| *** dolphm has joined #openstack | 00:13 | |
| *** reldan has quit IRC | 00:18 | |
| *** zns has quit IRC | 00:26 | |
| *** Ryan_Lane has quit IRC | 00:38 | |
| *** RJD22 is now known as RJD22|away | 00:44 | |
| ShermanBoyd | I'm trying to install openstack compute on natty, when I apt-get install nova-compute it stalls out on "0% [Waiting for headers]" | 00:47 |
|---|---|---|
| *** ziyadb has quit IRC | 00:47 | |
| *** RJD22|away is now known as RJD22 | 00:48 | |
| julian_c | Are you able to run "apt-get update" successfully? | 00:49 |
| ShermanBoyd | yeah | 00:49 |
| ShermanBoyd | it just started up but it failed to download curl | 00:49 |
| julian_c | Downloading from a standard repo? | 00:50 |
| ShermanBoyd | yeah | 00:50 |
| ShermanBoyd | I know it's bleeding edge | 00:50 |
| ShermanBoyd | but I heard it was working | 00:51 |
| ShermanBoyd | Failed to fetch http://us.archive.ubuntu.com/ubuntu/pool/main/c/curl/curl_7.21.3-1ubuntu1_amd64.deb Connection faile d [IP: 91.189.92.169 80] | 00:52 |
| julian_c | That looks like a network connection on one end (likely not on yours). | 00:52 |
| ShermanBoyd | hmm I'm able to download that file to my workstation | 00:53 |
| *** mahadev has joined #openstack | 00:56 | |
| julian_c | Try to install from aptitude, and see if it gives the same error. | 00:56 |
| ShermanBoyd | ends with current status: 0 broken | 00:58 |
| ShermanBoyd | I uploaded the .deb file and tried to install directly | 00:58 |
| * julian_c scratches his head... | 00:59 | |
| ShermanBoyd | that worked | 00:59 |
| ShermanBoyd | it's just bizarre ... I assume I got all the other files from the same server | 00:59 |
| julian_c | Any mention of a hash sum mismatch, by chance? | 01:00 |
| ShermanBoyd | nope | 01:01 |
| ShermanBoyd | maybe I'll look at the logs | 01:01 |
| ShermanBoyd | ahh it might have been a dependency thing "libcurl3" | 01:03 |
| ShermanBoyd | is there a list of ports that openstack uses anywhere? | 01:05 |
| ShermanBoyd | I suppose enabling UFW is a bad idea | 01:05 |
| *** dolphm has quit IRC | 01:05 | |
| julian_c | On its own, it's probably not outright a bad idea. | 01:06 |
| *** dendro-afk is now known as dendrobates | 01:06 | |
| julian_c | The ports are probably mentioned somewhere in the documentation. I haven't looked at it in great detail yet. | 01:06 |
| ShermanBoyd | I see a bunch of ports open for rabbitMQ | 01:08 |
| ShermanBoyd | I wonder if those jump around | 01:08 |
| julian_c | Probably a bunch of high-number ports, going between components of OpenStack. | 01:10 |
| julian_c | Multiple boxes, or single-box setup? | 01:10 |
| ShermanBoyd | right now just one | 01:14 |
| ShermanBoyd | but I have a cluster of four | 01:14 |
| *** cascone has quit IRC | 01:15 | |
| ShermanBoyd | I guess I'll start with one controller | 01:15 |
| ShermanBoyd | and 3 compute nodes | 01:15 |
| julian_c | The networking you mentioned earlier makes perfect sense. | 01:17 |
| *** cascone has joined #openstack | 01:19 | |
| ShermanBoyd | ? | 01:19 |
| julian_c | Earlier, you mentioned having three networks for your cloud. | 01:20 |
| ShermanBoyd | yeah | 01:20 |
| ShermanBoyd | the management one is mostly for IPMI | 01:20 |
| ShermanBoyd | and cobbler deployment for the physical nodes | 01:21 |
| ShermanBoyd | I also want to keep the management one non-essential | 01:21 |
| ShermanBoyd | so if another switch fails | 01:21 |
| ShermanBoyd | I can just sacrifice the management network for a short time | 01:22 |
