Ryan_Lane | Is there are reason why all the auth_unittest tests are being skipped when I run the tests? are they not supposed to be running, or do I have to do something to make them run? | 00:04 |
---|---|---|
eday | they run for me | 00:05 |
Ryan_Lane | I wonder what I'm doing wrong | 00:06 |
Ryan_Lane | it's specifically skipping the ones I want to run the most :) | 00:06 |
eday | heh | 00:07 |
eday | they commented out in run_tests.py for some reason? | 00:08 |
Ryan_Lane | nope | 00:08 |
openstackhudson | Project nova build #294: SUCCESS in 1 min 15 sec: http://hudson.openstack.org/job/nova/294/ | 00:09 |
*** Podilarius has left #openstack | 00:09 | |
*** dfg_ has quit IRC | 00:13 | |
Ryan_Lane | it's doing the DB auth tests, and not the LDAP ones | 00:13 |
*** bfschott has quit IRC | 00:24 | |
*** aryko has joined #openstack | 00:24 | |
Ryan_Lane | ah. it'll only do the ldap tests with redis. | 00:26 |
*** dubsquared has left #openstack | 00:29 | |
*** kashyapc has joined #openstack | 00:36 | |
* Ryan_Lane sighs | 00:40 | |
Ryan_Lane | the fake ldap driver doesn't implement wildcard searches | 00:40 |
*** joearnold has quit IRC | 00:42 | |
*** joearnold has joined #openstack | 00:43 | |
eday | its fake! | 00:43 |
Ryan_Lane | I know ;) | 00:44 |
Ryan_Lane | but I'm now using wildcard searches in the ldap driver | 00:44 |
*** joearnold has quit IRC | 00:44 | |
Ryan_Lane | so all the damn tests are failing :( | 00:44 |
eday | ahh | 00:47 |
Ryan_Lane | wildcard searches are needed for searches that check for attribute existence | 00:47 |
*** BK_man has joined #openstack | 00:47 | |
Ryan_Lane | such as "(owner=*)" which will match for objects that have the attribute, but will not for attributes that do not | 00:48 |
Ryan_Lane | I think I can fake it in the fake driver :) | 00:49 |
KyleM1 | I'm back again! | 00:50 |
KyleM1 | I tried using the script that alekibango pointed me at but had a couple of issues: | 00:50 |
KyleM1 | 1) Try as I might I couldn't find any references to PPA or VENV variables to disable | 00:51 |
*** rlucio has quit IRC | 00:51 | |
KyleM1 | 2) I went ahead with the branch and install commands anyway but when I tried to run it my SSH died and I couldn't reconnect to the server. | 00:51 |
KyleM1 | Does anyone know how I can prevent that from happening? | 00:52 |
alekibango | KyleM1: he changed the code, uh | 00:52 |
alekibango | dnt worry about venv, that is prolly gone for good | 00:53 |
KyleM1 | Hooray? | 00:53 |
alekibango | but sudo add-apt-repository ppa:nova-core/trunk will fail on debian | 00:53 |
alekibango | imho | 00:53 |
KyleM1 | Alright. | 00:53 |
alekibango | waht i did with that script - was placing set -e on top of it | 00:54 |
KyleM1 | What do you suggest changing it to? | 00:54 |
alekibango | set -xv | 00:54 |
alekibango | KyleM1: you are on squeeze right? | 00:54 |
KyleM1 | I believe so, yes. | 00:54 |
KyleM1 | Yep | 00:54 |
*** BK_man has quit IRC | 00:54 | |
*** BK_man has joined #openstack | 00:54 | |
KyleM1 | # cat /etc/debian_version | 00:54 |
KyleM1 | squeeze/sid | 00:54 |
*** dendro-afk is now known as dendrobates | 00:58 | |
* Ryan_Lane cheers | 01:01 | |
Ryan_Lane | that was easy enough to add :) | 01:01 |
*** robrightscale has left #openstack | 01:03 | |
*** jc_smith has quit IRC | 01:04 | |
alekibango | Ryan_Lane: ? | 01:04 |
Ryan_Lane | wildcard searches to the fake ldap driver | 01:04 |
alekibango | ah :) | 01:04 |
alekibango | python is easy | 01:04 |
Ryan_Lane | of course I just implemented them as all or nothing | 01:04 |
alekibango | :D | 01:04 |
* alekibango blames python | 01:05 | |
alekibango | it should be harder | 01:05 |
Ryan_Lane | since that's all I need :) | 01:05 |
alekibango | :D | 01:05 |
Ryan_Lane | heh | 01:05 |
Ryan_Lane | it should be php? | 01:05 |
Ryan_Lane | ;) | 01:05 |
*** jbryce_ has joined #openstack | 01:07 | |
*** jbryce has quit IRC | 01:11 | |
*** jbryce_ is now known as jbryce | 01:11 | |
KyleM1 | Okay, it looks like the "clean-vlans" command is what's killing my network; is there something I can do to prevent this from dropping the interfaces, or should I just comment it out? | 01:13 |
winston-d | What I did is comment it out... | 01:17 |
*** sirp1 has joined #openstack | 01:17 | |
KyleM1 | Okay. | 01:17 |
KyleM1 | Simple enough! | 01:17 |
*** littleidea has quit IRC | 01:18 | |
*** dragondm has quit IRC | 01:24 | |
*** BK_man_ has joined #openstack | 01:25 | |
*** BK_man has quit IRC | 01:25 | |
*** BK_man_ is now known as BK_man | 01:25 | |
*** zhengli has quit IRC | 01:26 | |
termie | eday: https://code.launchpad.net/~termie/nova/eventlet_merge is pretty much there, all the tests are passing currently (except ignoring objectstore at the moment) | 01:27 |
*** Ryan_Lane is now known as Ryan_Lane|away | 01:27 | |
termie | eday: still needs a variety of things to be cleaned up | 01:27 |
*** Ryan_Lane|away has quit IRC | 01:27 | |
termie | eday: and i don't _really_ trust the tests enough to say everything works | 01:28 |
termie | eday: but the end is squarely in sight | 01:28 |
*** dovetaildan has quit IRC | 01:33 | |
*** dovetaildan has joined #openstack | 01:35 | |
*** sirp1 has quit IRC | 01:36 | |
*** dendrobates is now known as dendro-afk | 01:41 | |
*** sophiap has joined #openstack | 01:42 | |
*** johnpur has quit IRC | 01:42 | |
*** mdomsch has joined #openstack | 01:55 | |
*** daleolds has quit IRC | 02:01 | |
*** ctennis has quit IRC | 02:08 | |
*** miclorb has quit IRC | 02:09 | |
*** KyleM1 has quit IRC | 02:13 | |
*** dovetaildan has quit IRC | 02:14 | |
*** bfschott has joined #openstack | 02:15 | |
*** dovetaildan has joined #openstack | 02:16 | |
*** johnbergoon has joined #openstack | 02:20 | |
*** aimon has quit IRC | 02:20 | |
*** aimon has joined #openstack | 02:20 | |
*** BK_man_ has joined #openstack | 02:25 | |
*** BK_man has quit IRC | 02:25 | |
*** BK_man_ is now known as BK_man | 02:25 | |
*** kashyapc has quit IRC | 02:30 | |
*** jimbaker has joined #openstack | 02:33 | |
*** johnbergoon has quit IRC | 02:33 | |
*** hazmat has quit IRC | 02:40 | |
*** BK_man has quit IRC | 02:46 | |
*** dendro-afk is now known as dendrobates | 02:48 | |
*** kashyapc has joined #openstack | 03:01 | |
*** maplebed has joined #openstack | 03:09 | |
*** sophiap has quit IRC | 03:23 | |
*** jdurgin has quit IRC | 03:28 | |
*** johnbergoon has joined #openstack | 03:29 | |
*** jimbaker has quit IRC | 03:41 | |
*** bfschott has quit IRC | 03:41 | |
*** maplebed has quit IRC | 03:53 | |
*** sirp1 has joined #openstack | 04:14 | |
*** joearnold has joined #openstack | 04:15 | |
*** sirp1 has joined #openstack | 04:15 | |
*** dendrobates is now known as dendro-afk | 04:22 | |
*** novahelp has joined #openstack | 04:28 | |
*** johnbergoon has left #openstack | 04:30 | |
*** jaypipes has quit IRC | 04:31 | |
*** dubsquared has joined #openstack | 04:40 | |
uvirtbot | New bug: #687661 in glance "Glance servers crash when using epoll" [High,New] https://launchpad.net/bugs/687661 | 04:41 |
*** omidhdl has joined #openstack | 04:42 | |
*** maplebed has joined #openstack | 04:43 | |
*** jimbaker has joined #openstack | 04:48 | |
*** ctennis has joined #openstack | 04:50 | |
*** maplebed has quit IRC | 04:51 | |
uvirtbot | New bug: #687665 in glance "Swift Backend URL Parser breaks for 2.6.5" [High,New] https://launchpad.net/bugs/687665 | 04:56 |
*** winston-d has quit IRC | 04:57 | |
*** mdomsch has quit IRC | 05:12 | |
*** johnbergoon has joined #openstack | 05:24 | |
*** joearnold has quit IRC | 05:25 | |
*** johnbergoon has left #openstack | 05:26 | |
*** joearnold has joined #openstack | 05:27 | |
*** ramkrsna has joined #openstack | 05:29 | |
*** jimbaker has quit IRC | 05:51 | |
openstackhudson | Project nova-tarmac build #42,003: FAILURE in 13 sec: http://hudson.openstack.org/job/nova-tarmac/42003/ | 05:58 |
*** eldarnugaev has joined #openstack | 05:58 | |
*** ArdRigh has joined #openstack | 06:02 | |
openstackhudson | Yippie, build fixed! | 06:03 |
openstackhudson | Project nova-tarmac build #42,004: FIXED in 3.7 sec: http://hudson.openstack.org/job/nova-tarmac/42004/ | 06:03 |
