 02:56 | |
| *** littleidea has quit IRC | 02:56 | |
| notmyname | anticw: how do you mean newish? wasn't everything equally new in austin? that was the first official release | 02:57 |
| anticw | there was code before that | 02:58 |
| anticw | that lacked the groups | 02:58 |
| notmyname | I guess I'm not understanding your question | 03:02 |
| anticw | notmyname: i must be retarded, but im really not parsing the details on that page very well at all | 03:02 |
| anticw | we have a diff auth server | 03:02 |
| anticw | that never did x-auth-groups | 03:02 |
| anticw | i need to contruct that in a way that works correctly | 03:02 |
| anticw | it seems x-auth-groups: user,,token | 03:03 |
| anticw | might work | 03:03 |
| anticw | not sure about the middle option | 03:03 |
| anticw | actually, no, user,, might be all that's needed | 03:04 |
| anticw | notmyname: so if i put foobar into a group option ... what parses/proceses that? | 03:04 |
| *** pvo_away is now known as pvo | 03:05 | |
| *** pvo is now known as pvo_away | 03:05 | |
| *** pvo_away is now known as pvo | 03:06 | |
| notmyname | anticw: you are trying to make your auth system compatible with swift's devauth? | 03:06 |
| anticw | yes | 03:07 |
| notmyname | let me look around to see what I can do to help. but the better person to talk to is gholt since he wrote the code | 03:09 |
| notmyname | in bexar we are including a swauth auth system that is production-ready as opposed to devauth which we recommend for dev use only | 03:10 |
| notmyname | anticw: ah. I see it now | 03:14 |
| *** littleidea has joined #openstack | 03:14 | |
| notmyname | anticw: looks like all you should need to do is return "account:user,account" at the bare minimum | 03:16 |
| notmyname | you could optionally return "account:user,account,another_account" for more ACL support | 03:17 |
| anticw | account is what they passed in X-Auth-User ? | 03:18 |
| anticw | or what swift uses internally? | 03:18 |
| notmyname | neither :-) | 03:18 |
| anticw | k | 03:18 |
| notmyname | X-Auth-User is "account:user" | 03:18 |
| anticw | you mean literally that? | 03:18 |
| anticw | ok | 03:18 |
| notmyname | devauth has the concept of auth accounts, each with one to many users | 03:19 |
| notmyname | the accounts are mapped to the swift accounts (eg AUTH_12423449j34234234) | 03:19 |
| notmyname | but the user auths wth "account:user" | 03:19 |
| notmyname | so if I had a user on your system, I would do something like "curl -i -H "X-Auth-User: someaccount:notmyname" -H "X-Auth-Key: mypassword" https://your.auth.system.com/v1.0 | 03:21 |
| notmyname | make sense? | 03:21 |
| anticw | sorry, im having parsing issues | 03:25 |
| anticw | we had something similar but different, and my brain isn't bridging the dots right this second | 03:26 |
| notmyname | how can I help you connect the dots? | 03:26 |
| anticw | honestly, seeing example byte streams works great right now | 03:27 |
| anticw | i just had someone setup an austin auth server | 03:27 |
| anticw | and got a token from it | 03:27 |
| anticw | so now im going to see what /token/ says for that | 03:27 |
| anticw | the resller prefix stuff i just dont see why it is the way it is | 03:28 |
| anticw | when you could demux in an auth layer | 03:28 |
| anticw | and avoid magic string prefixes | 03:28 |
| anticw | notmyname: ok, i partly see what's needed | 03:30 |
| notmyname | the reseller concept allows you to shard the account namespace across multiple non-interacting auth systems | 03:30 |
| anticw | wrt to the last argument, i see some token there | 03:30 |
| anticw | that's optional right? | 03:30 |
| notmyname | yes | 03:31 |
| anticw | seems it returns the auth acct string | 03:32 |
| anticw | well, cf acct name | 03:32 |
| anticw | ah, only for admin t | 03:33 |
| anticw | notmyname: why is env[...] used to pass state? | 03:35 |
| notmyname | where? in the wsgi middleware? | 03:36 |
| anticw | yeah | 03:36 |
| anticw | is that a wsgi thang? | 03:36 |
| notmyname | 'cause that's the way you do it with middleware. | 03:37 |
| notmyname | the env gets passed down the wsgi stack, so each layer's job is to modify the state by updating the env | 03:37 |
| anticw | why env though? | 03:37 |
| anticw | it's all the same process isn't it? | 03:38 |
| *** ksteward has quit IRC | 03:39 | |
| *** kashyapc has quit IRC | 03:41 | |
| *** kashyapc has joined #openstack | 03:42 | |
| notmyname | yes, but you don't have access to some controlling object. well, actually, I guess you do. it's the env | 03:42 |
| *** DubLo7 has joined #openstack | 03:45 | |
| anticw | yeah, all you have is env | 03:47 |
| anticw | and you need to do a get /token/xxx | 03:47 |
| anticw | because you dont actually know at that stage who/what you are talking to | 03:47 |
| *** DubLo7 has left #openstack | 03:50 | |
| *** kashyapc has quit IRC | 03:54 | |
| *** pvo is now known as pvo_away | 03:59 | |
| *** joearnol_ has joined #openstack | 04:01 | |
| *** MarkAt2od has quit IRC | 04:07 | |
| *** retr0h has joined #openstack | 04:17 | |
| *** MarkAt2od has joined #openstack | 04:17 | |
| *** maplebed has joined #openstack | 04:32 | |
| anticw | notmyname: still up? | 04:32 |
| notmyname | yes, but was actually just about to shut down for the night. what's up? | 04:33 |
| anticw | it can probably waut | 04:35 |
| anticw | wait | 04:35 |
| anticw | wrt to the groups | 04:35 |
| anticw | there is a,b,c | 04:35 |
| anticw | a is typically the acct | 04:35 |
| anticw | sorry, acct:user | 04:35 |
| anticw | b is the acct | 04:35 |
| anticw | and so depending on the container may or may not have access right? | 04:35 |
| notmyname | ya, I think that's the way it's supposed to work. perhaps gholt can enlighten us in the morning | 04:36 |
| anticw | ok | 04:37 |
| notmyname | anticw: all that being said, I didn't press you much about making something compatible with devauth | 04:37 |
| notmyname | I don't think that's a good idea | 04:37 |
| notmyname | devauth will probably go away in cactus | 04:37 |
| anticw | notmyname: what should middleware be compatible with? | 04:37 |
| notmyname | the plan is to completely replace it with swauth | 04:37 |
| anticw | not possible | 04:37 |
| notmyname | swauth and devauth are both in bexar | 04:37 |
| anticw | i know | 04:38 |
| anticw | but we can't use either | 04:38 |
| anticw | uam is external, that's something we can't avoid | 04:38 |
| notmyname | uam? | 04:38 |
| anticw | so if we move user knowledge deeply into swauth | 04:38 |
| anticw | user account management | 04:38 |
| notmyname | ah | 04:38 |
| anticw | if we move that into swift of sorts, we have sync/skew | 04:38 |
| anticw | so we have something that proxies devauth to their middleware | 04:38 |
| notmyname | no, the idea is that you should bring your own auth system. but there is no need to make it api compatible with anything else | 04:39 |
| anticw | what's the interface then? | 04:39 |
| notmyname | devauth is for example purposes. swauth is much better (prod ready), but not a requirement | 04:39 |
| anticw | this isn't devauth | 04:39 |
| anticw | it's just using the same rest interface | 04:39 |
| notmyname | your auth middleware needs to implement an authorize method that's it | 04:40 |
| anticw | this is gonna sound dumb | 04:40 |
| anticw | how does the state get into the auth middleware? | 04:40 |
| notmyname | each piece of middleware is given the environment dictionary (per the wsgi spec) | 04:41 |
| anticw | so middleware then has to be wsgi? | 04:41 |
| anticw | i would just put in a shim that proxies to a rest interface then | 04:41 |
| anticw | it still dont really grok the wsgi stuff .... it's the same process as everything else? | 04:42 |
| notmyname | i think you are getting the 2 pieces confused | 04:42 |
| notmyname | there are 2 parts | 04:42 |
| anticw | likely | 04:42 |
| anticw | the nouns & metaphores are causing buffer overflows in my head | 04:43 |
| notmyname | 1 is the middleware that implements an authorize method. the middleware takes the env. see swift/common/middleware/auth.py for the devauth example | 04:43 |
| anticw | ok ... that im somewhat familar with, but only in isolation ... exactly how the proxies communicate with it im vague on | 04:44 |
| notmyname | the other part is some service that handles account management (creation deletion etc). the middleware's authorize method could call out to this | 04:44 |
| anticw | right, so the first part | 04:44 |
| anticw | does it end up being it's own service? | 04:45 |
| notmyname | no | 04:45 |
| anticw | recall, i know nothing about web-frameworks and wsgi and what not, other than in large doeses it causes brain cancer | 04:45 |
| notmyname | think of it as a swift plugin | 04:45 |
| anticw | ok | 04:45 |
| anticw | plugin means what | 04:45 |
| anticw | in c you have shared libraries | 04:45 |
| notmyname | not like that | 04:45 |
| anticw | in python it can be but typically you import something | 04:46 |
| anticw | or it's something you talk to over a socket/pipe | 04:46 |
| notmyname | middleware is something that goes between the client and the server that examines and/or modifies the request and/or response | 04:46 |
| anticw | sure, i get that much | 04:46 |
| anticw | but i dont understand HOW the data flows | 04:46 |
| notmyname | the first piece of middleware calls the second which calls the third etc until the server is called | 04:47 |
| anticw | calls how | 04:47 |
| anticw | this seems to imply the middleware is python | 04:47 |
| anticw | always | 04:47 |
| notmyname | swift/common/middleware/catch_errors.py is a very simple example | 04:48 |
| notmyname | yes, it is. but it can call out to anything | 04:48 |
| notmyname | via whatever method you want | 04:48 |