| ShermanBoyd | so my thought process is to open ports on the back end network | 01:22 |
| ShermanBoyd | but not on the public one | 01:22 |
| julian_c | Makes sense. | 01:24 |
| julian_c | I can't say that I've done anything with either IPMI or Cobbler. | 01:24 |
| *** rdw has joined #openstack | 01:25 | |
| ShermanBoyd | I got these supermicro servers with IPMI, so far I've just been using the KVM and remote boot features | 01:27 |
| ShermanBoyd | I'm pretty stoked, it saves you a ton, you don't have to buy a KVM or a PDU with remote power cycle | 01:28 |
| ShermanBoyd | Cobbler was a little shaky running on Ubuntu and deploying ubuntu | 01:29 |
| *** miclorb_ has joined #openstack | 01:29 | |
| *** dijenerate has quit IRC | 01:31 | |
| *** sigmonsays has joined #openstack | 01:37 | |
| ShermanBoyd | should I avoid using 127.0.0.1 in the nova.conf? | 01:38 |
| julian_c | If you're just trying things out on a single box, you should be able to get away with it. | 01:44 |
| julian_c | (wouldn't be surprised if ::1 works properly, too) | 01:44 |
| ShermanBoyd | No manual entry for nova.conf =( | 01:45 |
| ShermanBoyd | what's this? --routing_source_ip= | 01:46 |
| julian_c | Check out: <http://docs.openstack.org/cactus/openstack-compute/admin/content/setting-flags-in-nova-conf-file.html>. | 01:47 |
| ShermanBoyd | that's what I'm following now | 01:49 |
| julian_c | In the example provided, it actually has a regular static IP in all the necessary fields. | 01:49 |
| ShermanBoyd | it doesn't explain what "routing_source_ip" is | 01:50 |
| julian_c | I'm still making sense of the networking options myself. | 01:50 |
| julian_c | I'm trying to switch from Eucalyptus. | 01:51 |
| julian_c | It looks like --routing_source_ip is the address for the bridge. | 01:53 |
| *** dijenerate has joined #openstack | 01:54 | |
| ShermanBoyd | that's kind of what I thought | 01:54 |
| *** foxtrotdelta has joined #openstack | 02:00 | |
| *** foxtrotgulf has quit IRC | 02:02 | |
| sigmonsays | is there any architecture documents wrt storage? | 02:02 |
| *** mahadev has quit IRC | 02:08 | |
| *** miclorb_ has quit IRC | 02:09 | |
| *** adjohn has joined #openstack | 02:10 | |
| *** ksteward has quit IRC | 02:11 | |
| *** mahadev has joined #openstack | 02:14 | |
| gholt | sigmonsays: If you mean Swift: http://dl.dropbox.com/u/19755215/swift/swift_overview.html and http://swift.openstack.org/ | 02:29 |
| sigmonsays | That makes sense. I was going to be suprised if they built their own .. =P | 02:31 |
| ShermanBoyd | I'm trying to install Openstack using docs.openstack.org. It's getting pretty obvious that these docs are skip steps and are incomplete. | 02:37 |
| ShermanBoyd | is there some other install doc out there? | 02:38 |
| sigmonsays | ShermanBoyd, what part? | 02:38 |
| ShermanBoyd | I've followed along up to: http://docs.openstack.org/cactus/openstack-compute/admin/content/enabling-access-to-vms-on-the-compute-node.html | 02:39 |
| julian_c | Also check out: <http://cssoss.wordpress.com/>. | 02:39 |
| ShermanBoyd | but I never really did anything with the API | 02:39 |
| ShermanBoyd | so I just kind of did this: --ec2_url=http://184.106.239.134:8773/services/Cloud | 02:40 |
| ShermanBoyd | with my own ip | 02:40 |
| julian_c | The same blog did really good guides for Eucalyptus last year for Ubuntu 10.04 and 10.10. | 02:40 |
| ShermanBoyd | when I hit that url I get a 404 | 02:40 |
| ShermanBoyd | whoa that blog looks nice | 02:41 |