*** soosfarm has quit IRC | 06:10 | |
*** GasbaKid has joined #openstack | 06:13 | |
*** GasbaKid has quit IRC | 06:17 | |
*** GasbaKid has joined #openstack | 06:18 | |
*** sirp1 has quit IRC | 06:26 | |
*** novaoss has joined #openstack | 06:34 | |
*** eldarnugaev has quit IRC | 06:43 | |
*** almaisan-away is now known as al-maisan | 06:43 | |
*** winston-d has joined #openstack | 06:43 | |
*** ctennis has quit IRC | 07:02 | |
*** ibarrera has joined #openstack | 07:08 | |
*** dubsquared has quit IRC | 07:16 | |
*** GasbaKid has quit IRC | 07:22 | |
*** schisamo has quit IRC | 07:38 | |
*** joearnold has quit IRC | 07:45 | |
*** larstobi has quit IRC | 07:55 | |
*** brd_from_italy has joined #openstack | 08:02 | |
*** guigui has joined #openstack | 08:04 | |
*** befreax has joined #openstack | 08:14 | |
*** calavera has joined #openstack | 08:15 | |
*** larstobi has joined #openstack | 08:19 | |
*** eldarnugaev has joined #openstack | 08:34 | |
*** GasbaKid has joined #openstack | 08:42 | |
*** eldarnugaev has quit IRC | 08:57 | |
*** irahgel1 has joined #openstack | 09:22 | |
*** deshantm_cosi has quit IRC | 09:32 | |
*** deshantm has quit IRC | 09:32 | |
*** miclorb_ has joined #openstack | 09:38 | |
*** eldarnugaev has joined #openstack | 09:41 | |
*** allsystemsarego has joined #openstack | 09:42 | |
*** deshantm has joined #openstack | 09:45 | |
*** deshantm_cosi has joined #openstack | 09:45 | |
*** novahelp has quit IRC | 09:47 | |
*** novaoss has quit IRC | 09:50 | |
*** GasbaKid has quit IRC | 09:50 | |
*** ramkrsna has quit IRC | 09:54 | |
*** zhengli has joined #openstack | 09:56 | |
*** eldarnugaev has left #openstack | 10:00 | |
*** reldan has joined #openstack | 10:01 | |
*** ramkrsna has joined #openstack | 10:07 | |
*** miclorb_ has quit IRC | 10:15 | |
chmouel | anybody know how to make a bit prettier the copy and paste in launchpad bugs (ie: having a tags doing a <pre> arounds or <tt> tags around) | 10:17 |
*** ArdRigh has quit IRC | 10:19 | |
soren | chmouel: Huh? | 10:21 |
chmouel | soren: have you used stackoverflow for example ? | 10:28 |
chmouel | they have this blockquote tag which make code snippets or copy and paste looks nice in html | 10:29 |
chmouel | (or at least nicer) | 10:29 |
*** befreax has quit IRC | 10:41 | |
*** stewart has quit IRC | 10:44 | |
*** sophiap has joined #openstack | 10:44 | |
uvirtbot | New bug: #687942 in glance "Missing authorization in Glance API" [Undecided,New] https://launchpad.net/bugs/687942 | 10:46 |
uvirtbot | New bug: #687944 in glance "missing get-image-size method in Teller client" [Undecided,New] https://launchpad.net/bugs/687944 | 10:56 |
*** stewart has joined #openstack | 10:57 | |
soren | chmouel: Everything on Launchpad is monospace. Code should look fine. | 10:57 |
soren | Err.. | 10:57 |
soren | chmouel: Not "everything", but all comments are monospace by default. | 10:57 |
chmouel | ok.. would be nice to have some kind of syntax highlight for code snippets tho but I guess it's an another discussion for #launchpad | 10:59 |
*** kashyapc has quit IRC | 11:02 | |
uvirtbot | New bug: #687949 in glance "Creating image with same name does not come back with error" [Undecided,New] https://launchpad.net/bugs/687949 | 11:06 |
chmouel | soren: is it my setup or is it normal if nova-api does not dump logs into /var/log/nova/ (using the daily package from this morning)? | 11:17 |
*** arthurc has joined #openstack | 11:17 | |
vivekbabs | chmouel: i'm seeing this too; there's no nova-api.log in /var/log/nova. other logs are present | 11:20 |
chmouel | it's the only process using eventlet and not twisted ... | 11:25 |
chmouel | seems like time to fill a bug :) | 11:26 |
larstobi | I'm having a problem with swift. It stopped working, and I can't find out what went wrong. It authenticates fine, but when fetching the X-Storage-Url it crashes. The stack trace received is here: http://paste.openstack.org/show/261/ | 11:27 |
uvirtbot | New bug: #687963 in nova "nova-api debian packages does not create /var/log/nova/nova-api.log" [Undecided,New] https://launchpad.net/bugs/687963 | 11:35 |
soren | chmouel: Yes. | 11:45 |
soren | chmouel: It's something I need to fix, but it's expected for now. | 11:45 |
chmouel | soren: cool, I know you are busy I can get a try on a patch if you like.. | 11:46 |
soren | It's pretty simple, really. | 11:46 |
chmouel | i thought they would be some fancy python 'logging' library stuff if its init script thing then yeah that should be trival | 11:47 |
*** rchavik has joined #openstack | 11:49 | |
soren | chmouel: I followed up on the bug explaining how to fix it. | 11:49 |
*** novahelp has joined #openstack | 11:51 | |
chmouel | is flags files are merged from different config files ? | 11:51 |
chmouel | ie: if i have a global nova.conf and another nova-api.conf | 11:51 |
chmouel | would it use only the nova-api.conf or parse as well the nova.conf | 11:52 |
soren | chmouel: At the moment, we explicitly point the various daemons at the right flagfile (from the init script). | 11:52 |
chmouel | soren: i heard flags file would be deprecated in the future, isnt it ? | 11:53 |
soren | chmouel: Eventually, yes. | 11:55 |
sandywalsh | o/ | 11:59 |
soren | 'ello. | 12:02 |
*** hazmat has joined #openstack | 12:07 | |
*** znake has joined #openstack | 12:13 | |
*** znake has left #openstack | 12:19 | |
*** sebastianstadil has joined #openstack | 12:20 | |
*** omidhdl has left #openstack | 12:21 | |
sebastianstadil | Hey everyone, is there any public Nova to use for testing? Like Eucalyptus Community Cloud. | 12:21 |
sebastianstadil | We'd like to finish adding support for it in Scalr | 12:21 |
alekibango | someone talked about this... but i am not sure how it ended. there sure should be some public nova for tests, i would say even few of them (stable, latest) | 12:22 |
alekibango | sebastianstadil: try asking 2-4 hours later when americans will be awake | 12:23 |
*** sophiap has quit IRC | 12:24 | |
*** sophiap has joined #openstack | 12:30 | |
soren | If anyone cares, natty ppa builds of nova are working again. | 12:36 |
alekibango | i do, thanks for info | 12:36 |
soren | sebastianstadil: I don't think there is yet. I'm trying to get some resources to set one up, but I don't know the status of it. | 12:37 |
*** irahgel2 has joined #openstack | 12:37 | |
*** irahgel1 has quit IRC | 12:38 | |
sebastianstadil | Could we get a dozen dedicated servers from Rackspace and expense them? | 12:40 |
sebastianstadil | Actually, what's the recommended minimum number of servers? | 12:41 |
soren | 1. | 12:41 |
soren | :) | 12:41 |
sebastianstadil | That should be cheaper | 12:41 |
alekibango | hmm those can even be virtual machines | 12:44 |
*** vivekbabs has quit IRC | 12:45 | |
*** vivekbabs has joined #openstack | 12:51 | |
*** fabiand has joined #openstack | 12:56 | |
*** arcane has quit IRC | 12:57 | |
vivekbabs | i am using nova compute build from ppa (rev 452) and it now contains xenapi changes from xenapi-refactoring branch, but i'm hitting a possible bug. this is after i have side-stepped the 'no network found for bridge br100' exception that sandywalsh said is a known issue. | 13:00 |
vivekbabs | i started getting "XenAPI.Failure: Xen-API failure: ['FIELD_TYPE_ERROR', 'self']" | 13:00 |
vivekbabs | there seems to be some problem in get_xenapi_host() in xenapi_conn.py when it calls get_this_host(self._session.handle) | 13:01 |
vivekbabs | it started working when i replaced 'self._session.handle' with 'self._session.getSession()' | 13:01 |
*** wiliam has joined #openstack | 13:02 | |
*** iRTermite has joined #openstack | 13:04 | |
*** ramkrsna has quit IRC | 13:09 | |
*** aryko has left #openstack | 13:13 | |
*** Cybodog has joined #openstack | 13:15 | |
*** Cybodog has quit IRC | 13:21 | |
chmouel | is nova needs a ramdisk to boot ? i get a ari-11111 not found (which is the default image name for flag) when i try to boot without ramdisk | 13:23 |
*** soosfarm has joined #openstack | 13:24 | |
ttx | chmouel: that's probably https://bugs.launchpad.net/nova/+bug/661159 | 13:25 |
uvirtbot | Launchpad bug 661159 in nova "Does not allow booting without a ramdisk" [Medium,In progress] | 13:25 |
chmouel | ttx: tks buddy, was going to fire up a bug report already 8-) | 13:26 |
ttx | np | 13:26 |