| *** kashyapc has joined #openstack | 04:49 | |
| notmyname | catch_errors is designed to be the outer-most middleware to catch any uncaught errors anywhere else | 04:49 |
| *** EdwinGrubbs has joined #openstack | 04:49 | |
| *** EdwinGrubbs has joined #openstack | 04:49 | |
| anticw | notmyname: you mean replace __call__ with something that marshals data over a socket for example? | 04:49 |
| creiht | vishy: what's up? | 04:52 |
| notmyname | anticw: replace the __call__ body | 04:53 |
| anticw | right, so in that case we can shim it to something that talks to whatever ReST interface we like | 04:53 |
| anticw | and make it pretty thin | 04:53 |
| notmyname | anticw: swift/common/middleware/memcache.py is another simple one to look at. it adds a memcache client to the env so it's available to the other middleware or the server itself | 04:54 |
| notmyname | yes, exactly | 04:54 |
| notmyname | anticw: the healthcheck middleware looks for a specific path and short-circuits the response | 04:57 |
| *** littleidea has quit IRC | 04:57 | |
| notmyname | the domain remap middleware takes elements from the host header and puts it in the URL path | 04:57 |
| anticw | im going to get angry turn green and rewrite all this is vax assembly | 04:58 |
| notmyname | so your auth middleware should look at the env variables you care about (headers or REMOTE_USER if using basic auth) and call your identity management service to return true or falsh | 04:58 |
| notmyname | sean connery python? true and falsh | 04:59 |
| notmyname | anticw: is it making any more sense? | 05:02 |
| anticw | yeah, up to a point | 05:02 |
| notmyname | good. I hope that point is not where you were a few hours ago :-) | 05:03 |
| anticw | no, im not doing too badly, i'm trying to figure out how to work around some code someone else did right now :-) | 05:03 |
| anticw | that is part of this maze of issues | 05:03 |
| notmyname | heh. we've all been there | 05:04 |
| notmyname | anticw: the only other important piece I can think of is that the "pipeline" config option is what defines the middleware and what order it's called in | 05:04 |
| *** burris has quit IRC | 05:05 | |
| notmyname | for the proxy, notice that catch_errors is first and proxy_server is last. the config sections at the bottom of the config file define how those pieces of middleware are called and what options they get | 05:06 |
| *** burris has joined #openstack | 05:08 | |
| *** littleidea has joined #openstack | 05:09 | |
| *** hadrian has quit IRC | 05:14 | |
| *** opengeard has quit IRC | 05:17 | |
| *** burris has quit IRC | 05:34 | |
| *** notmyname_ has joined #openstack | 05:36 | |
| *** notmyname has quit IRC | 05:36 | |
| *** notmyname_ is now known as notmyname | 05:36 | |
| *** ChanServ sets mode: +v notmyname | 05:36 | |
| *** notmyname has quit IRC | 05:38 | |
| *** notmyname has joined #openstack | 05:39 | |
| *** ChanServ sets mode: +v notmyname | 05:39 | |
| *** f4m8_ is now known as f4m8 | 05:41 | |
| *** f4m8 is now known as f4m8_ | 05:41 | |
| *** f4m8_ is now known as f4m8 | 05:41 | |
| *** notmyname has quit IRC | 05:45 | |
| *** notmyname has joined #openstack | 05:45 | |
| *** ChanServ sets mode: +v notmyname | 05:45 | |
| *** anticw has quit IRC | 05:47 | |
| *** anticw has joined #openstack | 05:47 | |
| *** littleidea has quit IRC | 06:01 | |
| *** damon__ has quit IRC | 06:12 | |
| *** EdwinGrubbs has quit IRC | 06:22 | |
| *** joearnol_ has quit IRC | 06:23 | |
| *** laurensell has quit IRC | 06:31 | |
| *** laurensell has joined #openstack | 06:32 | |
| *** alii_ has joined #openstack | 06:43 | |
| *** thimble has joined #openstack | 06:54 | |
| *** omidhdl has joined #openstack | 07:01 | |
| * JordanRinke yawns | 07:04 | |
| *** drico_ has quit IRC | 07:10 | |
| *** drico_ has joined #openstack | 07:11 | |
| *** arreyder has quit IRC | 07:14 | |
| *** nate_h has quit IRC | 07:14 | |
| *** tsharju has quit IRC | 07:14 | |
| *** befreax has joined #openstack | 07:20 | |
| *** nate_h has joined #openstack | 07:27 | |
| *** dendro-afk is now known as dendrobates | 07:31 | |
| *** dstromberg has quit IRC | 07:31 | |
| *** HugoKuo has joined #openstack | 07:34 | |
| *** brd_from_italy has quit IRC | 07:35 | |
| *** guigui1 has joined #openstack | 07:38 | |
| *** dstromberg has joined #openstack | 07:43 | |
| *** zenmatt_ has quit IRC | 07:47 | |
| *** maplebed has joined #openstack | 07:54 | |
| *** BK_man has quit IRC | 07:58 | |
| *** maplebed has quit IRC | 07:59 | |
| *** brd_from_italy has joined #openstack | 08:03 | |
| *** sandywalsh has quit IRC | 08:04 | |
| *** sandywalsh has joined #openstack | 08:05 | |
| *** herki has quit IRC | 08:06 | |
| *** herki has joined #openstack | 08:06 | |
| *** brd_from_italy has quit IRC | 08:12 | |
| ttx | soren: I closed the "create DB" bug and opened an "upgrade DB" bug to track remaining work (bug 705758) | 08:14 |
| uvirtbot | Launchpad bug 705758 in nova "Ubuntu packaging should handle (some) DB upgrades cases" [Wishlist,Confirmed] https://launchpad.net/bugs/705758 | 08:14 |
| *** miclorb_ has quit IRC | 08:14 | |
| *** herki has quit IRC | 08:15 | |
| *** greenisus has joined #openstack | 08:17 | |
| *** retr0h has quit IRC | 08:17 | |
| uvirtbot | New bug: #705758 in nova "Ubuntu packaging should handle (some) DB upgrades cases" [Wishlist,Confirmed] https://launchpad.net/bugs/705758 | 08:21 |
| *** dendrobates is now known as dendro-afk | 08:25 | |
| *** fabiand_ has joined #openstack | 08:27 | |
| *** allsystemsarego has joined #openstack | 08:29 | |
| *** calavera has joined #openstack | 08:30 | |
| *** fabiand_ has left #openstack | 08:31 | |
| *** fabiand_ has joined #openstack | 08:31 | |
| *** brd_from_italy has joined #openstack | 08:37 | |
| *** ibarrera has joined #openstack | 08:39 | |
| fabiand_ | soren: there is now something like timetable for fosdem .. | 08:41 |
| *** aimon_ has quit IRC | 08:50 | |
| *** brd_from_italy has quit IRC | 08:50 | |
| *** MarkAt2od has quit IRC | 08:51 | |
| uvirtbot | New bug: #705769 in nova "Error in injecting SSH public key during the creation of the VM" [Undecided,New] https://launchpad.net/bugs/705769 | 09:01 |
| *** MarkAt2od has joined #openstack | 09:12 | |
| *** ramkrsna has joined #openstack | 09:13 | |
| *** ramkrsna has joined #openstack | 09:13 | |
| *** brd_from_italy has joined #openstack | 09:20 | |
| soren | ttx: Awesome, I was just about to do that. | 09:21 |
| soren | ttx: Thanks! | 09:21 |
| soren | fabiand_: Yes, so people say. Thanks, though! | 09:21 |
| ttx | soren: i'm tagging "ubuntu" all the bugs that affect the packaging. I'm tired of crossfiling to ubuntu/nova while it does not affect distributed packages | 09:26 |
| ttx | if that makes sense | 09:26 |
| *** BK_man has joined #openstack | 09:38 | |
| *** fitzdsl has quit IRC | 09:42 | |
| *** pyros_xx has quit IRC | 09:47 | |
| *** jaypipes has quit IRC | 09:48 | |
| *** thimble has quit IRC | 09:49 | |
| *** thimble has joined #openstack | 09:50 | |
| *** jaypipes has joined #openstack | 10:02 | |
| uvirtbot | New bug: #705790 in nova "Parted cannot find device when using glance images in xenapi" [Undecided,New] https://launchpad.net/bugs/705790 | 10:11 |
| *** adiantum has joined #openstack | 10:20 | |
| *** reldan has joined #openstack | 10:33 | |
| *** thimble has quit IRC | 10:34 | |
| *** ramkrsna has quit IRC | 10:35 | |
| soren | ttx: Perhaps only file it against ubuntu/nova? | 10:36 |
| *** adiantum has quit IRC | 10:37 | |
| ttx | soren: we tend to lose track of it, and it doesn't really affect ubuntu packages in any release | 10:37 |
| soren | ttx: That's only because someone is failing to upload fresh nova's to Natty. | 10:38 |
| soren | ttx: That was supposed to happen (very) regularly. | 10:38 |
| ttx | soren: oh, then | 10:38 |
| * ttx fears adding another buglist to regularly scrub | 10:38 | |
| ttx | but makes sense, if it's in Natty. | 10:39 |
| *** thimble has joined #openstack | 10:42 | |
| *** adiantum has joined #openstack | 10:42 | |
| *** ramkrsna has joined #openstack | 10:49 | |
| *** ramkrsna has joined #openstack | 10:49 | |
| *** adiantum has quit IRC | 10:50 | |
| *** adiantum has joined #openstack | 10:51 | |
| uvirtbot | New bug: #705811 in nova "Cannot boot instances under use_ipv6=true and with FlatManager/FlatDHCPManager" [Undecided,In progress] https://launchpad.net/bugs/705811 | 11:06 |
| *** koji-iida has quit IRC | 11:12 | |
| *** reldan has quit IRC | 11:21 | |
| *** adiantum has quit IRC | 11:26 | |
| *** adiantum has joined #openstack | 11:26 | |
| * soren lunchesw | 11:31 | |
| soren | lunches, even. | 11:31 |
| soren | soeren: Do you play golf? | 11:31 |
| *** hansode has quit IRC | 11:33 | |
| *** adiantum has quit IRC | 11:33 | |
| *** adiantum has joined #openstack | 11:38 | |
| *** tsharju has joined #openstack | 11:53 | |
| *** reldan has joined #openstack | 11:53 | |
| *** arreyder has joined #openstack | 11:53 | |
| rcc | hi guys, i get in the volume session: | 11:55 |
| rcc | nova.exception.Error: volume group nova-volumes doesn't exist | 11:55 |
| rcc | i guess i need to manually create a logical volume | 11:56 |
| soren | Do you have a volume group by that name? | 11:56 |
| rcc | where's the doc for that? | 11:56 |
| soren | Not a logical volume. A volume group. | 11:56 |
| soren | I'm not sure. | 11:56 |
| rcc | erm, yeah, a vg | 11:56 |
| * soren hasn't really kept up with the docs | 11:56 | |
| *** adiantum has quit IRC | 11:56 | |
| rcc | heh | 11:56 |
| rcc | google doesn't help much either :/ | 11:57 |
| soren | Look for vgcreate on nova.openstack.org. | 11:57 |
| fraggeln | soren: lvm2? :) | 11:57 |
| soren | fraggeln: ? | 11:57 |
| fraggeln | vgcreate is a command in lvm2 :) | 11:57 |