| julian_c | The graphic in Chapter 2 makes the whole setup make sense. | 02:42 |
| sigmonsays | man, this eucalyptus user space makes no sense to me. Where do I stop an instance? | 02:48 |
| julian_c | The Eucalyptus tools? | 02:49 |
| sigmonsays | Yah, the dev doc shows using euca-run-instances to start but now i'm stuck on stopping | 02:50 |
| julian_c | euca-terminate-instances | 02:51 |
| *** mahadev has quit IRC | 02:52 | |
| sigmonsays | doh | 02:52 |
| *** mahadev has joined #openstack | 02:54 | |
| *** CloudChris has joined #openstack | 02:56 | |
| sigmonsays | I'm using nova-manage network delete to rmove the network but it says its bound to a project. However, I've already delete the project. Must I restart the stack? | 02:56 |
| *** mahadev has quit IRC | 03:07 | |
| julian_c | I haven't gotten that far yet... | 03:08 |
| sigmonsays | no worries, i'll rtfm or re-install. bbiab | 03:10 |
| *** mahadev has joined #openstack | 03:17 | |
| *** mahadev has quit IRC | 03:22 | |
| ShermanBoyd | "--network_size=8" is this number of hosts or more like a /8 ?? | 03:24 |
| *** adjohn has quit IRC | 03:26 | |
| *** mahadev has joined #openstack | 03:27 | |
| *** adjohn has joined #openstack | 03:29 | |
| *** mahadev has quit IRC | 03:40 | |
| *** julian_c has quit IRC | 03:43 | |
| *** mahadev has joined #openstack | 03:49 | |
| *** MarcMorata has joined #openstack | 03:55 | |
| *** MarcMorata has quit IRC | 03:56 | |
| *** osier has joined #openstack | 03:56 | |
| *** AimanA has quit IRC | 04:09 | |
| *** zns has joined #openstack | 04:31 | |
| *** taihen has quit IRC | 04:31 | |
| *** iRTermite has quit IRC | 04:32 | |
| *** adjohn has quit IRC | 04:43 | |
| *** ryker has joined #openstack | 04:43 | |
| *** iRTermite has joined #openstack | 04:49 | |
| *** zns has quit IRC | 04:54 | |
| *** zns has joined #openstack | 05:00 | |
| *** rchavik has joined #openstack | 05:03 | |
| *** rchavik has joined #openstack | 05:03 | |
| *** alice has joined #openstack | 05:05 | |
| *** Tweeda has quit IRC | 05:21 | |
| *** adjohn has joined #openstack | 05:25 | |
| *** CloudChris has left #openstack | 05:31 | |
| *** chiehwen has joined #openstack | 05:50 | |
| *** zenmatt has quit IRC | 05:57 | |
| *** chiehwen has left #openstack | 06:01 | |
| *** alice has quit IRC | 06:07 | |
| *** mahadev has quit IRC | 06:13 | |
| *** jdarcy has quit IRC | 06:14 | |
| *** zns has quit IRC | 06:27 | |
| *** koolhead17 has left #openstack | 06:33 | |
| *** adjohn has quit IRC | 07:03 | |
| *** allsystemsarego has joined #openstack | 07:09 | |
| *** jdarcy has joined #openstack | 07:12 | |
| *** newopen has joined #openstack | 07:23 | |
| *** cdbs has quit IRC | 07:38 | |
| *** koolhead17 has joined #openstack | 07:39 | |
| *** cdbs has joined #openstack | 07:40 | |
| *** taihen has joined #openstack | 08:24 | |
| *** redash has joined #openstack | 08:29 | |
| *** cascone has quit IRC | 08:30 | |
| *** krish|wired-in has joined #openstack | 08:31 | |
| krish|wired-in | hi , in /usr/bin/nova-manage network create 192.168.0.0/24 1 255 | 08:32 |
| krish|wired-in | What does the 1 255 signify? | 08:32 |
| *** krish|wired-in has quit IRC | 08:51 | |
| *** lionel has quit IRC | 08:59 | |
| *** katkee has joined #openstack | 09:04 | |
| *** ryker has quit IRC | 09:05 | |
| *** brd_from_italy has joined #openstack | 09:23 | |
| *** uuser123 has joined #openstack | 09:30 | |
| *** zenmatt has joined #openstack | 09:36 | |
| *** krish|wired-in has joined #openstack | 09:37 | |