chmouel | and when i try to crean an instance via euca with a wrong instance type i am getting 'NameError: global name 'exception' is not defined' is that already reported? | 13:28 |
chmouel | or i could just check the bug report :) | 13:28 |
soren | Nenova/compute/instance_types.py:40: undefined name 'exception' | 13:32 |
soren | Whoopes. | 13:32 |
soren | nova/compute/instance_types.py:40: undefined name 'exception' | 13:32 |
soren | From current trunk. | 13:32 |
chmouel | yeah figured out we just need an import in there isnt it ? | 13:33 |
*** ar1 has quit IRC | 13:34 | |
soren | Dunno. Haven't looked. | 13:35 |
soren | Yup. | 13:35 |
chmouel | cheers would provide a quick patch | 13:36 |
*** irahgel2 has quit IRC | 13:39 | |
sandywalsh | Hey guys ... anything holding up getting https://code.launchpad.net/~sandy-walsh/nova/admin-only-api merged into trunk? There are two approvals and I need it for xs-pause/xs-suspend | 13:41 |
*** kashyapc has joined #openstack | 13:42 | |
*** irahgel1 has joined #openstack | 13:45 | |
uvirtbot | New bug: #688032 in nova "creating a new instance via eucatools with wrong instance type get the error "global name 'exception' is not defined"" [Undecided,New] https://launchpad.net/bugs/688032 | 13:46 |
*** Adiantum has joined #openstack | 13:46 | |
*** sebastianstadil has quit IRC | 13:48 | |
*** larstobi has quit IRC | 13:49 | |
*** hggdh has quit IRC | 13:54 | |
*** hggdh has joined #openstack | 13:55 | |
*** soosfarm has quit IRC | 13:56 | |
*** hggdh has quit IRC | 13:58 | |
*** ppetraki has joined #openstack | 14:05 | |
*** soosfarm has joined #openstack | 14:06 | |
soren | sandywalsh: Not at all. Just approved it. | 14:08 |
*** alekibango_ has joined #openstack | 14:11 | |
*** alekibango_ has quit IRC | 14:11 | |
sandywalsh | thanks S! | 14:11 |
*** reldan has quit IRC | 14:14 | |
*** jimbaker has joined #openstack | 14:15 | |
*** murkk has quit IRC | 14:18 | |
*** hggdh has joined #openstack | 14:19 | |
*** rchavik has quit IRC | 14:21 | |
sandywalsh | soren, I've fixed the merge conflict and re-pushed. Do I need to re-issue the merge prop? | 14:25 |
soren | Nope. | 14:27 |
soren | sandywalsh: Re-approved. | 14:27 |
sandywalsh | soren, re-thank-you | 14:28 |
*** johnbergoon has joined #openstack | 14:34 | |
soren | sandywalsh: Merged succesfully. | 14:34 |
openstackhudson | Project nova build #295: SUCCESS in 1 min 12 sec: http://hudson.openstack.org/job/nova/295/ | 14:34 |
sandywalsh | soren, groovy. | 14:34 |
*** gondoi has joined #openstack | 14:41 | |
*** gondoi has joined #openstack | 14:43 | |
*** ctennis has joined #openstack | 14:44 | |
sandywalsh | how can I give someone else push permissions to my branch? | 14:45 |
*** calavera has quit IRC | 14:46 | |
*** calavera has joined #openstack | 14:48 | |
creiht | sandywalsh: Not sure about that, people can make merge proposals to your branch | 14:49 |
*** jfluhmann has quit IRC | 14:49 | |
sandywalsh | creiht, hmm, I was trying to avoid that. We're pairing on a bp and I just want to grant full access. | 14:50 |
*** jfluhmann has joined #openstack | 14:50 | |
soren | sandywalsh: No. | 14:50 |
soren | sandywalsh: If there's a team you're both a member of, you can make it owned by that team. | 14:51 |
sandywalsh | hmm, ok ... we'll ping pong between two branches then. thx! | 14:52 |
*** Podilarius has joined #openstack | 14:53 | |
soren | sandywalsh: Sure. | 14:57 |
*** sebastianstadil has joined #openstack | 14:57 | |
*** Hua has joined #openstack | 14:57 | |
*** skrusty has joined #openstack | 14:58 | |
*** sebastianstadil has quit IRC | 14:58 | |
*** daleolds has joined #openstack | 14:59 | |
*** f4m8 is now known as f4m8_ | 15:00 | |
skrusty | can i ask a really stupid question? Is Compute an alternative to things like hyper-v? Does it handle running VMs or is it a management platform? | 15:00 |
*** jfluhmann has quit IRC | 15:01 | |
*** reldan has joined #openstack | 15:02 | |
*** jfluhmann has joined #openstack | 15:02 | |
vishy | skrusty: management platform would be closer | 15:04 |
*** dubsquared has joined #openstack | 15:04 | |
vishy | skrusty: compute uses a hypervisor underneath | 15:04 |
vishy | skrusty: like hyper-v/xen/kvm | 15:05 |
sophiap | I hit bug 661159, how do I unregister an image? | 15:10 |
uvirtbot | Launchpad bug 661159 in nova "Does not allow booting without a ramdisk" [High,In progress] https://launchpad.net/bugs/661159 | 15:10 |
soren | euca-deregister | 15:10 |
*** reldan has quit IRC | 15:11 | |
*** reldan has joined #openstack | 15:11 | |
sophiap | and emi is the image-id? | 15:15 |
skrusty | can it be used with windows hyper-v server? | 15:15 |
skrusty | (the free one) | 15:15 |
soren | sophiap: Yes. | 15:16 |
sophiap | thanks soren | 15:16 |
dubsquared | soren: did you have a chance to revisit the python/nova package problem in the ppa? | 15:17 |
dubsquared | i haven't tried since got in, just getting around this AM :D | 15:17 |
soren | dubsquared: I can't reproduce your problem. | 15:17 |
soren | dubsquared: It works well for me. | 15:17 |
dubsquared | soren: interesting, ill check it out, thanks! | 15:18 |
soren | dubsquared: I didn't quite understand the pastebin you threw me. What command was that from? | 15:18 |
soren | http://paste.openstack.org/show/260/ | 15:18 |
dubsquared | apt-get install -y nova-common | 15:18 |
dubsquared | oh that one.. | 15:19 |
soren | apt-get's output does not look like that. | 15:19 |
dubsquared | well, thats my script, running an apt-get | 15:19 |
soren | Can I see it, please? | 15:19 |
dubsquared | absolutely | 15:19 |
dubsquared | just got rang into a con call, ill brb, sorry! | 15:21 |
dubsquared | conf* | 15:21 |
*** stewart has quit IRC | 15:22 | |
soren | np | 15:22 |
*** sebastianstadil has joined #openstack | 15:23 | |
*** jdarcy has quit IRC | 15:25 | |
*** jdarcy has joined #openstack | 15:25 | |
*** stewart has joined #openstack | 15:34 | |
*** rnirmal has joined #openstack | 15:34 | |
sophiap | is there way to get the log for an instance that is starting up? My instance is hungup on pending | 15:39 |
*** ChanServ sets mode: +o _cerberus_ | 15:40 | |
*** dfg_ has joined #openstack | 15:41 | |
*** jaypipes has joined #openstack | 15:42 | |
ttx | _cerberus_: see :) | 15:42 |
_cerberus_ | ttx: indeed. No longer necessary though | 15:43 |
*** ChanServ sets mode: -o _cerberus_ | 15:44 | |
ttx | jaypipes: ping me when you have 10 min | 15:44 |
*** gundlach has joined #openstack | 15:45 | |
*** schisamo has joined #openstack | 15:45 | |
*** johnbergoon has left #openstack | 15:55 | |
*** sirp1 has joined #openstack | 15:55 | |
*** Hua has quit IRC | 15:57 | |
*** dragondm has joined #openstack | 16:01 | |
*** dfg_ has quit IRC | 16:02 | |
sophiap | ttx: I'm following you article, but my instance never starts, it just stays pending. Is there a log I can look at? | 16:07 |
ttx | sophiap: what kind of image do you try to start ? | 16:07 |
ttx | sophiap: the ttylinux one ? | 16:07 |
ttx | or an ubuntu one ? | 16:08 |
sophiap | ttx: yes the ttylinux | 16:08 |
sophiap | the i686 arch | 16:08 |
ttx | You're running the tutorial on a physical machine or inside a VM ? | 16:08 |
sophiap | physical machine | 16:09 |
ttx | anything suspicious in /var/log/nova/* ? | 16:09 |
ttx | eday: could you ping me if you have 5 min. over the next hour ? | 16:10 |
sophiap | this looks suspicious: libvirtError: internal error no supported architecture for os type 'hvm' | 16:11 |
ttx | indeed | 16:11 |
ttx | what does "kvm-ok" return ? | 16:12 |
sophiap | Your CPU does not support KVM extensions | 16:12 |
sophiap | dang | 16:12 |
ttx | sophiap: sometimes it's just disabled in the BIOS | 16:12 |
ttx | sometimes you just bought the wrong Intel CPU. | 16:13 |
sophiap | I think it's the latter, it's an old IBM xseries 335 | 16:13 |
chmouel | you can check with the command kvm-ok by the way if it works | 16:13 |
chmouel | ttx: you may want to add that to your article... | 16:13 |
ttx | chmouel: I *do* mention that the machine needs to have VT extensions, but that might not be enough :) | 16:14 |
chmouel | hehe :) | 16:14 |
*** daleolds has quit IRC | 16:14 | |
*** johnpur has joined #openstack | 16:15 | |
*** ChanServ sets mode: +v johnpur | 16:15 | |
*** rds__ has joined #openstack | 16:16 | |
*** maplebed has joined #openstack | 16:17 | |
*** jfluhmann has quit IRC | 16:22 | |
*** ctennis has quit IRC | 16:25 | |
*** jfluhmann has joined #openstack | 16:26 | |
eday | ttx: ping | 16:28 |