| soren | Yes. | 11:57 |
| fraggeln | does nova use lvm? | 11:58 |
| soren | It's the default for nova-volume, yes. | 11:58 |
| fraggeln | check. | 11:58 |
| rcc | "Your search query "vgcreate" didn't return any results" QQ | 11:58 |
| rcc | oh wait, i searched on wiki.openstack.org | 11:58 |
| rcc | let me check | 11:59 |
| rcc | ahh | 11:59 |
| rcc | now | 11:59 |
| rcc | k, thx | 11:59 |
| *** omidhdl has left #openstack | 12:01 | |
| *** adiantum has joined #openstack | 12:03 | |
| rcc | up and running, thx again soren | 12:05 |
| rcc | it'd be nice if nova.sh checked for nova-volume vg existance and create if not. you think is a valid RFE ? | 12:06 |
| *** adiantum has quit IRC | 12:11 | |
| *** HugoKuo has quit IRC | 12:14 | |
| *** adiantum has joined #openstack | 12:18 | |
| *** ctennis has quit IRC | 12:23 | |
| *** hisaharu has quit IRC | 12:25 | |
| *** adiantum has quit IRC | 12:27 | |
| *** adiantum has joined #openstack | 12:32 | |
| *** ctennis has joined #openstack | 12:38 | |
| *** ctennis has joined #openstack | 12:38 | |
| *** blakeyeager has joined #openstack | 12:38 | |
| *** piezo has quit IRC | 12:41 | |
| *** piezo has joined #openstack | 12:47 | |
| *** calavera has quit IRC | 13:01 | |
| soren | rcc: It's kind of hard to just create a volume group. | 13:03 |
| soren | rcc: You need block storage available to do so. | 13:03 |
| *** westmaas has quit IRC | 13:04 | |
| *** MarkAt2od has quit IRC | 13:05 | |
| *** dendro-afk is now known as dendrobates | 13:06 | |
| Adri2000 | hi everyone | 13:06 |
| Adri2000 | I'm trying to understand how network should work with nova | 13:07 |
| Adri2000 | are compute nodes supposed to be in a separate private network? | 13:07 |
| *** arthurc has joined #openstack | 13:16 | |
| *** DubLo7 has joined #openstack | 13:20 | |
| *** adiantum has quit IRC | 13:23 | |
| soren | Alright, I'm all Nattyed up. | 13:27 |
| *** adiantum has joined #openstack | 13:27 | |
| *** kashyapc has quit IRC | 13:28 | |
| *** drico_ has quit IRC | 13:36 | |
| *** aliguori has joined #openstack | 13:41 | |
| rackerhacker | soren: nattyed up? does that involve this? http://ciceronito.files.wordpress.com/2008/11/495.jpg | 13:42 |
| *** drico has joined #openstack | 13:42 | |
| BK_man | ppl, SHOULD I ping one instance from another? My network is VlanManager and these two instances are in the same project (==VLAN). | 13:43 |
| *** hadrian has joined #openstack | 13:44 | |
| *** arthurc has quit IRC | 13:45 | |
| *** arthurc has joined #openstack | 13:45 | |
| *** westmaas has joined #openstack | 13:49 | |
| *** zul has quit IRC | 13:52 | |
| *** zul has joined #openstack | 13:53 | |
| soren | rackerhacker: Please tell me that isn't a real product. | 13:53 |
| rackerhacker | soren: it is a real product | 13:53 |
| soren | rackerhacker: Light mineral water? srsly? | 13:53 |
| rackerhacker | soren: nah, it's beer. awful beer that happens to be very cheap. | 13:54 |
| soren | Ah. | 13:56 |
| *** DubLo7 has quit IRC | 13:57 | |
| dabo | rackerhacker: it's a stretch to call that stuff 'beer' | 13:58 |
| rackerhacker | you make a valid point | 13:58 |
| *** befreax has quit IRC | 13:59 | |
| zul | soren: im going to do an upload today? | 14:01 |
| zul | soren: er...im going to do an upload today | 14:02 |
| soren | zul: GREAT. | 14:02 |
| soren | MY CAPS LOCK KEY IS BROKEN... | 14:02 |
| zul | heh | 14:02 |
| soren | Yay! Fixed. | 14:04 |
| *** ramkrsna has quit IRC | 14:06 | |
| rcc | soren, do you know if there's an issue on lucid euca2ools ? no matter what command I run , i get xml parser errors | 14:07 |
| rcc | http://paste.openstack.org/show/533/ | 14:09 |
| *** ppetraki has joined #openstack | 14:10 | |
| soren | rcc: "EC2_URL not specified. Trying http://localhost:8773/services/Eucalyptus" doesn't look promising. | 14:10 |
| rcc | it defaults to that URL | 14:11 |
| rcc | and I can indeed access it via a browser | 14:11 |
| *** fitzdsl has joined #openstack | 14:11 | |
| soren | Whuh? | 14:11 |
| soren | Are you running eucalyptus, too? | 14:11 |
| jaypipes | ttx: "Doing a morning stand-up meeting with a team of 60 in widely different timezones is a bit difficult." Heh, how true. :) | 14:12 |
| *** blakeyeager has quit IRC | 14:12 | |
| rcc | no, i'm not, nova is... | 14:12 |
| ttx | jaypipes: heh | 14:13 |
| *** befreax has joined #openstack | 14:13 | |
| soren | rcc: Hmm.. You're right, that does work. I'm surprised. | 14:14 |
| soren | rcc: I'm not sure what could cause that. | 14:15 |
| rcc | erm | 14:17 |
| rcc | o-k | 14:17 |
| rcc | so | 14:17 |
| rcc | i was running euca2ools outside the virtualenv, getting that error | 14:17 |
| rcc | but now, I copied euca2ools site-packages within the virtualenv and commands work | 14:18 |
| rcc | damn, there are a few steps that are missing the wiki quickstart guide for newbies like me :/ | 14:18 |
| rcc | i'll add a note to the wiki page | 14:18 |
| *** dirakx has quit IRC | 14:25 | |
| *** thimble has quit IRC | 14:27 | |
| ttx | nova-core: please give https://code.launchpad.net/~ttx/nova/lp704896/+merge/46892 a review, if we change the log format better do it early than late | 14:33 |
| *** littleidea has joined #openstack | 14:34 | |
| *** littleidea has quit IRC | 14:42 | |
| *** littleidea has joined #openstack | 14:44 | |
| *** littleidea has quit IRC | 14:46 | |
| uvirtbot | New bug: #705914 in nova "nova-network should stop dnsmasq when it stops" [Undecided,New] https://launchpad.net/bugs/705914 | 14:46 |
| *** aliguori has quit IRC | 14:47 | |
| *** arthurc has quit IRC | 14:48 | |
| *** arthurc has joined #openstack | 14:48 | |
| *** omidhdl has joined #openstack | 14:48 | |
| *** littleidea has joined #openstack | 14:52 | |
| *** adiantum has quit IRC | 14:57 | |
| * Adri2000 still trying to setup network correctly | 14:58 | |
| Adri2000 | the br100 on my compute node doesn't have an IP address, I guess there's something wrong? | 14:59 |
| Adri2000 | br100 interface* | 14:59 |
| *** aliguori has joined #openstack | 15:01 | |
| *** kashyapc has joined #openstack | 15:03 | |
| *** EdwinGrubbs has joined #openstack | 15:05 | |
| *** EdwinGrubbs has joined #openstack | 15:05 | |
| *** littleidea has quit IRC | 15:05 | |
| *** adiantum has joined #openstack | 15:11 | |
| *** littleidea has joined #openstack | 15:11 | |
| *** dprince has joined #openstack | 15:15 | |
| *** adiantum has quit IRC | 15:19 | |
| openstackhudson | Project nova-tarmac build #54,393: FAILURE in 24 sec: http://hudson.openstack.org/job/nova-tarmac/54393/ | 15:21 |
| soren | LP codehosting seems to be down. | 15:22 |
| annegentle | soren: heh. I was just gonna ask if anyone else saw it | 15:24 |
| openstackhudson | Project nova-tarmac build #54,394: STILL FAILING in 12 sec: http://hudson.openstack.org/job/nova-tarmac/54394/ | 15:25 |
| *** reldan has quit IRC | 15:26 | |
| *** gondoi has joined #openstack | 15:26 | |
| annegentle | Adri2000: I think you have to integrate the br100 interface into your network - check the bottom of this page: http://wiki.openstack.org/InstallFromSource as well as the Network Configuration section on this page: http://wiki.openstack.org/NovaInstall/MultipleServer and let me know how it goes. | 15:27 |
| *** reldan has joined #openstack | 15:30 | |
| *** omidhdl has quit IRC | 15:30 | |
| openstackhudson | Project nova-tarmac build #54,395: STILL FAILING in 26 sec: http://hudson.openstack.org/job/nova-tarmac/54395/ | 15:31 |
| *** adiantum has joined #openstack | 15:32 | |
| *** ibarrera has quit IRC | 15:33 | |
| openstackhudson | Project nova-tarmac build #54,396: STILL FAILING in 13 sec: http://hudson.openstack.org/job/nova-tarmac/54396/ | 15:35 |
| openstackhudson | Project nova-tarmac build #54,397: STILL FAILING in 26 sec: http://hudson.openstack.org/job/nova-tarmac/54397/ | 15:41 |
| *** daleolds has joined #openstack | 15:43 | |
| *** zenmatt has joined #openstack | 15:45 | |
| *** dirakx has joined #openstack | 15:45 | |
| openstackhudson | Project nova-tarmac build #54,398: STILL FAILING in 25 sec: http://hudson.openstack.org/job/nova-tarmac/54398/ | 15:46 |
| *** pvo_away is now known as pvo | 15:46 | |
| *** johnpur has joined #openstack | 15:47 | |
| *** ChanServ sets mode: +v johnpur | 15:47 | |
| *** omidhdl has joined #openstack | 15:49 | |
| *** dirakx has quit IRC | 15:51 | |
| openstackhudson | Project nova-tarmac build #54,399: STILL FAILING in 27 sec: http://hudson.openstack.org/job/nova-tarmac/54399/ | 15:51 |
| *** drico has quit IRC | 15:51 | |
| *** drico has joined #openstack | 15:51 | |
| *** rnirmal has joined #openstack | 15:52 | |
| openstackhudson | Project nova-tarmac build #54,400: STILL FAILING in 24 sec: http://hudson.openstack.org/job/nova-tarmac/54400/ | 15:56 |
| *** dragondm has joined #openstack | 15:57 | |
| *** fabiand_ has quit IRC | 15:57 | |
| *** j05h1 has joined #openstack | 15:58 | |
| *** j05h has quit IRC | 15:58 | |
| openstackhudson | Project nova-tarmac build #54,401: STILL FAILING in 25 sec: http://hudson.openstack.org/job/nova-tarmac/54401/ | 16:01 |
| jaypipes | thank you openstackhudson for your nagging. | 16:02 |
| *** iRTermite has joined #openstack | 16:02 | |
| soren | jaypipes: I've disabled the tarmac job. | 16:04 |
| jaypipes | soren: ah, thx :) | 16:05 |
| creiht | yay lp | 16:05 |
| creiht | :/ | 16:05 |
| *** littleidea has quit IRC | 16:09 | |
| *** pharkmillups has joined #openstack | 16:11 | |
| *** retr0h has joined #openstack | 16:15 | |
| *** retr0h has joined #openstack | 16:15 | |
| BK_man | lp is back online | 16:15 |
| *** joearnold has joined #openstack | 16:16 | |
| *** joearnold has quit IRC | 16:18 | |
| ttx | soren: you can reenable tarmac | 16:18 |
| soren | Yay. | 16:19 |
| soren | Done. | 16:19 |
| *** aimon has joined #openstack | 16:19 | |