| *** cascone has joined #openstack | 09:39 | |
| *** j^ has left #openstack | 09:39 | |
| *** seb_delta has joined #openstack | 09:56 | |
| seb_delta | hi, is there anyone who could test openstack "nova" with deltacloud? | 09:58 |
| *** ejat has joined #openstack | 10:02 | |
| *** seb_delta has quit IRC | 10:04 | |
| *** ziyadb has joined #openstack | 10:15 | |
| *** ziyadb has joined #openstack | 10:15 | |
| *** katkee has quit IRC | 10:22 | |
| *** zenmatt has quit IRC | 10:24 | |
| *** zenmatt has joined #openstack | 10:30 | |
| *** ejat has quit IRC | 10:34 | |
| *** krish|wired-in has quit IRC | 10:34 | |
| *** brd_from_italy has quit IRC | 10:40 | |
| *** ejat has joined #openstack | 11:00 | |
| *** ejat has joined #openstack | 11:00 | |
| *** reldan has joined #openstack | 11:04 | |
| *** zenmatt has quit IRC | 11:20 | |
| *** cascone has quit IRC | 11:33 | |
| *** ejat has quit IRC | 11:34 | |
| *** fayce has joined #openstack | 11:39 | |
| *** cascone has joined #openstack | 11:53 | |
| *** uuser123 has quit IRC | 12:00 | |
| *** fayce has quit IRC | 12:14 | |
| *** redash has quit IRC | 12:25 | |
| *** RobertLaptop has quit IRC | 12:32 | |
| *** cascone has quit IRC | 12:33 | |
| *** Eyk^off is now known as Eyk | 12:39 | |
| *** fayce has joined #openstack | 12:42 | |
| *** fayce has quit IRC | 12:45 | |
| *** cascone has joined #openstack | 12:51 | |
| *** osier has quit IRC | 12:55 | |
| *** ziyadb_ has joined #openstack | 12:56 | |
| *** cascone has quit IRC | 12:58 | |
| *** ziyadb has quit IRC | 13:00 | |
| *** cascone has joined #openstack | 13:15 | |
| *** reldan has quit IRC | 13:15 | |
| *** cascone has quit IRC | 13:20 | |
| *** alice has joined #openstack | 13:25 | |
| *** stewart has quit IRC | 13:27 | |
| *** newopen has quit IRC | 13:28 | |
| *** stewart has joined #openstack | 13:30 | |
| *** zns has joined #openstack | 13:31 | |
| *** Eyk is now known as Eyk^off | 13:38 | |
| *** shermanboyd_ has joined #openstack | 13:42 | |
| shermanboyd_ | I'm not sure I understand what S3 is ... is this Glance? | 13:47 |
| notmyname | S3 is a commercial object store provided by Amazon. the equivalent in Openstack is swift (http://openstack.org/projects/storage/) | 13:49 |
| *** zns has quit IRC | 13:55 | |
| *** reldan has joined #openstack | 13:56 | |
| *** adjohn has joined #openstack | 14:05 | |
| *** MarkAtwood has joined #openstack | 14:08 | |
| shermanboyd_ | My end goal is store the images locally on the node where they run. I would also like copies of each image to be stored on other nodes for redundancy. | 14:15 |
| shermanboyd_ | So "--s3_host=" in nova.conf is supposed to point to a swift install right? | 14:17 |
| shermanboyd_ | I guess my question is when compute fires up an instance, where is that vm stored? Just ram? | 14:23 |
| *** JuanPerez has joined #openstack | 14:33 | |
| *** rchavik has quit IRC | 14:37 | |
| *** adjohn has quit IRC | 14:38 | |
| *** reldan has quit IRC | 14:44 | |
| *** Eyk^off is now known as Eyk | 14:45 | |
| *** sigmonsays has quit IRC | 14:51 | |
| *** newopen has joined #openstack | 15:01 | |
| *** zenmatt has joined #openstack | 15:03 | |
| *** sigmonsays has joined #openstack | 15:03 | |
| *** RobertLaptop has joined #openstack | 15:06 | |
| *** zenmatt has quit IRC | 15:09 | |
| *** RobertLaptop has quit IRC | 15:12 | |
| *** osier has joined #openstack | 15:17 | |
| *** sigmonsays has quit IRC | 15:20 | |
| *** Eyk is now known as Eyk^off | 15:27 | |
| *** arreyder has quit IRC | 15:28 | |