czajkowski | soren: congrats re fosdem | 16:28 |
ttx | eday: was looking at your nova specs and was wondering which, if any, we should target for bexar | 16:28 |
ttx | eday: anything you are working on or plan to deliver by Jan 6 ? | 16:29 |
*** zhengli has quit IRC | 16:29 | |
ttx | bexar-shared-ip-groups, bexar-openstack-api-floating-ips... | 16:29 |
ttx | bexar-distributed-scheduler... | 16:30 |
johnpur | ttx: do we have a spec or target for bexar that indicates that all implemented nova functionality be exposed through the openstack api? ie i don't need to go to the ec2 api (unless i want to)? | 16:33 |
*** ctennis has joined #openstack | 16:33 | |
eday | ttx: I converted the share-ip-groups/floating-ip bps since it was previously a bug.. I can work on it if they're high priority, but I'm mainly working on dist-sched | 16:33 |
eday | ttx: for dist-sched, I'm most likely not going to get it all done by jan6, and instead will have a number of steps done (I have 5 or 6 listed in the spec) | 16:34 |
*** kashyapc has quit IRC | 16:34 | |
eday | so, we'll need a cactus-dist-sched pointing to the same spec | 16:34 |
ttx | eday: ok, so I'll keep them out of bexar for now | 16:34 |
ttx | ah hm | 16:34 |
eday | I could break it into more bps if we really wanted, but seemed like extra work | 16:35 |
ttx | eday: sounds good. Unfortunately, LP prevents you from pointing two BP to the same URL. But there are ways around that :) | 16:35 |
eday | ttx: yeah, I found that out :) | 16:36 |
ttx | eday: two options (based on how much the intermediary results at bexar time make sense): two BPs and deliver one at Bexar... or just consider it work in progress and deliver at Cactus | 16:36 |
*** novahelp has quit IRC | 16:36 | |
*** maplebed has quit IRC | 16:36 | |
*** sebastianstadil has quit IRC | 16:37 | |
soren | czajkowski: Thanks! Should be fun! | 16:37 |
czajkowski | soren: yup, just submitted a lightning talk, but going no matter what | 16:37 |
soren | ttx: You could do that, too. | 16:37 |
ttx | johnpur: still need to catch up with dendrobates on that | 16:37 |
soren | ttx: A lightning talk, that is. | 16:37 |
czajkowski | ttx: http://fosdem.org/2011/call_for_lightningtalks | 16:38 |
johnpur | ttx: ok. we are running out of time tho | 16:38 |
ttx | soren: on what ? I was thinking of bitching about Java in the java devroom | 16:38 |
czajkowski | ttx: don't hold back :) | 16:38 |
eday | ttx: I'll leave it to LP release experts (ie, you) to determine the best method :) Just let me know | 16:38 |
soren | ttx: Sounds like a party :) | 16:38 |
*** guigui has quit IRC | 16:39 | |
alekibango | ttx: get back alive and in one piece! | 16:39 |
spectorclan | czajkowski: please let me know times of your speaking slot, etc so I can promote to wider audience | 16:39 |
ttx | eday: hm. do you already know what will be in bexar, or is it "as much as I can" ? | 16:40 |
* soren wanders off for dinner | 16:40 | |
sirp1 | ttx: if it's alright with you, i'd like to split the xs-snapshots blueprint into a couple of smaller sub-blueprints (e.g. a separate one for implementing the backup scheduler) | 16:41 |
ttx | johnpur: I doubt such a spec would be fully implemented for bexar, since we'll rely on some features that would appear only in Cactus | 16:41 |
ttx | johnpur: but I agree on the need to track that completion | 16:41 |
ttx | johnpur: just unsure how much of that work was already done by dendrobates | 16:42 |
ttx | or pvo. | 16:42 |
*** dragondm has quit IRC | 16:42 | |
ttx | sirp1: works for me. Just keep the main one and create another that depends on the first one completion | 16:42 |
eday | ttx: hard to say where the line will be drawn exactly, lots of unknowns in the rathole still. I have a few things that WILL be done, and a few things I'm pretty sure won't, but in between is up in the air | 16:43 |
ttx | sirp1: would the backup scheduler one still be targeted for bexar ? At the same priority ? | 16:43 |
ttx | eday: I'd go the "partial implementation in Bexar" route and just target the whole spec delivery to cactus | 16:43 |
johnpur | ttx: sure, we can only expose in the API what exists in the release. we need to be evolving the API and documenting the dev guide as we go. i mentioned some of this to annegentle yesterday. | 16:43 |
*** calavera has quit IRC | 16:44 | |
alekibango | ttx: btw.. about java... does java license still require you to delete all java installations if sun tells you to? (this might be nice rant) | 16:44 |
ttx | alekibango: I don't need more reasons for ranting, I already have my bag full | 16:45 |
alekibango | :) | 16:45 |
alekibango | ttx: dont forget about my spec | 16:45 |
ttx | Will ping you tomorrow. | 16:45 |
alekibango | i would like to push it forward, and i am willing to implement it -- if someone will help me a bit... | 16:45 |
ttx | alekibango: you're next on my big list of pings | 16:45 |
alekibango | k, will wait | 16:46 |
*** hggdh has quit IRC | 16:46 | |
sirp1 | ttx: for now, i would think that nova-backup-scheduler would be lower priority (if something has to slip, this would be a good candidate), but we could keep it on the schedule for bexar | 16:47 |
ttx | ok, will prioritize it as Low. Just subscribe me to it when you created it | 16:48 |
*** hggdh has joined #openstack | 16:48 | |
*** kashyapc has joined #openstack | 16:50 | |
*** miclorb has joined #openstack | 16:52 | |
*** vivekbabs_ has joined #openstack | 16:55 | |
*** dragondm has joined #openstack | 16:59 | |
*** hggdh has quit IRC | 17:01 | |
*** reldan has quit IRC | 17:01 | |
pvo | ttx the bexar blueprints for ozone should all be claimed | 17:02 |
ttx | pvo: cool thanks | 17:03 |
*** hggdh has joined #openstack | 17:05 | |
vivekbabs_ | a question to people running nova with XenServer, has anoyone seen "XenAPI.Failure: Xen-API failure: ['FIELD_TYPE_ERROR', 'self']" while running an instance? this is revision 452. | 17:06 |
*** kashyapc has quit IRC | 17:06 | |
*** al-maisan is now known as almaisan-away | 17:07 | |
*** rlucio has joined #openstack | 17:08 | |
*** fitzdsl has joined #openstack | 17:09 | |
*** mjmac has left #openstack | 17:09 | |
fitzdsl | Hello | 17:14 |
*** aliguori has quit IRC | 17:15 | |
uvirtbot | New bug: #688151 in nova "Describe volumes in the ec2 api does not filter results by instance_id" [Undecided,New] https://launchpad.net/bugs/688151 | 17:16 |
fitzdsl | I'd like to know if it's possible to use iSCSI target with Object Store ? | 17:16 |
*** zhengli has joined #openstack | 17:17 | |
*** daleolds has joined #openstack | 17:20 | |
*** joearnold has joined #openstack | 17:22 | |
jk0 | can I get someone on nova-core to look at this for me please? https://code.launchpad.net/~jk0/nova/diagnostics-per-instance/+merge/43132 | 17:22 |
jaypipes | jk0: looking... | 17:23 |
jk0 | thanks man | 17:23 |
jaypipes | anyone know armando's IRC nick? | 17:24 |
jk0 | good question.. I wonder if he's in here | 17:24 |
jk0 | he's in the UK, I know that | 17:24 |
dabo | jk0: in vm_utils.py you have: from xml.dom.minidom import parseString. Isn't the standard for openstack that we only import modules, not functions? | 17:26 |
jk0 | they are already importing defer right above that | 17:27 |
jk0 | unless defer is a module (?) | 17:28 |
dabo | yes, it's a module | 17:28 |
jk0 | ok, I'll change | 17:28 |
dabo | jk0: the only other thing I see (stylistically) is in vmops.py, you are raising an Exception, instead of a specific type of Exception. | 17:30 |
jk0 | that's the trend from Citrix | 17:30 |
dabo | Not knowing the code, it looks like a ValueError or a RuntimeError would probably be a better choice | 17:31 |
jk0 | I'm just following what Citrix is doing for now | 17:31 |
dabo | Well, if they're catching Exception, it'll catch Value/RuntimeError too, since they are subclasses of Exception | 17:32 |
jk0 | pushed import change | 17:32 |
* jk0 can't speak on why they decided to do it that way | 17:33 | |
pvo | jk0: you might ask if any citrix-ies are in #openstack | 17:33 |
pvo | citrix-ites | 17:33 |
jk0 | we are in #openstack :) | 17:33 |
pvo | ha | 17:33 |
pvo | er | 17:33 |