| * soren heads dinnerwards | 16:21 | |
| *** guigui1 has quit IRC | 16:22 | |
| openstackhudson | Yippie, build fixed! | 16:24 |
| openstackhudson | Project nova-tarmac build #54,402: FIXED in 4 min 12 sec: http://hudson.openstack.org/job/nova-tarmac/54402/ | 16:24 |
| *** littleidea has joined #openstack | 16:26 | |
| openstackhudson | Project nova build #438: SUCCESS in 1 min 28 sec: http://hudson.openstack.org/job/nova/438/ | 16:28 |
| openstackhudson | Tarmac: Change default log formats so that: | 16:28 |
| openstackhudson | ?* they include a timestamp (necessary to correlate logs) | 16:28 |
| openstackhudson | ?* no longer display version on every line (shorter lines) | 16:28 |
| openstackhudson | ?* use [-] instead of [N/A] (shorter lines, less scary-looking) | 16:28 |
| openstackhudson | ?* show level before logger name (better human-readability) | 16:28 |
| openstackhudson | Display version once at node startup time instead. | 16:28 |
| openstackhudson | Fix description of the log format flags. | 16:28 |
| openstackhudson | Project swift build #170: SUCCESS in 27 sec: http://hudson.openstack.org/job/swift/170/ | 16:31 |
| openstackhudson | Tarmac: During a CNAME lookup, exception NXDOMAIN isn't caught by the check for DNSException, so need to also check for it. | 16:31 |
| *** aimon has quit IRC | 16:32 | |
| *** kashyapc has quit IRC | 16:34 | |
| *** MarkAtwood has joined #openstack | 16:34 | |
| jaypipes | sirp_: ping | 16:35 |
| sirp_ | jaypipes: hey | 16:36 |
| jaypipes | sirp_: heyo. mind pulling lp:~jaypipes/glance/bug700162 again and verifying the upload of ubuntu-lucid now works without a MemoryError? | 16:36 |
| colinnich | creiht: woohoo | 16:37 |
| creiht | :) | 16:37 |
| sirp_ | jaypipes: sure will do | 16:38 |
| jaypipes | sirp_: cheers! | 16:38 |
| * jk0 can't wait for these changes to make it to trunk :D | 16:38 | |
| jaypipes | creiht: so, yes, looks like just doing chunked transfer encoding has solved our 2G limit issues in Glance... | 16:38 |
| *** ascovel has joined #openstack | 16:38 | |
| jaypipes | jk0: the glance ones? :) | 16:38 |
| jk0 | yeah | 16:39 |
| jaypipes | koo. | 16:39 |
| ascovel | quick question for any devs out there | 16:39 |
| jaypipes | ascovel: shoot. | 16:39 |
| creiht | jaypipes: woot! | 16:39 |
| jk0 | I just upped my domU to 4GB to bypass the memory issue, so I'm mostly looking forward to the other fixes :) | 16:39 |
| ascovel | what is your take on the stability of the nightly builds? | 16:39 |
| jaypipes | ascovel: of Nova, Swift, or Glance? | 16:39 |
| ascovel | nova | 16:39 |
| jaypipes | ascovel: and I assume you mean "bzr branch lp:nova" as "nightly builds" :) We don't tarball nightly... | 16:40 |
| ascovel | yes | 16:40 |
| *** dendrobates is now known as dendro-afk | 16:40 | |
| jaypipes | ascovel: ok. so, I would say the stability right now is not the greatest. There have been a crap-ton (technical term) of features added in Bexar, so we're in heavy bug-fixing mode right now. I think in Cactus (next release), the nighly builds should be a bit more stable. | 16:41 |
| jaypipes | ascovel: the above applies to Glance as well, but not really for Swift, which is much more mature than either Nova or Glance. | 16:41 |
| jaypipes | ascovel: of course, it behooves you to get the opinions of others in the channel other than me, too :) | 16:42 |
| ascovel | cool, thanks man | 16:42 |
| jaypipes | np | 16:42 |
| jaypipes | sirp_: so, with that bug fixed, it looks like we're pretty good with Glance for Bexar, right? | 16:43 |
| *** dendro-afk is now known as dendrobates | 16:44 | |
| jaypipes | jk0: what's our opinion about Glance? Stabilizing OK for Bexar? | 16:44 |
| jk0 | yeah, I think so, especially after these merges hit | 16:44 |
| jk0 | it seems to have been holding up pretty well over the past few days | 16:45 |
| jaypipes | jk0: cool. I'm working on beefing up the docs today and tomorrow, including a few nice little diagrams to explain the architecture. | 16:46 |
| jk0 | I can think of more than one person who would love to look those over :) | 16:46 |
| *** mray has joined #openstack | 16:46 | |
| jaypipes | jk0: well, only a couple weeks ago, Glance had no docs at all. Now we've got a decent start at http://glance.openstack.org :) | 16:49 |
| *** kashyapc has joined #openstack | 16:51 | |
| sirp_ | jaypipes: yeah, other than that bug, glance has been working really pretty well for me | 16:51 |
| zul | i noticed you guys added glance to the debian packaging does that mean its actually usable now? | 16:52 |
| jaypipes | zul: hehe, yes :) | 16:52 |
| jaypipes | zul: it's in PyPI now, too... easy_install or pip install it.. | 16:52 |
| zul | jaypipes: blah i rather use apt ;) | 16:53 |
| jaypipes | zul: that said, the next tarball (today or tomorrow) will be REALLY usable. :) | 16:53 |
| * jaypipes notes that there really shouldn't have degrees of usability... | 16:54 | |
| *** mdomsch has joined #openstack | 16:54 | |
| BK_man | jaypipes: hi! Does glance need a special setup? We tried it today and it failed. http://paste.openstack.org/show/535/ | 16:57 |
| BK_man | jaypipes: after failback to default image_service we got api working again | 16:57 |
| jaypipes | BK_man: one sec, looking.. | 16:58 |
| jaypipes | BK_man: what revno are you on? That looks like a bug that was fixed this last couple days... | 16:59 |
| sirp_ | jaypipes: looks like in unit-tests it needs to be req.body and functionally it needs to be req.body_file | 16:59 |
| sirp_ | jaypipes: may be a try/except fallback is in order | 16:59 |
| BK_man | jaypipes: glance is 0.1.2 | 16:59 |
| jaypipes | sirp_: :( grr... | 16:59 |
| mtaylor | soren: morning | 16:59 |
| BK_man | jaypipes: nova was bzr597 | 16:59 |
| *** brd_from_italy has quit IRC | 16:59 | |
| jaypipes | BK_man: so you are using glance from packaging and not bzr? | 17:00 |
| BK_man | jaypipes: you can ask adiantum for details | 17:00 |
| jaypipes | BK_man: forget that, I see you are from the paste output :) | 17:00 |
| BK_man | jaypipes: corrent | 17:00 |
| jaypipes | BK_man: there is a fix for that bug in glance's trunk. Or you can wait ~2-3 hours until we cut a new tarball and upload to PyPI and LP... | 17:01 |
| jaypipes | BK_man: think you can wait? | 17:01 |
| BK_man | jaypipes: sure. will wait and than build it again from trunk | 17:01 |
| jaypipes | BK_man: sounds good. thx mate! | 17:01 |
| jaypipes | sirp_: try/catch is a bit messy, but if it works... | 17:02 |
| jaypipes | sirp_: that's the problem with Webob.Request != httplib.HTTPResponse :) | 17:02 |
| sirp_ | jaypipes: yeah, makes me wonder if we're using the wrong WebOb method, does it have a better read() abstraction | 17:02 |
| sirp_ | yeah | 17:02 |
| mtaylor | soeren: I was just looking through the ppa_script on the hudson box and noticed that it sets name and email to Soren and soren@openstack.org - can I suggest changing those to "OpenStack Hudson" and "hudson@openstack.org"? | 17:02 |
| sirp_ | jaypipes: there should be a WebOb.safe_read which always returns a file-like, or something *grumble grumble* | 17:03 |
| *** aimon has joined #openstack | 17:04 | |
| sirp_ | clarify ^ that doesn't exist, though, to my knowledge | 17:04 |
| jaypipes | creiht: any insight from your vast wealth of webob experience? :) we're having issues that in testing, webob.Request.body is used, while in normal running, httplib.HTTPRequest.body_file is used... | 17:06 |
| jaypipes | annegentle: you're all set. should be merging docs as we speak.. | 17:07 |
| *** Pentheus has quit IRC | 17:09 | |
| *** maplebed has joined #openstack | 17:10 | |
| *** dfg has joined #openstack | 17:10 | |
| creiht | jaypipes: ok, so what is the problem that you run into? | 17:10 |
| annegentle | jaypipes: thanks! you have mad bzr skillz | 17:10 |
| johnpur | mtaylor: sounds good | 17:11 |
| mtaylor | hey johnpur ! | 17:11 |
| johnpur | hey! | 17:12 |
| openstackhudson | Project nova build #439: SUCCESS in 1 min 29 sec: http://hudson.openstack.org/job/nova/439/ | 17:12 |
| openstackhudson | Tarmac: Updated docs for db sync requirements; merged with Vish's similar doc updates | 17:12 |
| jaypipes | creiht: so, in the Glance server we read from request.body_file and pass that body_file object to the backend store, which iterates over it, writing chunks from the request.body_file object. | 17:12 |
| creiht | k | 17:12 |
| jaypipes | creiht: unfortunately, in testing, we use webob.Request, which has no body_file attribute... | 17:12 |
| jaypipes | creiht: so we're stuck trying to figure out a way that we can accurately test body_file when webob.Request has no such attribute... | 17:13 |
| creiht | jaypipes: can you point me to some code? | 17:13 |
| jaypipes | creiht: sure, one sec. | 17:13 |
| jaypipes | creiht: http://bazaar.launchpad.net/~rconradharris/glance/bug700162/revision/45 | 17:14 |
| jaypipes | creiht: expand the glance/server.py diff there... | 17:14 |
| creiht | and to make sure I understand correctly, you are pretty sure that the code works, but you aren't sure how to test it | 17:14 |
| jaypipes | creiht: one sec, here's the test stubout code of relevance: http://bazaar.launchpad.net/~rconradharris/glance/bug700162/view/head:/tests/stubs.py (lines 252-295 | 17:15 |
| jaypipes | creiht: so, when rick changes that code in glance/server.py from if req.body: to if req.body_file:, then functionally, Glance works perfectly. But the tests break. When it's if req.body and not if req.body_file, the tests work and the server fails functionally.. | 17:16 |
| jaypipes | sirp_: is that a good description of the problem? | 17:17 |
| creiht | ahh | 17:17 |
| *** kbringard has joined #openstack | 17:17 | |
| sirp_ | the issue jay i think is that body_file always exists | 17:17 |