| *** arreyder has joined #openstack | 15:28 | |
| *** RobertLaptop has joined #openstack | 15:29 | |
| *** RobertLaptop has quit IRC | 15:34 | |
| *** romero has joined #openstack | 15:37 | |
| *** zns has joined #openstack | 15:38 | |
| *** Pathin_ has quit IRC | 15:42 | |
| *** Pathin has joined #openstack | 15:44 | |
| *** RobertLaptop has joined #openstack | 15:46 | |
| *** RobertLaptop has quit IRC | 15:51 | |
| *** osier has quit IRC | 15:53 | |
| *** ejat has joined #openstack | 15:56 | |
| *** ejat has joined #openstack | 15:56 | |
| *** Pathin has quit IRC | 15:58 | |
| *** ejat has quit IRC | 16:06 | |
| *** JuanPerez has quit IRC | 16:11 | |
| *** ziyadb_ has quit IRC | 16:14 | |
| *** ziyadb has joined #openstack | 16:18 | |
| *** RobertLaptop has joined #openstack | 16:21 | |
| *** ziyadb_ has joined #openstack | 16:21 | |
| *** ziyadb has quit IRC | 16:24 | |
| *** RobertLaptop has quit IRC | 16:26 | |
| *** zns has quit IRC | 16:31 | |
| *** ziyadb has joined #openstack | 16:32 | |
| *** ziyadb has joined #openstack | 16:32 | |
| *** uuser123 has joined #openstack | 16:32 | |
| *** londo has quit IRC | 16:33 | |
| *** ziyadb_ has quit IRC | 16:35 | |
| *** alice has quit IRC | 16:37 | |
| *** RobertLaptop has joined #openstack | 16:38 | |
| *** RobertLaptop has quit IRC | 16:42 | |
| *** vernhart has joined #openstack | 16:53 | |
| *** ziyadb has quit IRC | 16:56 | |
| *** zns has joined #openstack | 16:56 | |
| *** nelson has quit IRC | 17:04 | |
| *** nelson has joined #openstack | 17:04 | |
| *** reldan has joined #openstack | 17:06 | |
| *** MarkAtwood has quit IRC | 17:14 | |
| *** shermanboyd_ has quit IRC | 17:23 | |
| *** koolhead17 has joined #openstack | 17:39 | |
| *** obino has joined #openstack | 17:51 | |
| *** zenmatt has joined #openstack | 17:56 | |
| *** sigmonsays has joined #openstack | 17:56 | |
| *** ShermanBoyd has quit IRC | 17:59 | |
| sigmonsays | Is there any example code of python swift api I can quickly groK? | 18:04 |
| *** hggdh has quit IRC | 18:04 | |
| *** hggdh has joined #openstack | 18:05 | |
| sigmonsays | it'd be nice if they included example invocations in the pydoc | 18:06 |
| sigmonsays | is swift.common.client.Connection the wrong thing to use? | 18:09 |
| * sigmonsays thinks it's really lame to put "rackspace" in the cloudfiles name. | 18:11 | |
| sigmonsays | Ugh, it's even more lame they dont have a host argument on the connection class. | 18:14 |
| vishy | krish: it should be 1 256 | 18:15 |
| vishy | krish: 1 is number of networks and 256 is network size | 18:15 |
| vishy | so a single /24 is 1 256 | 18:15 |
| *** fenris- has joined #openstack | 18:15 | |
| *** fenris- has joined #openstack | 18:15 | |
| vishy | sigmonsays: in the future nova-manage project scrub | 18:16 |
| *** fenris- has quit IRC | 18:16 | |
| *** ejat has joined #openstack | 18:17 | |
| *** ejat has joined #openstack | 18:17 | |
| RichiH | swift defaults to using xfs (very good!), but does it have extra checksums to catch silent data corruption on top? | 18:19 |
| RichiH | i.e. zfs-esque | 18:19 |
| sigmonsays | who in their right mind hardcodes hosts into an api that's supposed to be open.. | 18:22 |
| sigmonsays | http://nopaste.info/71a54b5578_nl.html anyone ever use cloudfiles for their own setup? | 18:23 |
| sigmonsays | am I missing something here? | 18:25 |
| *** MarcMorata has joined #openstack | 18:29 | |
| * sigmonsays searches for a real API that's not stuck to rackspace | 18:29 | |