pvo | so... yea | 17:33 |
dabo | pvo: I like "citrixies" | 17:34 |
comstud | LOL | 17:34 |
* pvo needs to check his tabs more often | 17:34 | |
*** mdomsch has joined #openstack | 17:40 | |
*** joearnold has quit IRC | 17:40 | |
uvirtbot | New bug: #688160 in glance "Parallax image index returns empty erroneously" [Undecided,New] https://launchpad.net/bugs/688160 | 17:41 |
*** brd_from_italy has quit IRC | 17:41 | |
*** zhengli has quit IRC | 17:44 | |
jaypipes | sirp1: so, was it just that the image was not marked public? | 17:53 |
sirp1 | hmm good question, haven't debugged it yet... | 17:54 |
sirp1 | lemme check | 17:54 |
jk0 | jaypipes: thanks for the review | 17:57 |
jaypipes | jk0: no worries :) | 17:57 |
sirp1 | jaypipes: that was it, need to update my tests/test_data script | 17:59 |
*** hggdh has quit IRC | 18:00 | |
*** Ryan_Lane has joined #openstack | 18:00 | |
*** MarkAtwood has quit IRC | 18:01 | |
*** hggdh has joined #openstack | 18:02 | |
*** sophiap has quit IRC | 18:03 | |
*** joearnold has joined #openstack | 18:08 | |
*** daleolds1 has joined #openstack | 18:08 | |
*** daleolds has quit IRC | 18:09 | |
*** KyleM1 has joined #openstack | 18:12 | |
*** reldan has joined #openstack | 18:13 | |
*** skrusty has quit IRC | 18:14 | |
*** skrusty has joined #openstack | 18:14 | |
*** zhengli has joined #openstack | 18:15 | |
*** guynaor has joined #openstack | 18:16 | |
*** guynaor has left #openstack | 18:16 | |
*** daleolds1 has quit IRC | 18:18 | |
*** daleolds has joined #openstack | 18:18 | |
eday | vishy: so, how would you feel about removing 'internal_id' from the instance table, and making 'id', which is a 32bit auto_increment now, a 64bit random int? could also make it a UUID. this also means the ec2 names would be a bit larger | 18:19 |
alekibango | +1 | 18:19 |
vishy | seems fine to me | 18:20 |
vishy | seemed like sqlalchemy + 64 is a little tricky | 18:20 |
eday | you think we should go 64 or 128 bit (128 being a proper uuid()). uiud also means a less than ideal primary key type (binary or varchar) | 18:20 |
*** vivekbabs_ has quit IRC | 18:21 | |
eday | really? I thought BitInteger just worked for sqlalchemy | 18:21 |
eday | err, Big | 18:21 |
vishy | don't know, perhaps it does | 18:21 |
alekibango | i think it should work | 18:21 |
alekibango | if it doesnt, we should bug alchemy people | 18:21 |
eday | at least it's listed in it's core generic types | 18:21 |
vishy | ah, k | 18:21 |
eday | so, thoughts on uuid vs 64bit? we can make a 64bit int pretty random and probably won't collide, but you never know | 18:22 |
vishy | i'm not a db expert but i understand nonint primary keys is a no-no | 18:23 |
eday | I was thinking 32bit epoch + 20bit microsecond + 12bit rand() | 18:23 |
eday | vishy: it's painful for large tables yes, but they can be done | 18:23 |
alekibango | uuid is 128 bits | 18:24 |
*** sebastianstadil has joined #openstack | 18:24 | |
eday | alekibango: yeah, so we'dneed a varchar or a binary type, if it's supported by sqlalchemy | 18:24 |
alekibango | i am not sure about support ,will check | 18:24 |
eday | alekibango: we'd need to use string | 18:26 |
eday | alekibango: no generic types for binary (just large binary, but we don't want blob overhead, if any) | 18:26 |
alekibango | or binary uuid.. why not 128bit value? | 18:26 |
alekibango | ic | 18:26 |
eday | I don't see any types that would work with that. we can do string(16) and pack it | 18:27 |
alekibango | you are prolly right | 18:27 |
alekibango | 2^64 is pretty big number ;) | 18:27 |
eday | but, as vishy said, string(16) PKs are not ideal | 18:27 |
alekibango | that would eat resources... right | 18:28 |
eday | it's not so much that we'll ever fill the space, it's just if it's big enough to avoid collisions when distributed nodes are generating them independently | 18:29 |
alekibango | so lets compute probability | 18:29 |
eday | if we do the time-based value I suggested above, that's 4k random per microsecond | 18:30 |
eday | probably good enough. what UUID does is put in host-specific portion (mac address usually) so you only need to worry about host collisions | 18:31 |
alekibango | do we really need so big epoch? | 18:31 |
alekibango | hmm but for longer run... | 18:31 |
eday | we could do 31 bits, but that doesn't help | 18:32 |
alekibango | or we can partition the space | 18:32 |
alekibango | so we will not use random :) | 18:32 |
eday | we won't know all hosts ahead of time to partition, at least not in the current model | 18:33 |
alekibango | or we could use 2 columns together as a PK | 18:33 |
eday | or we could do the central ID assignment, and have each thing creating ID's pull blocks from it, but that is a bit of extra work (new service, ...) | 18:33 |
alekibango | (heh) | 18:33 |
pvo | probalby a dumb idea but we make their ipv6 address? 128bit, unique and re-useable? | 18:34 |
alekibango | central is problem | 18:34 |
alekibango | we want to go async | 18:34 |
pvo | ok | 18:34 |
alekibango | ah - pulling blocks ahead - that could be nice | 18:34 |
alekibango | sort of dynamic partitioning | 18:35 |
eday | pvo: that would work, but do we want to make the ID the ipv6 address? can't really change ips once we do that (which may be fine) | 18:35 |
alekibango | but debugging this :) might be hell | 18:35 |
alekibango | eday: what happens in case of conflict> | 18:35 |
alekibango | :? | 18:35 |
eday | alekibango: for which option? | 18:36 |
alekibango | when you will use random | 18:36 |
alekibango | and there will be conflict | 18:36 |
eday | last update wins I guess :) | 18:36 |
eday | unless we add logic to detect the case and handle it differently | 18:36 |
alekibango | yes thats my question, is that possible/?/ | 18:37 |
pvo | eday: I can see the case going either way... we don't remove primary but we do recycle. with ipv6 we wouldn't recycle. could use the customer ipv6 prefix as a filter | 18:37 |
alekibango | i would like to vote for dynamic partitioning, but it might be hell to debug if there will be problem | 18:37 |
eday | sure, unless the rest of the row data has the EXACT same payload, which would be unlikely. also this means having to read before every insert/update | 18:37 |
alekibango | (not only for devs, but for admins as well) | 18:38 |
pvo | caveat, haven't thought this through... just thinking out loud | 18:38 |
eday | also, dynamic partitioning requires a central authority == SPoF. prefer not to go that route | 18:38 |
alekibango | hmm right | 18:38 |
alekibango | unless its cooperative, peer to peer solution | 18:39 |
alekibango | hehe (complexity over roof) | 18:39 |
eday | pvo: if we were ipv6 ready i'd say yes.. but to start now, there is too much not done | 18:39 |
alekibango | whatever we do, this should be able to run for a week without sync... | 18:40 |
alekibango | and then to sync and continue operation without problem | 18:40 |
pvo | eday: well... not to use ipv6 as the networking... but we could use the scheme to provision ids and just use those when we shift. | 18:40 |
eday | alekibango: there are projects out there that handle this, ie: http://engineering.twitter.com/2010/06/announcing-snowflake.html | 18:40 |
eday | alekibango: don't really want to write our own, or make a dependency like that :) | 18:40 |
alekibango | eday: i can see why :D | 18:41 |
eday | pvo: well, we would need ipv6 provisioning in there, even without the networking | 18:42 |
eday | pvo: I suppose we can just pass arbitrary ranges, but if we have customer prefixes and other such logic, thats a bit more | 18:42 |
* eday doesn't know enough about how it will work | 18:42 | |
alekibango | maybe using uuid is not that bad idea after all | 18:43 |
alekibango | as it can contain the host | 18:43 |
pvo | I understand it to be a customer woudl get a prefix on signup, then vms spun would get the mac added to the prefix + some bits to get the ipv6 addresses. the + bits could be an random | 18:43 |
eday | well, if we were going uuid, ipv6 has a lot better context for the same size :) | 18:43 |
alekibango | so collision will never happen | 18:43 |
pvo | prefix + bits + mac | 18:43 |
pvo | eday: rigjt | 18:44 |
pvo | right | 18:44 |
eday | bits == random, or something we'd need to manage? | 18:44 |