| openstackhudson | Project nova build #440: SUCCESS in 1 min 31 sec: http://hudson.openstack.org/job/nova/440/ | 17:17 |
| sirp_ | its just an empty cStringIO | 17:17 |
| openstackhudson | Tarmac: No longer chmod 0777 instance directories, since nova works just fine without them. | 17:17 |
| sirp_ | but that still evaulates to truthy | 17:17 |
| sirp_ | which causes the logic to break | 17:17 |
| jaypipes | creiht: I'm thinking we either need to hack the stubs more, or switch to an alternate way of testing? | 17:18 |
| jaypipes | mmmm, truthiness. | 17:18 |
| creiht | hrm | 17:18 |
| sirp_ | i think this is a problem in code not tests | 17:18 |
| *** mray has quit IRC | 17:19 | |
| sirp_ | since body_file always has an object… maybe we should test content-lenght to determine if an entity body is present | 17:19 |
| sirp_ | Content-Length: 0 or Transfer-Encoding: chunked would indicate | 17:19 |
| jaypipes | sirp_: hmm, that might work. | 17:20 |
| creiht | yeah I agree with checking the content_length | 17:21 |
| jaypipes | sirp_: wanna try that approach and let me know if that works please? | 17:22 |
| sirp_ | yep, working on it right now :) | 17:22 |
| jaypipes | coolio :) | 17:22 |
| creiht | jaypipes: if it is the test failure, then setting the body on your request in the test, should make body_file available | 17:24 |
| creiht | jaypipes: also another interesting possibility for testing is someting like: | 17:24 |
| creiht | http://code.google.com/p/wsgi-intercept/ | 17:24 |
| jaypipes | creiht: that's how I read the docs, too, but for some reason I was running into issues. Prolly was PEBCAK, though ;) | 17:24 |
| jaypipes | intersting, I'll check that out. | 17:25 |
| creiht | It creates fake wsgi apps, and patches httplib, so that you make calls to it like you normally would rather than just faking webob requests | 17:25 |
| jaypipes | creiht: ooooh, that sounds nice. | 17:25 |
| creiht | jaypipes: I tested the body thing in a shell, and it seems to work well | 17:25 |
| jaypipes | creiht: w00t. | 17:26 |
| creiht | I've used wsgi-intercept on another project, and it worked nicely | 17:26 |
| creiht | though it adds an extra dependency | 17:26 |
| sirp_ | jaypipes: this seems to work in all cases: http://paste.openstack.org/show/536/ | 17:28 |
| creiht | sirp_: Why pull that out to has_body? | 17:31 |
| creiht | or why not just put all that in the if statement? | 17:31 |
| *** joearnold has joined #openstack | 17:31 | |
| sirp_ | creiht: just thought it read a little clearer (not a line-miser ;) ) | 17:32 |
| *** EdwinGrubbs has quit IRC | 17:32 | |
| jaypipes | creiht: nothing wrong with using an extra line for clarity :) | 17:33 |
| jaypipes | sirp_: cool, I will use that code and push to the merge proposal branch of mine... | 17:33 |
| creiht | jaypipes: true | 17:33 |
| creiht | :) | 17:33 |
| *** mray has joined #openstack | 17:33 | |
| *** adamsih300u has joined #openstack | 17:40 | |
| *** dirakx has joined #openstack | 17:40 | |
| *** reldan has quit IRC | 17:43 | |
| *** adiantum has quit IRC | 17:44 | |
| *** retr0h has quit IRC | 17:50 | |
| *** blueadept has joined #openstack | 17:52 | |
| jarrod | man, just ordered 3 more dell r610s for compute nodes | 17:54 |
| jarrod | im impressed with openstack | 17:54 |
| *** aimon has quit IRC | 17:56 | |
| blueadept | i heard InterNAP just released a new storage cloud using open stack | 17:57 |
| *** jdurgin has joined #openstack | 17:57 | |
| *** adiantum has joined #openstack | 17:58 | |
| PiotrSikora | blueadept: yeah, announcement is here: http://www.internap.com/2011/01/18/internap-expands-it-infrastructure-services-offering-with-enterprise-cloud-storage/ | 17:58 |
| blueadept | cool, so this openstack is all written in python, correct? | 18:00 |
| jarrod | yes | 18:00 |
| vishy | soren: http://jaggermafia.com/?p=8359 | 18:00 |
| jarrod | internap is using cloudfiles | 18:00 |
| blueadept | so what how does InternNAP's product integrate with the storage system? has internap built their own control panel for the underlying logic? | 18:01 |
| jarrod | yea, they implemented cloudfiles and wrote their own wrapper, interface around it using the API | 18:02 |
| PiotrSikora | jarrod: cloudfiles as in they are reselling rackspace? | 18:02 |
| PiotrSikora | jarrod: or did you mean swift? | 18:02 |
| jarrod | swift | 18:02 |
| PiotrSikora | ok, you've got me confused for a moment :P | 18:02 |
| blueadept | rackspace opensourced the core server and distribution logic for the nodes, but you still have to write a CRUD for the services it runs | 18:03 |
| blueadept | thats my basic understanding | 18:03 |
| blueadept | im going to get swift running on my servers tonight and see what i can do with it | 18:03 |
| blueadept | i have so many projects going, but this really interests me | 18:03 |
| *** aimon has joined #openstack | 18:07 | |
| *** westmaas has quit IRC | 18:07 | |
| *** mray has joined #openstack | 18:10 | |
| *** adiantum has quit IRC | 18:10 | |
| *** Ryan_Lane has joined #openstack | 18:12 | |
| *** dudeman has joined #openstack | 18:13 | |
| dudeman | i tried testing, http://www.ansolabs.com/articles/hacking-nova, and ssh fails with no route to host 10.0.0.3 | 18:14 |
| dudeman | everything looks good except that | 18:15 |
| dudeman | reinstalled and get same error | 18:16 |
| dudeman | any pointers from anyone on this? | 18:17 |
| dudeman | I watched vish install this with no issue last night, has anyone fixed this error before? | 18:21 |
| dudeman | quiet in here, is this irc client working? | 18:21 |
| dudeman | hello? | 18:22 |
| *** arthurc has quit IRC | 18:23 | |
| dudeman | syn | 18:23 |
| dudeman | ack | 18:23 |
| *** retr0h has joined #openstack | 18:23 | |
| *** adiantum has joined #openstack | 18:24 | |
| dudeman | i tried testing, http://www.ansolabs.com/articles/hacking-nova, and ssh fails with no route to host 10.0.0.3 | 18:24 |
| dudeman | reinstalled and get same error | 18:25 |
| dudeman | has anyone seen this before? | 18:25 |
| *** BK_man has quit IRC | 18:26 | |
| dudeman | anyone? | 18:26 |
| dudeman | bueller? | 18:26 |
| kbringard | sorry dudeman, I've not seen that before | 18:27 |
| annegentle | dudeman: hm. I'm not sure SSH is installed on the image that the nova.sh script downloads. | 18:28 |
| vishy | redbo: here? | 18:28 |
| kbringard | can you ping 10.0.0.3? | 18:29 |
| dudeman | no cant ping, its a routing error, but route -n shows the 10.0.0.0 route | 18:29 |
| johnpur | vishy: redbo is afk | 18:30 |
| kbringard | if you do an ifconfig, do you see the vnet adapter? | 18:30 |
| kbringard | does brctl show it attached to the bridge? | 18:30 |
| vishy | redbo: it seems that you are responsible for the patch that fixes eventlet / ipv6? Or is that a different redbo? redbo = Mike Barton? | 18:30 |
| vishy | johnpur: ^^ is redbo Mike Barton? | 18:31 |
| dudeman | kbringard: it does seem to be attached | 18:31 |
| annegentle | vishy: yep, that's right. They're probably at lunch now, is my guess. | 18:31 |
| retr0h | johnpur: /who redbo agrees it is ;) | 18:31 |
| vishy | ok I'll wait for him to return then | 18:32 |
| retr0h | howdy kbringard :) | 18:32 |
| kbringard | hai retr0h | 18:32 |
| kbringard | oh, dudeman... I seem to recall having network issues due to ipv6 stuff | 18:33 |
| kbringard | although I think it worked when I used nova.sh, it was only when I was running it from init that I had problems | 18:33 |
| dudeman | I'mnot using v6 on this system | 18:34 |
| *** nelson has quit IRC | 18:34 | |
| kbringard | right, that was the problem I had | 18:34 |
| kbringard | it was something about writing out iptables ipv6 rules | 18:34 |
| kbringard | I had to set --use_ipv6=false in my nova.conf | 18:34 |
| annegentle | kbringard: ah there you go | 18:35 |
| kbringard | no idea if that's your issue | 18:35 |
| dudeman | i'll try that | 18:35 |
| dudeman | thanks! | 18:35 |
| retr0h | dudeman: u have an epic handle! | 18:35 |
| kbringard | no problem, hopefully it works :-D | 18:35 |
| kbringard | so, I have a question about the FlatManager | 18:37 |
| *** tonywolf has joined #openstack | 18:37 | |
| kbringard | in the docs, it says | 18:37 |
| kbringard | Creating multiple networks for for one manager is currently not supported | 18:37 |
| *** kpepple has joined #openstack | 18:38 | |
| kbringard | so, does that mean, that I couldn't create 10.1.100.0/24 vlan tagged as 100 through br_100 and 10.1.101.0/24 vlan tagged as 101 through br_101? | 18:39 |
| kbringard | I want to break my network up into /24s, and map a bridge device to each network... then have the network manager attach to the appropriate bridge based on the IP it assigned to the VM | 18:40 |
| *** blamar has joined #openstack | 18:42 | |
| *** adiantum has quit IRC | 18:44 | |
| *** EdwinGrubbs has joined #openstack | 18:44 | |
| retr0h | yeah | 18:46 |
| *** morfeas has quit IRC | 18:48 | |
| *** adiantum has joined #openstack | 18:48 | |
| *** morfeas has joined #openstack | 18:49 | |
| *** pharkmillups has quit IRC | 18:49 | |
| retr0h | along the lines of what kbringard said, we 802.1Q tag our vnets (in OpenNebula) so we can bring up a guest on any host in our DC.. trying to tackle that in openstack | 18:49 |
| retr0h | our neteng likes to assign us /24 vnets vs one large netblock | 18:50 |
| kbringard | dudeman: disabling ipv6 solve your problem? | 18:51 |
| kbringard | I think I tumbleweeded the room | 18:55 |
| redbo | vishy: that's me | 18:56 |
| vishy | redbo: cool, nova needs the ipv6 fix to fully support ipv6. Any idea how we can get it into eventlet proper? In the meantime we probably need to build a custom deb package with the fix and get it into the ppa | 18:57 |