| *** MarcMorata has quit IRC | 18:30 | |
| RichiH | sigmonsays: send patch? | 18:31 |
| RichiH | (without having looked at the paste or anything else) | 18:31 |
| *** irclog_ has quit IRC | 18:42 | |
| *** irclog_ has joined #openstack | 18:43 | |
| *** uuser123 has quit IRC | 18:45 | |
| sigmonsays | oh boy, might have to. They use httplib which reads 1 byte at a time. Lovely | 18:45 |
| sigmonsays | even if it wasn't limited and badly designed it also performs poorly | 18:45 |
| gholt | sigmonsays: Cloud Files existed before Open Stack, hence why python-cloudfiles is 'stuck' to Rackspace. | 18:46 |
| gholt | sigmonsays: For my own stuff, I use swift.common.client from the swift code base. | 18:46 |
| gholt | sigmonsays: And, for that, there's an authurl that you point to ... whatever. | 18:47 |
| *** sophiap has joined #openstack | 18:48 | |
| sigmonsays | /usr/bin/st copy n pasted the api into its file.. just seems unstable and also uses httplib w/ performs poorly | 18:57 |
| gholt | sigmonsays: Yes, that is so it can be distributed independently of everything, as it states in the comment. | 18:57 |
| gholt | And why do you feel httplib reads only 1 byte at time? | 18:58 |
| *** Eyk^off is now known as Eyk | 18:58 | |
| sigmonsays | i've seen it, probably only w/ http keep alives if you know it doesn't | 18:59 |
| *** ziyadb has joined #openstack | 19:03 | |
| *** ziyadb has joined #openstack | 19:03 | |
| gholt | It could just be an older version of httplib, not sure. I haven't noticed any problems with Python 2.6 and httplib usage, excepting a bit in their header reading. One of our guys wrote swift.common.bufferedhttp to work around that. | 19:07 |
| sigmonsays | rgr. I am on cactus and am getting a really strange experience with calling get_account() twice. It works fine once then dies | 19:13 |
| *** Eyk is now known as Eyk^off | 19:13 | |
| *** stewart has quit IRC | 19:13 | |
| *** AimanA has joined #openstack | 19:15 | |
| sigmonsays | http://nopaste.dk/p3352 Is this odd? It's like the return value is being set in the class instance for subsequent calls.... | 19:22 |
| * sigmonsays gives up for the day | 19:26 | |
| gholt | That is odd, which codebase of Swift are you using? | 19:26 |
| sigmonsays | cactus | 19:27 |
| gholt | I'll try with that. It worked okay from trunk, but that's not horribly useful to you I'd suspect. | 19:27 |
| sigmonsays | well.. actually not entirely sure. using the ppa. deb http://ppa.launchpad.net/swift-core/ppa/ubuntu lucid main --- If that's cactus. Assume it is cause that's what compute was using also | 19:28 |
| sigmonsays | probably just butcher the code from /usr/bin/st then | 19:28 |
| sigmonsays | not now tho, gotta run | 19:29 |
| gholt | Gah. That's /still/ 1.2 (Bexar) after all this time. | 19:29 |
| sigmonsays | doh | 19:29 |
| * gholt looks for who to bang over the head. | 19:29 | |
| * gholt takes the easy way out and pings the PTL. | 19:30 | |
| gholt | notmyname: The Swift packages still seem to be stuck at 1.2. | 19:30 |
| *** brd_from_italy has joined #openstack | 19:38 | |
| *** ShermanBoyd has joined #openstack | 19:44 | |
| *** koolhead17 has quit IRC | 20:01 | |
| *** koolhead17 has joined #openstack | 20:02 | |
| *** monaDeveloper has joined #openstack | 20:07 | |
| monaDeveloper | hi there | 20:07 |
| *** AimanA has quit IRC | 20:08 | |
| monaDeveloper | I just wanted your advice how to start my way into open stack | 20:08 |