pvo | still need to think it all teh way through, but if we're looking at a 128 bit unique field... seems wrong to not consider it | 18:44 |
eday | I suppose if we just made it a 128bit rand, we can switch to ipv6 values once we know what we're doing | 18:45 |
pvo | if the mac is 48 bits and the prefix is 64 bits, just need the extra part to be random, I'd think | 18:45 |
eday | for now do ipv4 IP + rand :) | 18:45 |
eday | also, from a security perspective, I wonder if these id's should have info like IPs, timestamps, ... had to say where they will be seen | 18:46 |
eday | hard | 18:46 |
alekibango | you can separate uuid into 2 bigInts... | 18:49 |
pvo | eday: agree that might be questionable.... but I don't know thats showstopper. just need to think it all the way though. fe80226bbfffe1da120 or fe80::226:bbff:fe1d:a120 | 18:49 |
pvo | I think we'd need the ":" | 18:49 |
*** jc_smith has joined #openstack | 18:49 | |
eday | pvo: we'd pack it into 128bits | 18:49 |
eday | string(16) | 18:50 |
pvo | right, right | 18:50 |
eday | vishy: any thoughts on using ipv6 address for primary key in db? | 18:51 |
alekibango | we can even generate that ipv6 address now even if we will not use it for network | 18:51 |
eday | I just feel like something will bite us in the ass with it, just not sure what yet | 18:51 |
alekibango | we will give you 'doomsayer award' when that happens... and it will, if we will rely on randomness of random... | 18:55 |
alekibango | or on the connectivity of the network | 18:56 |
eday | as long as we make the random space large enough so that I have a better chance of premature death (lightning, plane crash, ...), nothing to worry about :) | 18:58 |
*** daleolds has quit IRC | 19:00 | |
alekibango | that means it should as probable as dying by terrorist action... | 19:00 |
alekibango | lol | 19:00 |
*** aliguori has joined #openstack | 19:01 | |
alekibango | and they still are spending so much to fight it.. i would rather fight doctors - who kill many more people every year | 19:01 |
*** MarkAtwood has joined #openstack | 19:01 | |
alekibango | 4096 for microsecond? | 19:03 |
alekibango | if we do 1 server every second, how long it might take to get bingo | 19:04 |
eday | MarkAtwood: hey, you might have an opinion on this... what would think about using an ipv6 address as the primary key in a database containing all virtual machines for a cluster? don't need a central 'auto_increment' this way or worry about collisions | 19:04 |
eday | alekibango: rand(4096) for every microsecond | 19:04 |
eday | 20 bits are for microsecond value | 19:04 |
alekibango | so what is probabily of getting collision in one year? :) | 19:08 |
alekibango | or 100 years rather | 19:08 |
eday | depends on how many hosts/cores are generating ids concurrently | 19:08 |
alekibango | lets say 10 ids per second | 19:08 |
alekibango | or 10000 to make it tough | 19:09 |
eday | can't be done on a single core, the ops to put together one of these ids > 1 microsecond | 19:09 |
alekibango | eday: imagine cloud sized 10^6 computers | 19:10 |
eday | well, we can't average here, we need to think of spikes | 19:11 |
alekibango | yes right | 19:11 |
alekibango | so what those spikes might be? on 10^6 sized cloud? | 19:11 |
eday | well, if ID generation is in the API servers, thats less | 19:12 |
uvirtbot | New bug: #688193 in nova "Running nova-scheduler results in a traceback." [Undecided,New] https://launchpad.net/bugs/688193 | 19:12 |
alekibango | that might be 10^4 right? | 19:12 |
alekibango | i am just trying to get the number, as i feel it should be enumerated somehow - where the limit of the 64bit approach is | 19:13 |
eday | probably not that many, but sure | 19:13 |
*** BK_man has joined #openstack | 19:13 | |
alekibango | how many ids per second those hosts can create? | 19:13 |
alekibango | :) | 19:13 |
eday | let me test :) | 19:14 |
alekibango | or per microsecond | 19:14 |
MarkAtwood | eday: i like that idea, but i would like to spend some time with interested people and a whiteboard to work it out | 19:16 |
MarkAtwood | and bluntly, while i like the idea, and would love that to be implemented, i dont want hecklers who havnt read the IPv6 RFCs there when we're working the whiteboard | 19:17 |
alekibango | MarkAtwood: i agree, that makes me excluded.. :) | 19:18 |
MarkAtwood | as long as you dont heckle | 19:18 |
MarkAtwood | ^_^ | 19:18 |
uvirtbot | MarkAtwood: Error: "_^" is not a valid command. | 19:18 |
MarkAtwood | :) | 19:18 |
alekibango | hehe | 19:18 |
BK_man | hi all! what's wrong with my objectstore? http://paste.openstack.org/show/264/ | 19:18 |
MarkAtwood | ive been thinking for a few years now about how IPv6's routing classes would work well with cloud computing | 19:19 |
MarkAtwood | there are the well known link local, site local, and global route addresses | 19:19 |
MarkAtwood | corresponding to IPv4 169.*, 1918, and routable addresses | 19:19 |
eday | alekibango: 200k/sec on a 2.4 ghz core | 19:20 |
* alekibango is reading heckler definition in dict | 19:20 | |
MarkAtwood | but there are also system definable route classes, which would map well to "all instances in this rack/shelf" "all instances in this application" "all instances owned by this user" "all instances in this availablity zone" etc | 19:20 |
alekibango | eday: ty | 19:21 |
alekibango | that means 0.2 per microsec -> which means ~ 1:20.000 to get collision in every microsecond during spikes | 19:23 |
alekibango | which is not that great -> would require some care if that happens | 19:23 |
*** BK_man has quit IRC | 19:23 | |
*** BK_man_ has joined #openstack | 19:23 | |
alekibango | maximim potential is cca 100k collisions per year | 19:24 |
alekibango | if we will just do generate ids all the time | 19:24 |
alekibango | i think thats a number we can use for discussion | 19:24 |
*** maplebed has joined #openstack | 19:25 | |
alekibango | imagine talking to 100k customers who got server hackled | 19:25 |
alekibango | heckled i mean | 19:26 |
*** maple_bed has joined #openstack | 19:26 | |
alekibango | by another user unintentionally | 19:26 |
alekibango | i think it will be much lower number, but still 1:20k every microsecond during peak.. | 19:27 |
MarkAtwood | ive often wondered about using the in CPU crypto stuff for generating high speed unique IDs | 19:29 |
MarkAtwood | but ive never seen it exposed at the OS level | 19:29 |
alekibango | uuids can use network mac + time | 19:29 |
*** maplebed has quit IRC | 19:30 | |
eday | so | 19:33 |
eday | I'm thinking we should go 128 bit to be safe. for now we can just use uuid/rand, but could switch to ipv6 if it makes sense later on | 19:34 |
MarkAtwood | the advantage of ipv6 is that its a known thing, and will be useful for connecting to machines with just a primary key lookup | 19:34 |
MarkAtwood | ^_^ | 19:34 |
uvirtbot | MarkAtwood: Error: "_^" is not a valid command. | 19:34 |
MarkAtwood | ^_^ | 19:34 |
*** reldan has quit IRC | 19:35 | |
eday | yeah | 19:35 |
*** mjmac has joined #openstack | 19:35 | |
eday | the ec2 ids are going to get quite a bit larger now :) | 19:35 |
* eday wonders if this will break any apis | 19:36 | |
alekibango | MarkAtwood: if that will not become some kind of security problem | 19:36 |
*** jfluhmann has quit IRC | 19:36 | |
*** jfluhmann has joined #openstack | 19:36 | |
alekibango | 128bits are not directly supported by SA.. :( | 19:38 |
creiht | 64K^Hbits should be enough for anyone | 19:39 |
MarkAtwood | if knowing something's IP address is a security problem, then someone is depending on obscurity for security, which is a fundamental mistake | 19:39 |
sirp1 | jaypipes: i'm also working on a fix or the circular ref, i'lll use my patch to get by for the time being, then we can compare fixes when you're all wrapped | 19:40 |
*** Ryan_Lane is now known as Ryan_Lane|food | 19:40 | |
alekibango | MarkAtwood: while i agree that relying on it for security of the server is bad idea, its sometimes not only about servers and services, but also about people... and relations | 19:41 |
*** schisamo has quit IRC | 19:41 | |
MarkAtwood | the same data would exist in some other mapping table of machines to IPs | 19:42 |
*** sebastianstadil has quit IRC | 19:42 | |
alekibango | i mean, somtimes you have to mail about this | 19:42 |