| retr0h | kbringard: <whistle>ooo wa ooo wa oooo - da da daaaaaaa!</whistle> | 18:57 |
| *** spectorclan has joined #openstack | 18:57 | |
| *** pvo is now known as pvo_away | 18:57 | |
| vishy | dudeman: hey | 18:58 |
| vishy | dudeman: check the output in nova-network and see if it is reporting the ip getting leased | 18:58 |
| vishy | dudeman: also verify that euca-get-console-output i-1 isn't showing any strange errors | 18:59 |
| *** pvo_away is now known as pvo | 18:59 | |
| jaypipes | sirp_: alrighty, just waiting for the final docs/man page merge to go into trunk and then let's cut a new tarball... | 18:59 |
| redbo | vishy: it's already in the eventlet trunk, he just hasn't done a release since then. I don't know what it'd take to do a ppa .deb from their trunk, mtaylor has been handling those. | 19:00 |
| dudeman | thanks retr0h | 19:00 |
| dudeman | kringard: disabling v6 did not fix it | 19:00 |
| kbringard | bummer | 19:00 |
| kbringard | vishy is the man, if you do what he tells you, it'll likely get fixed :-) | 19:01 |
| vishy | dudeman: is your host 64 bit? | 19:01 |
| vishy | redbo: ok, hopefully mtaylor or soren can give us a trunk eventlet then. I think they will be much faster than I :) | 19:02 |
| creiht | blueadept: cool... feel free to let me know if you run into any issues, or have any questions | 19:02 |
| vishy | s/I/me | 19:02 |
| dudeman | vish: euca-get-console-output looks ok | 19:03 |
| dudeman | checking for lease | 19:03 |
| dudeman | using 32bit maverick | 19:03 |
| dudeman | is 64 required? | 19:03 |
| creiht | vishy: did you still have a question from last night? | 19:03 |
| vishy | creiht: I can't remember why I was pinging you, let me see if anotherjesse can jog my memory | 19:04 |
| creiht | hehe | 19:04 |
| vishy | dudeman: should be | 19:04 |
| vishy | dudeman: the image is 64 bit | 19:04 |
| vishy | dudeman: does get console output show any output at all? | 19:05 |
| vishy | dudeman: generally when you try to run 64 bit on 32 bit host it just doesn't actually run | 19:05 |
| dudeman | vish: ok, using pae kernel. will reinstall 64 | 19:06 |
| dudeman | vish: no lease for 10.0.0.3 in network screen window | 19:06 |
| vishy | dudeman: that will be easier. I was going to say you could try grabbing a 32 bit ttylinux image | 19:06 |
| kbringard | vishy: if you have a moment, I have a question about the network manager | 19:07 |
| vishy | kbringard: sure | 19:07 |
| dudeman | vish: thanks! | 19:07 |
| *** westmaas has joined #openstack | 19:08 | |
| vishy | dudeman: np | 19:08 |
| dudeman | off to grab a new maverick... | 19:08 |
| kbringard | our netops guys don't like to give out more than a /24 or /23 to control broadcast, so what I'm wanting to do is | 19:08 |
| *** dudeman has left #openstack | 19:08 | |
| kbringard | create lets say 4 networks, each a /24 | 19:08 |
| kbringard | and assign them to 4 different bridges | 19:09 |
| kbringard | each bridge will VLAN tag for the network it's physically connected to | 19:09 |
| kbringard | the ideally, using FlatManager or FlatDHCPManager | 19:09 |
| kbringard | when a VM gets assigned an IP, it'll attach to the appropriate bridge (as defined in the network config in the DB) | 19:10 |
| kbringard | then the vlan trunking on the switches will allow everything to talk to each other | 19:10 |
| kbringard | so, my question is, in the docs, it says something about only one network being supported in the network manager | 19:10 |
| kbringard | does that mean I can't do what I want to do? | 19:11 |
| *** EdwinGrubbs has quit IRC | 19:11 | |
| vishy | kbringard: so each rack has its own vlan? | 19:11 |
| kbringard | yea | 19:11 |
| vishy | kbringard: easier might be to just stick one network-manager on each rack | 19:12 |
| kbringard | I have OpenNebula setup this way now, and it works fine... so I know I can tag at the bridge level... from an OS perspective | 19:12 |
| vishy | kbringard: s/network-manager/nova-network | 19:12 |
| kbringard | well, actually, sorry, I misspoke | 19:12 |
| kbringard | each rack will be configured for all of the VLANs | 19:13 |
| vishy | kbringard: then run all of the computes in the rack with --network_host=<this racks compute> | 19:13 |
| kbringard | so any random host machine will have a br_100 br_101, br_102... | 19:13 |
| kbringard | and the VM will attach to the correct bridge depending on the /24 it's IP was pulled from | 19:14 |
| vishy | kbringard: ah ok but the bridge it needs to go into is dependent on the ip it gets? | 19:14 |
| * vishy thinks for a second | 19:14 | |
| kbringard | hehe | 19:14 |
| kbringard | basically, I'm working around a netops restriction | 19:15 |
| kbringard | logically, it's the same thing as a giant /16 or whatever | 19:15 |
| kbringard | but they want to keep broadcast domains limited | 19:15 |
| kbringard | so they trunk VLANs | 19:15 |
| uvirtbot | New bug: #706029 in nova "pluginlib_nova needs gettext" [Undecided,Fix committed] https://launchpad.net/bugs/706029 | 19:16 |
| *** reldan has joined #openstack | 19:16 | |
| vishy | kbringard: the only reason multiple networks aren't supported is get_network_host | 19:16 |
| kbringard | yea, I saw that in the nova.network.manager module documentation | 19:16 |
| kbringard | Creating multiple networks for for one manager is currently not supported, but could be added by modifying allocate_fixed_ip and get_network to get the a network with new logic instead of network_get_by_bridge. | 19:17 |
| kbringard | so I guess that is still the case then :-) | 19:17 |
| vishy | kbringard: yes i believe I thought of this use case, I just didn't know the best way to implement it | 19:17 |
| kbringard | cool, makes sense | 19:18 |
| kbringard | well, I'll ponder on it as well | 19:18 |
| kbringard | and hopefully we can come up with something :-) | 19:18 |
| vishy | kbringarrd: i think if you just change the logic to get the least full network | 19:18 |
| vishy | instead of get_by_bridge | 19:18 |
| kbringard | makes sense | 19:19 |
| vishy | i can't see any reason why that wouldn't work, although you would probably need some supporting functions to create networks with different vlans without hacking the db directly | 19:19 |
| kbringard | I've not delved into the code much yet... but it sounds like the way it works now is it just looks at the bridge and assigns an IP based on that? | 19:20 |
| vishy | kbringard: in vlan mode, each project has its own bridge | 19:20 |
| kbringard | yea, I was hoping to get away from that because I don't necessarily want to create projects for everything... that's why I was looking at FlatManager | 19:21 |
| vishy | kbringard: in the other modes there is one network for the manager, so it always returns the only network | 19:21 |
| kbringard | got it | 19:21 |
| kbringard | cool, thanks for the info, I'll dig into it and see what I come up with :-) | 19:21 |
| vishy | kbringard: FlatDhcpManager is a little easier to set up as it handles routing and such, but it does have a couple gotchas | 19:22 |
| *** nelson has joined #openstack | 19:22 | |
| nelson | Let me guess: the pseudo-directory scheme in swift is VERY much like using a prefix=, correct? | 19:23 |
| nelson | Except that there's a file which has the same name as the prefix. | 19:24 |
| kbringard | vishy: OK, thanks for the info! | 19:25 |
| creiht | nelson: are you asking from the user API perspective? | 19:26 |
| jaypipes | sirp_, jk0: OK, going to cut 0.1.3pre2 tarball for Glance. last chance for objections :) | 19:26 |
| creiht | If so, if you include your directory structure in the object names, you can do pretty effective directory like queries | 19:26 |
| jk0 | jaypipes: I'm game | 19:26 |
| vishy | kbringard: np, we are testing different networking scenarios as well, so we might run into some of the same pain points and hopefully we have good docs and fixes soon | 19:27 |
| nelson | well yes, I'm reading the API, but Im guessing that the implementation is pretty transparent to the existance of these directory names. | 19:27 |
| sirp_ | jaypipes: can't think of any off hand, but really haven't had a chance to dig in yet, really fighting hard to get nova working in DomU | 19:27 |
| creiht | the prefix= queries are just translated to queries in the listing database | 19:28 |
| *** adiantum has quit IRC | 19:29 | |
| *** kpepple has left #openstack | 19:29 | |
| jaypipes | sirp_: is this something you want me to hold off the tarball for? | 19:29 |
| mtaylor | redbo: I thought we already had a trunk eventlet package somewhere | 19:30 |
| jk0 | jaypipes: this is the only problem I've been seeing: http://paste.openstack.org/show/537/ | 19:31 |
| mtaylor | redbo: and could someone kick the eventlet guy in the pants to do a new release? | 19:31 |
| jk0 | jaypipes: but I don't think it's related | 19:31 |
| redbo | rdw: do a release | 19:31 |
| creiht | hehe | 19:32 |
| creiht | rdw: you have full permission to kick mtaylor back :) | 19:32 |
| *** pharkmillups has joined #openstack | 19:33 | |
| mtaylor | rdw: kick | 19:33 |
| *** mdomsch has quit IRC | 19:34 | |
| *** pharkmillups has quit IRC | 19:34 | |
| jk0 | jaypipes: also, fwiw http://paste.openstack.org/show/538/ | 19:34 |
| *** adiantum has joined #openstack | 19:34 | |
| jaypipes | sirp_: hmm, that's odd. anyone have any ideas on why that is happening? this only happens when nova is running in dom0? | 19:34 |
| sirp_ | jaypipes: off hand i don't, maybe i could jump in a screen session with jk0 and dig in, jk0 up for that? | 19:35 |
| jk0 | yeah, definitely | 19:35 |
| kbringard | how do I delete networks created with nova-manage network create? | 19:35 |
| kbringard | do I need to just blast it from the DB? | 19:35 |
| sirp_ | jk0: ok im gonna grab a quick bite, one do that in 30 min or so | 19:35 |