| sigmonsays | monaDeveloper, dive in using http://wiki.openstack.org/NovaInstall/DevInstallScript -- i'm up and running in dev w/ both compute and storage in 1 days time | 20:12 |
| sigmonsays | less than really, this morning to be precise =P but now it's afternoon so been at it a while.. | 20:12 |
| monaDeveloper | oh great thank you so much | 20:12 |
| ShermanBoyd | if only the docs were that good | 20:14 |
| ShermanBoyd | =\ | 20:14 |
| sigmonsays | they are imo =P | 20:16 |
| ShermanBoyd | you can't follow the docs and arrive at a working cloud | 20:16 |
| sigmonsays | not blindly no | 20:17 |
| ShermanBoyd | or even with your eyes open | 20:17 |
| sigmonsays | even if there were docs or a script called setup-cloud.sh for one environment, at the end you'd have no idea what you just did and no idea how to fix or debug any of it... | 20:18 |
| ShermanBoyd | well that's where you are right? | 20:18 |
| ShermanBoyd | you just run a script | 20:18 |
| ShermanBoyd | that's why I'm trying to do it with docs | 20:19 |
| ShermanBoyd | so I can understand everything | 20:19 |
| *** AimanA has joined #openstack | 20:21 | |
| notmyname | gholt: *sigh* on it | 20:25 |
| *** AimanA has quit IRC | 20:27 | |
| *** stewart has joined #openstack | 20:29 | |
| *** katkee has joined #openstack | 20:35 | |
| *** monaDeveloper has quit IRC | 20:38 | |
| *** hggdh has quit IRC | 20:40 | |
| *** hggdh has joined #openstack | 20:40 | |
| *** zns has quit IRC | 20:43 | |
| *** stewart has quit IRC | 20:49 | |
| *** joearnold has joined #openstack | 20:53 | |
| *** AimanA has joined #openstack | 21:01 | |
| *** joearnold has quit IRC | 21:03 | |
| *** AimanA has quit IRC | 21:03 | |
| *** AimanA has joined #openstack | 21:06 | |
| *** JuanPerez has joined #openstack | 21:07 | |
| *** joearnold has joined #openstack | 21:07 | |
| *** arun_ has joined #openstack | 21:16 | |
| *** allsystemsarego has quit IRC | 21:19 | |
| *** katkee has quit IRC | 21:20 | |
| *** JuanPerez has quit IRC | 21:34 | |
| *** joearnold has quit IRC | 21:36 | |
| *** katkee has joined #openstack | 21:41 | |
| *** ziyadb_ has joined #openstack | 21:48 | |
| *** ziyadb has quit IRC | 21:51 | |
| *** brd_from_italy has quit IRC | 21:59 | |
| *** ziyadb has joined #openstack | 22:03 | |
| *** ziyadb has joined #openstack | 22:03 | |
| *** ziyadb_ has quit IRC | 22:04 | |
| *** mahadev has joined #openstack | 22:08 | |
| *** katkee has quit IRC | 22:12 | |
| ShermanBoyd | ahh now I find wiki.openstack.org | 22:13 |
| *** ziyadb_ has joined #openstack | 22:16 | |
| *** mahadev has quit IRC | 22:18 | |
| *** ziyadb has quit IRC | 22:19 | |
| *** ziyadb_ has quit IRC | 22:25 | |
| *** jgb has joined #openstack | 22:27 | |
| sigmonsays | is it possible to use a object server on a box w/o xattr supported file system? | 22:30 |
| *** zeph1ro has joined #openstack | 22:32 | |
| *** ejat has quit IRC | 22:32 | |
| zeph1ro | hi all, I'm digging into the code and architecture of nova, but I can' find in the documentation a high level description of the orchestration of events... how's a machine provisioned? is it an atomic chain of function calls? or an event is created in a queue and picked up? and how's the allocation of the VM decided? (means, on which hw machine gets pushed? and based on what?) | 22:34 |
| *** thevoke has joined #openstack | 22:38 | |
| *** joearnold has joined #openstack | 22:41 | |
| *** julian_c has joined #openstack | 22:44 | |
| *** joearnold has quit IRC | 22:45 | |
| *** zns has joined #openstack | 22:56 | |