alekibango | its associating you to a server directly for readers of the mail | 19:42 |
alekibango | having number protects privacy a bit | 19:43 |
*** joearnold has quit IRC | 19:43 | |
alekibango | its not always an issue, but might be for some | 19:43 |
alekibango | take julian assagne fro example... | 19:43 |
alekibango | assange is it spelled right. sometimes i think he just made his name by removing L from the end. | 19:45 |
*** Joseph has joined #openstack | 19:45 | |
*** Joseph is now known as Guest17514 | 19:46 | |
*** abecc has joined #openstack | 19:46 | |
alekibango | wikileaks had problems from the start -- they wanted to hide identity, to keep privacy... (i am not sure if they are nice guys or not, taking them just as example) | 19:47 |
jaypipes | sirp1: sure, sounds good | 19:49 |
*** jinwoosuh has joined #openstack | 19:52 | |
*** Guest17514 has left #openstack | 19:53 | |
*** daleolds has joined #openstack | 19:53 | |
*** BK_man has joined #openstack | 19:53 | |
*** BK_man_ has quit IRC | 19:53 | |
BK_man | gents, you MUST have twisted in pip-requires. An 8.2.0 version is not OK! 10.1.0 work ok. Should I create a bug for this? | 19:54 |
*** joearnold has joined #openstack | 19:56 | |
*** ibarrera has quit IRC | 19:56 | |
*** EdwinGrubbs is now known as Edwin-afk2 | 19:57 | |
MarkAtwood | that IS a good point | 19:58 |
MarkAtwood | :/ | 19:58 |
alekibango | BK_man: bug report, not another bug :) and if possible, also branch the code and make it yourself and do merge request :) | 19:58 |
*** jinwoosuh is now known as Suh | 19:58 | |
alekibango | MarkAtwood: you mean my privacy one? | 19:58 |
*** Suh has quit IRC | 19:59 | |
BK_man | alekibango: ok. Do you have a guide for newbies - how to create a branch, etc using bzr? I only use git and cvs before :-) | 19:59 |
alekibango | BK_man: will give you url with tutorial | 19:59 |
KyleM1 | BK_man: try this: http://wiki.openstack.org/LifeWithBzrAndLaunchpad | 20:00 |
alekibango | yes that one | 20:00 |
alekibango | by this you might also become known as openstack coauthor :D | 20:01 |
MarkAtwood | im of two minds... ive spend the past couple of weeks trying to walk some cops and a court clerk thru the process of getting a sender IP from gmail, and then mapping that IP to a subscriber at an ISP | 20:02 |
MarkAtwood | and i miss the days when i could just look it all up from Received and whois | 20:02 |
MarkAtwood | but then, there is the wikileaks issue, which i think is muchly a demonstration of some flaws in HTTP (im a fan of building a DHT document client into web browsers, and Ending the wikileaks shutdown issue) | 20:03 |
alekibango | MarkAtwood: big brother (and in america its sis) wants to know everything and limit your access to information... | 20:03 |
alekibango | MarkAtwood: do you know how internet2 will look like? | 20:04 |
creiht | http://www.renesys.com/blog/2010/12/wikileaks-moving-target.shtml | 20:04 |
MarkAtwood | there are 47 different DHTs running, andall of them currently suck | 20:04 |
creiht | you guys may find that blog post interesting (from my old boss) | 20:04 |
alekibango | i think wikileaks will be used to shut down free internet | 20:05 |
alekibango | i mean used as example which we need to fight against | 20:05 |
MarkAtwood | shut down open HTTP more like | 20:05 |
MarkAtwood | http and the current DNS is far too centeralized, but that is more or less off topic here | 20:06 |
alekibango | because we need to be dumb, we need to not know how goverments do wars and how they do false flag terror atacks (operations like paperclip, northwoods, gladio, ajax, .... gulf of tonkin incident, 911, 7/7 etc) | 20:06 |
*** joearnold has quit IRC | 20:07 | |
alekibango | they even talked here in eu - about allowing internet access only via buro stamped software -- which means 'trusted computing' and end of all freedom | 20:07 |
creiht | you guys can finish the conversation on #tinfoilhat :) | 20:07 |
*** dendro-afk is now known as dendrobates | 20:08 | |
KyleM1 | I'm trying to set up a single-machine nova install and just ran the following command to create a project: | 20:08 |
KyleM1 | nova-manage project create test_project test_user | 20:08 |
alekibango | creiht: what i named are actual names of operations which were publicly admitted to be done!!! its not a secret now | 20:08 |
alekibango | + some recent ones which stink a lot | 20:09 |
creiht | alekibango: what I was trying to say fact or not, this isn't the place :) | 20:09 |
alekibango | but ok, its OT | 20:09 |
KyleM1 | The instructions I'm following suggest that it should print some RSA key stuff and then spit out a zip file. | 20:09 |
KyleM1 | Instead I get no output at all. | 20:09 |
openstackhudson | Project nova build #296: SUCCESS in 1 min 19 sec: http://hudson.openstack.org/job/nova/296/ | 20:09 |
KyleM1 | if I try `nova-manage project list` I can see "test_project" there | 20:09 |
alekibango | creiht: ok, just dont call me tinfoilhatter please | 20:10 |
KyleM1 | But if I try `nova-manage project zipfile test_project test_user` it starts printing RSA key stuff and then dies with the following error: | 20:10 |
KyleM1 | OSError: [Errno 2] No such file or directory: '/usr/local/lib/python2.6/dist-packages/nova-2011.1-py2.6.egg/nova/..//CA' | 20:10 |
KyleM1 | Any suggestions? | 20:10 |
sandywalsh | KyleM1, did you create your CA directory? | 20:12 |
KyleM1 | There's one in nova/tests, but not one in python2.6/dist-packages/.../ | 20:13 |
KyleM1 | I did not create one manually. | 20:13 |
dubsquared | kylem1: the default location is "/usr/local/lib/python2.6/dist-packages/nova-2011.1-py2.6.egg/nova/..//CA'" … in your nova.conf, you need to add ''--state_path=/var/lib/nova" in your nova.conf, and restart nova-compute | 20:13 |
KyleM1 | Thanks. | 20:13 |
dubsquared | I'm /pretty/ sure that will fix ya up, let me know if it doesn't :D | 20:14 |
*** westmaas has joined #openstack | 20:15 | |
uvirtbot | New bug: #688220 in nova "Need to add a dependency for python-twisted >= 10.1.0" [Undecided,New] https://launchpad.net/bugs/688220 | 20:16 |
KyleM1 | Still no dice. I'm following the instructions here (http://nova.openstack.org/adminguide/single.node.install.html) to try to get nova running. | 20:16 |
KyleM1 | My end goal is to contribute to the code (specifically volume/driver.py) so I don't want to install from packages, but I haven't managed to get nova running at all yet and I've tried a couple different ways. | 20:17 |
dubsquared | are you using packages? | 20:17 |
KyleM1 | Nope. | 20:17 |
dubsquared | ah :D | 20:17 |
KyleM1 | Branched from trunk. | 20:17 |
KyleM1 | I tried Vishy's script as alekibango suggested yesterday but couldn't get that working (I'm installing on debian/squeeze rather than ubuntu)> | 20:18 |
alekibango | KyleM1: mine failed for you as those were packages which need another group | 20:18 |
alekibango | but create that group | 20:18 |
dubsquared | I've been more focused on the package side of nova so far; i did run into that, and the state_path was the fix | 20:18 |
KyleM1 | Hrm. | 20:19 |
alekibango | as alias of the libvirt group and it will work :) | 20:19 |
alekibango | sorry i forgot to tell you | 20:19 |
*** westmaas has left #openstack | 20:19 | |
alekibango | or make the group and add nova user to make sure nova user is in libvirt group | 20:19 |
alekibango | add nova to the group i mean | 20:19 |
alekibango | KyleM1: i will build new ones for you :) | 20:20 |
KyleM1 | alekibango: but I want to be able to test my changes to the code; can I do that if I install from packages rather than source? | 20:20 |
alekibango | that depends,. python still is source | 20:21 |
alekibango | you can use git + packages :) and be happy | 20:21 |
alekibango | or bzr from nova trunk | 20:22 |
alekibango | its your choice | 20:22 |
*** GasbaKid has joined #openstack | 20:22 | |
alekibango | in the end you will need bzr to push it back to your branch in launchpad to ask for merge | 20:22 |
alekibango | creiht: btw the article is nice, but not groundbreaking | 20:23 |
KyleM1 | I'd like to install from source if possible, rather than packages. | 20:24 |