| *** GasbaKid has joined #openstack | 19:35 | |
| sirp_ | *wanna | 19:36 |
| jk0 | sirp_: sounds good | 19:36 |
| *** rossi_j has quit IRC | 19:36 | |
| *** rossij has joined #openstack | 19:36 | |
| jaypipes | jk0: everything looks fine on that second paste... | 19:37 |
| jk0 | jaypipes: the funny part is curling the image directly works fine, too | 19:37 |
| jk0 | I'm thinking it's likely a problem somewhere else, not in glance | 19:37 |
| jaypipes | jk0: hmm, ok. | 19:38 |
| *** littleidea has quit IRC | 19:38 | |
| *** retr0h has quit IRC | 19:39 | |
| *** GasbaKid has quit IRC | 19:41 | |
| nelson | Hmmmmm .... I'm thinking that whomever made the Swift API didn't consider the diff between PUT and POST in the REST paradigm. | 19:42 |
| creiht | heh... here goes this discussion again :) | 19:42 |
| kbringard | lol | 19:43 |
| redbo | oh good god... | 19:43 |
| *** adiantum has quit IRC | 19:43 | |
| *** greenisus has quit IRC | 19:44 | |
| johnpur | redbo: dirty laundry being aired! | 19:44 |
| creiht | psh | 19:44 |
| creiht | :) | 19:44 |
| *** greenisus has joined #openstack | 19:45 | |
| *** retr0h has joined #openstack | 19:45 | |
| nelson | Not criticizing ... just noting. | 19:45 |
| creiht | hehe | 19:46 |
| redbo | Now someone needs to say that having the API version in the URL isn't restful. | 19:46 |
| nelson | Oh, well fuck that idea. API version NEEDS to be in the URL or lossage ensues. | 19:46 |
| creiht | maybe there is some script that tests the restfulness of apis that we could have hudson run against our code at each checkin | 19:46 |
| creiht | :) | 19:46 |
| jaypipes | creiht: lol. | 19:47 |
| *** elasticdog has quit IRC | 19:47 | |
| creiht | >:) | 19:47 |
| openstackhudson | Project nova build #441: SUCCESS in 1 min 32 sec: http://hudson.openstack.org/job/nova/441/ | 19:47 |
| openstackhudson | Tarmac: Adds gettext to pluginlib_nova.py. Fixes #706029. | 19:47 |
| jaypipes | creiht: I still think replacing that Chuck Norris plugin with a Chuck Thier plugin is a good idea. | 19:47 |
| nelson | and anyway, I always thought that the idea of REST was that all sorts of crap was supposed to go into the URL. That's how you know exactly what service you'[re accessing. | 19:47 |
| *** omidhdl has left #openstack | 19:48 | |
| creiht | jaypipes: :) | 19:48 |
| redbo | I think swift's API is very restful. | 19:48 |
| nelson | anyway ... back to reading.... | 19:48 |
| *** mdomsch has joined #openstack | 19:48 | |
| creiht | nelson: if you are curious, feel free to ask away | 19:48 |
| *** BK_man has joined #openstack | 19:48 | |
| nelson | redbo: except that REST expects PUT for writes, and POST for creates. PUT fails if it doesn't exist; POST fails if it does. | 19:49 |
| creiht | sorry about the sirly remarks, but some of the same questions come up all the time :) | 19:49 |
| *** adiantum has joined #openstack | 19:50 | |
| redbo | I think you're making up rest concepts. | 19:50 |
| creiht | lol | 19:50 |
| redbo | "To create or modify a resource, the client sends a PUT request that usually includes an entity-body." | 19:50 |
| creiht | nelson: read this book: http://oreilly.com/catalog/9780596529260 | 19:51 |
| creiht | :) | 19:51 |
| nelson | Oh, it's possible that I'm wrong, but I read that somewhere on the Internet. | 19:51 |
| *** elasticdog has joined #openstack | 19:51 | |
| nelson | And as we all know, someone on the Internet is wrong. | 19:51 |
| creiht | hehe | 19:51 |
| spectorclan | tr3buchet: Trey, who do you work for? thanks | 19:51 |
| jk0 | if it's on the Internet, it has to be true ;) | 19:51 |
| creiht | we base most of our RESTful ideas from that book | 19:51 |
| creiht | and it is a very good book | 19:51 |
| nelson | http://xkcd.com/386/ | 19:51 |
| jk0 | spectorclan: tr3buchet is in Rackspace Ozone | 19:51 |
| spectorclan | jk0: Thanks! Much appreciated | 19:52 |
| jk0 | np | 19:52 |
| jk0 | spectorclan: he likes to go by "Sprinkles", though | 19:52 |
| *** EdwinGrubbs has joined #openstack | 19:52 | |
| spectorclan | jk0: I am not even going to touch that at all | 19:52 |
| jk0 | ;) | 19:52 |
| redbo | yes, I identify with characters in XKCD comics far too often | 19:53 |
| mtaylor | redbo: which PPA do we need new eventlet in? | 19:53 |
| redbo | I don't know, it was vishy's request. | 19:53 |
| mtaylor | I'm gonna guess nova then | 19:54 |
| vishy | mtaylor: nova | 19:55 |
| mtaylor | k | 19:55 |
| *** dprince has quit IRC | 19:56 | |
| mtaylor | vishy: and you do need trunk and not just 0.9.13? | 19:58 |
| vishy | mtaylor: yes or 9.13 with redbo's patch | 19:59 |
| spectorclan | looking for michael rodriguez -> company?? | 19:59 |
| mtaylor | redbo: do you have a patch location? | 19:59 |
| vishy | mtaylor: here - https://bitbucket.org/which_linden/eventlet/changeset/a8b259177024 | 19:59 |
| redbo | The tricky part you have to fix *both* lines of code. | 20:00 |
| creiht | nelson: and to answer your question, POST and PUT can both be used when you want to create something, POST is usually used when you are creating something that you don't know what the final URL endpoint of that object is going to be, PUT is for when you do know what the final URL is going to be | 20:00 |
| vishy | redbo: *chuckle* | 20:00 |
| BK_man | jaypipes: did you committed fix into glance's trunk | 20:04 |
| BK_man | jaypipes: ? | 20:04 |
| spectorclan | KnightHacker: company you work for? Thanks | 20:04 |
| creiht | spectorclan: KnightHacker is a cloud architect | 20:05 |
| creiht | for RS | 20:05 |
| spectorclan | creiht: thanks | 20:05 |
| *** pvo is now known as pvo_away | 20:07 | |
| *** Pentheus has joined #openstack | 20:09 | |
| *** ctennis has quit IRC | 20:11 | |
| *** dubsquared has joined #openstack | 20:16 | |
| *** gondoi has quit IRC | 20:19 | |
| *** gondoi has joined #openstack | 20:19 | |
| JordanRinke | ohai | 20:24 |
| jk0 | hiya | 20:25 |
| dabo | JordanRinke: ohaiurself | 20:26 |
| *** littleidea has joined #openstack | 20:26 | |
| *** aimon has quit IRC | 20:31 | |
| *** gondoi has quit IRC | 20:35 | |
| *** adiantum has quit IRC | 20:37 | |
| *** gondoi has joined #openstack | 20:37 | |
| *** aimon has joined #openstack | 20:37 | |
| *** mdomsch has quit IRC | 20:39 | |
| *** adiantum has joined #openstack | 20:43 | |
| *** pvo_away is now known as pvo | 20:47 | |
| notmyname | colinnich: I haven't forgotten about he domain_remap code. I'm working on it now. I think we want to change it a little | 20:48 |
| jk0 | jaypipes: turns out that conn refused problem is caused by an issue in compute | 20:49 |
| colinnich | notmyname: I've just come back after dinner.... so I'll be here for a few hours. Change what you like, I'm no expert :-) | 20:50 |
| *** reldan has quit IRC | 20:50 | |
| *** pvo is now known as pvo_away | 20:51 | |
| notmyname | colinnich: pretty simple really. I think the thought is to restrict what will be written to the list of explicit prefixes in the config. IOW, if you define reseller_prefixes, only remap those and fail for others | 20:51 |
| uvirtbot | New bug: #706072 in nova "Instance NotFound error : Instance Id is not correct from user perspective" [Undecided,New] https://launchpad.net/bugs/706072 | 20:51 |
| *** adiantum has quit IRC | 20:53 | |
| *** j05h1 has quit IRC | 20:54 | |
| colinnich | notmyname: after I'd pushed the branch, I also wondered if there would be any merit in taking a lowercase copy of account before the loop | 20:54 |
| *** pvo_away is now known as pvo | 20:54 | |
| colinnich | notmyname: or maybe one would not have enough reseller prefixes for it to matter much...? | 20:55 |
| notmyname | colinnich: ya, I'm adding some docs on that too. it's the same with container names. if you use domain_remap, be dns compatible | 20:56 |
| *** adiantum has joined #openstack | 20:57 | |
| *** j05h has joined #openstack | 20:58 | |
| *** greenisus has quit IRC | 20:58 | |
| *** greenisus has joined #openstack | 20:59 | |
| *** dfg has quit IRC | 20:59 | |
| *** greenisus has quit IRC | 20:59 | |
| *** greenisus has joined #openstack | 21:00 | |
| openstackhudson | Project swift build #171: SUCCESS in 29 sec: http://hudson.openstack.org/job/swift/171/ | 21:01 |
| openstackhudson | Tarmac: Fixes to get audit to run cleanly over datasets where people have i18n strings (ie. bit 7 set utf-8 bytes) in container and object names. | 21:01 |
| openstackhudson | Arguably these fixes aren't the most elegant or complete, but for now it suffices to get things working until we re-factor things later. | 21:01 |
| openstackhudson | Also tweak output from foo... to "foo" (so cut & paste works a little better). | 21:01 |
| *** pharkmillups has joined #openstack | 21:02 | |
| *** adiantum has quit IRC | 21:03 | |
| *** Cybodog has joined #openstack | 21:04 | |
| *** adamsih300u has quit IRC | 21:07 | |
| *** adiantum has joined #openstack | 21:08 | |
| *** dirakx has quit IRC | 21:10 | |
| *** tonywolf has quit IRC | 21:14 | |
| *** dfg has joined #openstack | 21:15 | |
| *** pvo is now known as pvo_away | 21:17 | |
| *** pvo_away is now known as pvo | 21:18 | |
| *** reldan has joined #openstack | 21:18 | |
| *** ctennis has joined #openstack | 21:18 | |
| *** reldan has quit IRC | 21:19 | |
| *** pharkmillups has quit IRC | 21:20 | |
| *** westmaas has quit IRC | 21:22 | |
| *** adiantum has quit IRC | 21:23 | |
| *** pvo is now known as pvo_away | 21:30 | |
| *** kbringard has quit IRC | 21:32 | |
| *** dendrobates is now known as dendro-afk | 21:37 | |
| *** dendro-afk is now known as dendrobates | 21:38 | |
| *** pvo_away is now known as pvo | 21:39 | |
| *** MarkAtwood has quit IRC | 21:40 | |
| *** adiantum has joined #openstack | 21:45 | |
| sirp_ | Proposed https://code.launchpad.net/~rconradharris/nova/bug705790/+merge/47116 would love comments | 21:47 |