| *** irclog_ has quit IRC | 22:57 | |
| *** ovidwu_ has quit IRC | 22:57 | |
| *** zeph1ro has quit IRC | 23:05 | |
| *** ryker has joined #openstack | 23:09 | |
| notmyname | sigmonsays: no. xattr is required | 23:10 |
| sigmonsays | Is that a strict requirement? I would be curious of seeing that limitation removed | 23:13 |
| notmyname | it's a fairly hard rule. all of the object metadata is stored in xattrs (content-length, content-size, last-modified time, all the other headers that can be set, etc) | 23:14 |
| *** gungadin has joined #openstack | 23:16 | |
| *** RobertLaptop has joined #openstack | 23:16 | |
| notmyname | theoretically, it could be refactored to store that in another file as a sibling to the object, but that would require more disk hits to read the additional file, and we really haven't seen a use case for systems that don't support xattrs | 23:16 |
| gungadin | hi | 23:16 |
| notmyname | but if you have a good, compelling use case, we'd love to hear it. alternatively, patches are always welcome | 23:17 |
| sigmonsays | limitations within my infrastructure do not allow me to enable xattrs, not w/o a lot of prior work | 23:20 |
| *** gungadin has left #openstack | 23:21 | |
| *** cloudnod has joined #openstack | 23:22 | |
| *** cloudnod has left #openstack | 23:22 | |
| *** reldan has quit IRC | 23:22 | |
| *** cloudnod has joined #openstack | 23:23 | |
| *** stewart has joined #openstack | 23:25 | |
| notmyname | sigmonsays: using xattrs in the system greatly simplifies the code and allows us to use the underlying file system for some of the elements of scale. honestly, we'd need a very compelling reason to consider not using xattrs | 23:28 |
| *** zenmatt has quit IRC | 23:29 | |
| sigmonsays | Yep, i'm cool with that, although a patch which pads a object file seems reasonable too (in absence of xattr) | 23:31 |
| sigmonsays | still researching, so it's not a big deal yet =P | 23:32 |
| cloudnod | so um... if I want to contribute and don't yet know much python... other than "learn more python" what advice might you give me? :) | 23:35 |
| *** stewart has quit IRC | 23:37 | |
| *** rchavik has joined #openstack | 23:42 | |
| notmyname | cloudnod: read over the code of whatever project you are interested in. the swift code is probably the most complete right now (with failure handling, production ready, etc). try to see if you can understand what is happening in the code | 23:43 |
| notmyname | follow the mailing list to see what's happening in the projects and openstack in general | 23:43 |
| notmyname | and ask questions | 23:43 |
| notmyname | I think the basic rule is to have no fear. jump in with both feet. find something small that you want to change, and work on a patch. if you hop over to #openstack-dev, I'm sure there will be people who could help you work on a patch | 23:44 |
| sigmonsays | Im trying to find out node failures are recovered from, how nodes rebalance (if at all) and how objects are balanced. Lots of details specific to the system design are missing. Is there a doc i'm missing? | 23:50 |
| notmyname | sigmonsays: http://swift.openstack.org/overview_architecture.html http://swift.openstack.org/overview_ring.html | 23:55 |
| notmyname | look specifically at the ring | 23:55 |
| notmyname | sigmonsays: I'd also recommend some of the articles at http://tlohg.com/tag/openstack | 23:56 |
| sigmonsays | thanks, i'll take a look | 23:57 |
| *** obino has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!