*** BK_man_ has joined #openstack | 20:24 | |
*** BK_man has quit IRC | 20:24 | |
*** BK_man_ is now known as BK_man | 20:24 | |
*** brd_from_italy has joined #openstack | 20:24 | |
*** daleolds has quit IRC | 20:24 | |
alekibango | ok | 20:24 |
*** daleolds has joined #openstack | 20:24 | |
KyleM1 | Unfortunately the single-machine setup instructions seem to be out of date or incomplete. | 20:24 |
alekibango | might be... | 20:24 |
alekibango | please update them.. try asking here - ask real questions we will try to help | 20:24 |
alekibango | and update those pages, its moving so fast that we need to hold our hair (hats are gone now) | 20:25 |
KyleM1 | Hah. | 20:25 |
KyleM1 | Yes... | 20:25 |
KyleM1 | So yesterday I tried a couple different installation methods, but I seem to remember that following the instructions for single-machine setup didn't create /etc/nova/nova.conf. I have that file now but I'm not sure if there's another location that nova might look for nova.conf before checking /etc/nova that might be overriding the one in /etc. | 20:27 |
KyleM1 | Is there any other place that I should check for nova.conf? | 20:27 |
alekibango | if you have your scripts in /srv/cloud/ | 20:27 |
alekibango | delete /etc/nova | 20:28 |
eday | argh, sqlalchemy doesn't bytes for it's string types | 20:28 |
alekibango | eday: maybe try to talk to them :) | 20:28 |
alekibango | but problem with SA is to be implementation neutral | 20:28 |
KyleM1 | alekibango: What scripts should I have in /srv/cloud? | 20:28 |
alekibango | i checked my nova to /srv/cloud/ | 20:29 |
alekibango | as vishy did | 20:29 |
alekibango | his script somehow had this wired in in some time | 20:29 |
alekibango | but thats weeks ago | 20:29 |
KyleM1 | I don't understand what you mean by this: "i checked my nova to /srv/cloud/" | 20:29 |
alekibango | ha, i meant bzr pull | 20:29 |
KyleM1 | Ah. | 20:29 |
KyleM1 | okay. | 20:29 |
alekibango | checkout in svn world :) | 20:30 |
KyleM1 | Gotcha. | 20:30 |
alekibango | or bzr clone? i still fail to remember bzr commands :) | 20:30 |
KyleM1 | I think i've use bzr branch. | 20:30 |
KyleM1 | *shrug* | 20:30 |
alekibango | all will be in the directory. including conf | 20:30 |
alekibango | and images | 20:31 |
KyleM1 | Okay. I'm gonna head to lunch now, and I'll try that after I get back. | 20:31 |
KyleM1 | Thanks for your help. | 20:31 |
*** gondoi has quit IRC | 20:31 | |
alekibango | np | 20:31 |
*** Podilarius has left #openstack | 20:40 | |
*** Ryan_Lane|food is now known as Ryan_Lane | 20:47 | |
vishy | eday: 'the ec2 ids are going to get quite a bit larger now' we could allow people to specify ec2_ids as a prefix to make it easier to use the system without typing out i-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx. We could provide a get_by_project_and_prefix that would do a LIKE i-xxx% query | 20:47 |
vishy | KyleM1: Those instrucitons appear to be incorrect | 20:54 |
*** BK_man has quit IRC | 20:54 | |
*** BK_man has joined #openstack | 20:54 | |
*** Podilarius has joined #openstack | 20:56 | |
annegentle | vishy: I've got a merge request in that updates the single mode install instructions, https://code.launchpad.net/~annegentle/nova/installdoc/+merge/43022 | 20:56 |
annegentle | vishy: there are probably more changes to be made but at least it avoids confusion about multiple .conf files | 20:57 |
vishy | KyleM1: Or perhaps the packaging is wrong | 20:57 |
*** iammartian has joined #openstack | 21:01 | |
*** pothos has quit IRC | 21:02 | |
*** pothos has joined #openstack | 21:04 | |
*** pothos has quit IRC | 21:07 | |
*** pothos has joined #openstack | 21:07 | |
*** gundlach has left #openstack | 21:22 | |
*** westmaas1 has joined #openstack | 21:24 | |
*** BK_man has quit IRC | 21:24 | |
*** BK_man has joined #openstack | 21:24 | |
*** jkakar has joined #openstack | 21:26 | |
*** aliguori has quit IRC | 21:27 | |
*** miclorb has quit IRC | 21:32 | |
*** irahgel1 has quit IRC | 21:46 | |
*** piken is now known as piken_train | 21:50 | |
*** BK_man_ has joined #openstack | 21:54 | |
*** BK_man has quit IRC | 21:54 | |
*** BK_man_ is now known as BK_man | 21:54 | |
*** westmaas1 has quit IRC | 21:55 | |
*** aliguori has joined #openstack | 22:01 | |
*** arthurc has quit IRC | 22:07 | |
*** joearnold has joined #openstack | 22:07 | |
*** joearnold has quit IRC | 22:09 | |
*** aemdb has joined #openstack | 22:10 | |
*** joearnold has joined #openstack | 22:11 | |
eday | alekibango, vishy: https://code.launchpad.net/~eday/nova/instance-uuids/+merge/43301 used uuids for now, can replace with ipv6 if/when we want later | 22:12 |
aemdb | Hi everyone, could anybody help me a bit? I'm running openstack-nova under ubuntu 10.04 using the novascript, and have done everything to the point of having my ami-tiny instance running. But I can't ssh to it, I get no route to host. Any ideas? | 22:12 |
*** maple_bed has quit IRC | 22:14 | |
*** dfg_ has joined #openstack | 22:17 | |
*** abecc has joined #openstack | 22:17 | |
*** abecc has joined #openstack | 22:18 | |
*** jdarcy has quit IRC | 22:19 | |
*** BK_man has quit IRC | 22:23 | |
chmouel | if people using zsh and get annoyed to get to type/findout the instance name here is some completion for zsh and some euca functions https://github.com/chmouel/zsh-config/blob/master/functions/_euca-instances.zsh | 22:23 |
*** aemdb has quit IRC | 22:32 | |
vishy | aemdb: ps ax | grep dnsmasq: how many dnsmasq processes are running? | 22:36 |
jeremyb | vishy: he left before you asked | 22:39 |
*** hggdh has quit IRC | 22:39 | |
*** jimbaker has quit IRC | 22:48 | |
*** ppetraki has quit IRC | 22:48 | |
*** jk0 is now known as jk0_ | 22:54 | |
*** jk0_ is now known as jk0 | 22:54 | |
*** brd_from_italy has quit IRC | 23:03 | |
*** dfg_ has quit IRC | 23:05 | |
*** JuanPerez has joined #openstack | 23:07 | |
*** arcane has joined #openstack | 23:08 | |
*** AimanA is now known as HouseAway | 23:09 | |
*** jimbaker has joined #openstack | 23:10 | |
JuanPerez | Hi, long time listener, first time posting. :) Does anyone know if there are any plans to have a Forum site for OpenStack, in addition to the #openstack channel? | 23:11 |
JordanRinke | Hit up stephen.spector@openstack.org | 23:12 |
JordanRinke | he is the community manager | 23:12 |
JuanPerez | cool, thank you Jordan | 23:12 |
JordanRinke | You are welcome, here to help | 23:13 |
*** mdomsch has quit IRC | 23:15 | |
*** allsystemsarego has quit IRC | 23:15 | |
*** abecc_ has joined #openstack | 23:15 | |
eday | JuanPerez: I'd ask things on the mailing list too | 23:16 |
annegentle | JuanPerez: yes, definitely use http://launchpad.net/~openstack (openstack@lists.launchpad.net) as if it were a forum. | 23:17 |
*** abecc has quit IRC | 23:19 | |
*** abecc_ is now known as abecc | 23:19 | |
*** rnirmal has quit IRC | 23:19 | |
dubsquared | It's JuanPerez! | 23:22 |
dubsquared | About time you came out to play... | 23:22 |
JuanPerez | thank you eday and anne. i had been reading the meeting minutes for openstack, and it noted that only 3 messages had been posted to the 8 mailing-lists, so I thought it may be beneficial to have a search-able Forum just for OpenStack. | 23:22 |
eday | JuanPerez: the launchpad mailing list is more active | 23:22 |
JordanRinke | We should consolidate the mailing lists down to just openstack, openstack-clientdev (after talking to Mike Mayo) and openstack-users | 23:25 |
JordanRinke | a forum is a good idea though, potentially more for users than developers since the pace is significantly slower for user implementations | 23:26 |
JordanRinke | and the ability to search a question before asking it | 23:27 |
JordanRinke | hm | 23:27 |
JuanPerez | eday: is the answers section in launchpad the same as the launchpad mailing list? https://answers.launchpad.net/nova | 23:35 |
eday | JuanPerez: no, it's separate, but both places are fairly active | 23:40 |
JuanPerez | fyi... Stephen just replied that the forum idea is being discussed | 23:40 |
JuanPerez | eday: thank you for the pointer. found the mailing list archive: https://lists.launchpad.net/openstack/. irc logs are here also: http://eavesdrop.openstack.org/irclogs/ | 23:47 |
*** JuanPerez has quit IRC | 23:56 | |
*** dubsquared has quit IRC | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!