| blueadept | is openstack POSIX compatible? | 21:52 |
| *** miclorb_ has joined #openstack | 21:53 | |
| *** bradshaw has joined #openstack | 21:55 | |
| jero | what do you mean by posix compatible blueadept ? | 21:55 |
| *** westmaas has joined #openstack | 21:55 | |
| *** jschopp-- has joined #openstack | 21:55 | |
| sirp_ | jaypipes: if you get a sec, could you give https://code.launchpad.net/~rconradharris/nova/bug705790/+merge/47116 a once over | 21:57 |
| dabo | sirp_: lines 49-50 should be mappings, e.g.: LOG.debug(_('VBD %(vbd)s plugged into wrong dev, remapping to %(dev)s') % locals()) | 21:58 |
| sirp_ | dabo: is that an official coding standard, seems rather inconsistent with the rest of nova | 22:00 |
| dabo | sirp_: yes, it's a standard | 22:01 |
| *** MarkAtwood has joined #openstack | 22:01 | |
| dabo | It will be consistent once my merge prop is merged | 22:01 |
| dabo | https://code.launchpad.net/~ed-leafe/nova/lp703041/+merge/46823 | 22:01 |
| blueadept | jero: can you mount it as a file system and perform operations on it? | 22:01 |
| dabo | I need one more core dev to ok that one (hint, hint!) | 22:02 |
| sirp_ | dabo: ah, ok, it *will* be standard, gotcha | 22:02 |
| dabo | sirp_: and it isn't an arbitrary standard. Translations can't always work when positions are fixed. | 22:03 |
| sirp_ | dabo: not arguing whether its good or bad, its just not standard in current trunk | 22:03 |
| sirp_ | dabo: when you're stuff hits, then it becomes the best practice | 22:03 |
| dabo | It's been noted as a bug in current trunk | 22:03 |
| dabo | No sense writing even more bugs | 22:04 |
| Xenith | blueadept: No, that is not what openstack does. | 22:09 |
| jero | blueadept: well, I guess you ask about swift, so it provides an rest api (think http) to access the storage elements. | 22:10 |
| jero | -n* | 22:10 |
| mtaylor | anybody in here a bit rabbitmq person? | 22:10 |
| mtaylor | s/bit/big/ | 22:12 |
| blueadept | jero: right, ok, yeah i've gone over some of the documentation that talks about the REST setup, so i get it now | 22:15 |
| eday | mtaylor: I might be able to help | 22:15 |
| mtaylor | eday: I'm trying to figure out how to launch and then stop a rabbitmq in a sandbox not as root | 22:16 |
| mtaylor | eday: I've got launching down - but there doesn't seem to be any way to inform rabbitmqctl about having launched rabbit on a different port | 22:16 |
| *** opengeard has joined #openstack | 22:16 | |
| *** westmaas has quit IRC | 22:16 | |
| eday | mtaylor: so, you may want to address it using the erlang node, -n <node> | 22:17 |
| tr3buchet | spectorclan: i do not like to go by sprinkles | 22:18 |
| eday | mtaylor: and each port you start can have testX@blah nodename | 22:18 |
| eday | mtaylor: you can use RABBITMQ_NODENAME env var too | 22:18 |
| mtaylor | yeah. I set RABBITMQ_NODENAME | 22:18 |
| jaypipes | sirp_: looking now mate :) | 22:19 |
| jaypipes | sirp_: oops, too late! | 22:19 |
| sirp_ | heh | 22:19 |
| mtaylor | eday: http://paste.drizzle.org/show/312/ | 22:19 |
| sirp_ | i pushed a new fix for the logging, wonders whether that will make it into the merge | 22:19 |
| mtaylor | eday: rabbitmqlctl seems to not grok I'm not talking about the main rabbitmq node | 22:20 |
| eday | mtaylor: remove the @camelot on the command line arg | 22:20 |
| *** adiantum has quit IRC | 22:21 | |
| mtaylor | eday: same thing | 22:21 |
| eday | hm | 22:21 |
| mtaylor | eday: I did this: http://paste.drizzle.org/show/313/ | 22:22 |
| jero | blueadept: it's possible to implement user-space filesystems that provide a posix set of features in the frontend, and talk http with the swift backend accordingly | 22:22 |
| eday | mtaylor: ahh, might be because it's out of a diff dir, and the .erlang_cookie file doesn't match | 22:23 |
| mtaylor | eday: that makes sense - so I guess it's just a bug that rabbitmqctl doesn't also respect those env vars | 22:23 |
| eday | mtaylor: actually, try appending @camelot onto the env var before starting (nodename) | 22:23 |
| mtaylor | eday: k | 22:24 |
| eday | erlang is picky about short vs long names | 22:24 |
| jk0 | sirp_: yeah, didn't make it | 22:24 |
| sirp_ | ah well | 22:25 |
| mtaylor | eday: nope. I think the telling thing is: | 22:25 |
| mtaylor | - current node home dir: /var/lib/rabbitmq | 22:25 |
| jk0 | sirp_: flipped back to needs review | 22:25 |
| mtaylor | which is very much the wrong node home dir | 22:25 |
| eday | mtaylor: that's fine, the mnesia and log files don't need to be in there | 22:25 |
| sirp_ | jaypipes: could you glance at the little logging change, and if looks good, flip it back to 'Approved'? | 22:25 |
| mtaylor | but I'm wondering if it's finding the wrong cookie because it's looking in there | 22:26 |
| eday | mtaylor: the only thing that needs to be shared is .erlang_cookie, which looks like it is | 22:26 |
| jaypipes | sirp_: done | 22:26 |
| mtaylor | eday: ok. well that's good at least | 22:26 |
| sirp_ | jaypipes: thanks! | 22:27 |
| eday | mtaylor: did you restart the server with the new nodename? | 22:27 |
| mtaylor | eday: yes | 22:27 |
| mtaylor | eday: and tried both short and long versions to -n to rabbitmqctl | 22:27 |
| mtaylor | (of course, I won't even get started with the fact that it won't let me run that command as not-root or not rabbitmq-user) | 22:27 |
| *** adiantum has joined #openstack | 22:28 | |
| nelson | creiht: okay, this is annoying. I have a container whose contents I can GET (it's empty), and yet I can't PUT it to re-create it. I get a "404 Not Found". | 22:28 |
| *** ppetraki has quit IRC | 22:28 | |
| mtaylor | eday: I may be back to "capture the pid and kill it" | 22:29 |
| nelson | API documentation doesn't list that as one of the possible error returns. | 22:29 |
| *** te has joined #openstack | 22:30 | |
| eday | mtaylor: are you starting/stopping as the same user? | 22:30 |
| eday | mtaylor: because it may be looking in your home dir for the cookie | 22:30 |
| eday | mtaylor: rather than /var/lib | 22:31 |
| eday | I just tested and I have diff cookies | 22:31 |
| *** rpedde has joined #openstack | 22:31 | |
| mtaylor | eday: ah. that's probably it - it won't let me run the command as me. lemme see if fakeroot fakes it out | 22:31 |
| *** aimon has quit IRC | 22:32 | |
| *** tonywolf has joined #openstack | 22:33 | |
| mtaylor | eday: AHA! | 22:33 |
| mtaylor | eday: /usr/sbin/rabbitmqctl is just a stupid-ass wrapper shell script | 22:34 |
| *** allsystemsarego has quit IRC | 22:34 | |
| eday | mtaylor: fixed? | 22:34 |
| mtaylor | eday: which blocks use by not root ... if I skip it and call /usr/lib/rabbitmq/bin/rabbitmqctl directly, it works | 22:34 |
| eday | cool :) | 22:34 |
| mtaylor | eday: sweet. thanks man. | 22:36 |
| eday | np | 22:36 |
| *** blueadept has quit IRC | 22:37 | |
| openstackhudson | Project nova build #442: SUCCESS in 1 min 30 sec: http://hudson.openstack.org/job/nova/442/ | 22:37 |
| openstackhudson | Tarmac: This patch adds two flags: | 22:37 |
| openstackhudson | --xenapi_remap_vbd_dev | 22:37 |
| openstackhudson | --xenapi_remap_vbd_dev_prefix | 22:37 |
| openstackhudson | If the plugged-in VBD dev is wrong, these configs let your remap it on the fly. This works around a bug in Ubuntu Maverick: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/684875 | 22:37 |
| uvirtbot | Launchpad bug 684875 in linux "Patch to Natty 2.6.37-virtual breaks non-EC2 users" [High,Fix released] | 22:37 |
| ascovel | hey guys, got a question about user roles regarding starting instances without beign an admin | 22:45 |
| uvirtbot | New bug: #706115 in nova "XenServer destroy error on halted instance" [Undecided,New] https://launchpad.net/bugs/706115 | 22:47 |
| *** maplebed has quit IRC | 22:48 | |
| *** adiantum has quit IRC | 22:50 | |
| *** rnirmal has quit IRC | 22:51 | |
| *** dfg has quit IRC | 22:51 | |
| *** maplebed has joined #openstack | 22:51 | |
| *** rcc has quit IRC | 22:56 | |
| *** adiantum has joined #openstack | 22:56 | |
| *** spectorclan has quit IRC | 22:59 | |
| *** littleidea_ has joined #openstack | 22:59 | |
| uvirtbot | New bug: #706122 in nova "openstack api should filter image types" [Undecided,New] https://launchpad.net/bugs/706122 | 23:01 |
| anticw | gholt: where is the container acl checking done? | 23:03 |
| *** littleidea has quit IRC | 23:03 | |
| *** littleidea_ is now known as littleidea | 23:03 | |
| *** adiantum has quit IRC | 23:05 | |
| *** daleolds has quit IRC | 23:05 | |
| *** bradshaw has quit IRC | 23:06 | |
| *** dfg has joined #openstack | 23:07 | |
| *** ascovel has quit IRC | 23:09 | |
| *** adiantum has joined #openstack | 23:12 | |
| *** tonywolf has left #openstack | 23:12 | |
| *** pvo is now known as pvo_away | 23:16 | |
| *** dubsquared has quit IRC | 23:22 | |
| *** ferdy has joined #openstack | 23:22 | |
| *** ferdy has quit IRC | 23:23 | |
| *** adiantum has quit IRC | 23:24 | |
| *** jschopp-- has quit IRC | 23:27 | |
| *** cyonyx has joined #openstack | 23:27 | |
| *** dendrobates is now known as dendro-afk | 23:28 | |
| *** adiantum has joined #openstack | 23:30 | |
| *** retr0h has quit IRC | 23:30 | |
| *** gondoi has quit IRC | 23:33 | |
| *** cynb has quit IRC | 23:36 | |
| *** mray has quit IRC | 23:45 | |
| *** EdwinGrubbs has quit IRC | 23:47 | |
| *** MarkAtwood has quit IRC | 23:47 | |
| *** MarkAtwood has joined #openstack | 23:48 | |
| *** johnpur has quit IRC | 23:54 | |
| gholt | anticw: Via the swift.authorize callback to middleware. The webob.Request passed back will have an acl attr on it. For a bit of example code sww swift/common/middleware/auth.py line 183 or so. | 23:57 |
| *** rpedde has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
