*** t0mb0 has joined #akanda | 00:27 | |
*** t0mb0|work has joined #akanda | 00:32 | |
*** puranamr has quit IRC | 01:11 | |
*** puranamr has joined #akanda | 04:34 | |
*** puranamr has quit IRC | 05:02 | |
*** t0mb0 has quit IRC | 08:13 | |
*** t0mb0|work has quit IRC | 08:26 | |
*** t0mb0 has joined #akanda | 10:51 | |
*** t0mb0 has quit IRC | 10:57 | |
*** t0mb0 has joined #akanda | 11:01 | |
*** t0mb0 has quit IRC | 12:54 | |
*** t0mb0 has joined #akanda | 12:55 | |
*** t0mb0 has quit IRC | 13:07 | |
sarob | adam_g, davidlenwell still having problems | 15:40 |
---|---|---|
davidlenwell | come into the office .. lets work through it from the begining | 15:41 |
davidlenwell | sarob: ^^ | 15:41 |
sarob | davidlenwell: yeah | 15:41 |
*** puranamr has joined #akanda | 17:01 | |
*** puranamr has quit IRC | 17:02 | |
*** puranamr has joined #akanda | 17:03 | |
puranamr | drwahl: ping | 17:03 |
drwahl | pong | 17:03 |
puranamr | here is the way to get started | 17:05 |
puranamr | can u pl tear down the compute node and recreate it | 17:05 |
puranamr | only recreation, nothing else | 17:05 |
drwahl | yup | 17:05 |
puranamr | do not assign any Floating ip | 17:05 |
drwahl | sure | 17:05 |
puranamr | i will access from the control node | 17:05 |
puranamr | is there a way for you to watch what I am doing? or do you want to wait till i complete it? | 17:08 |
drwahl | depends on what you are doing exactly :) | 17:09 |
drwahl | if it's just modifying local.conf stuff, i can diff the changes easily enough | 17:09 |
drwahl | if you are modifying the system at all (creating network interfaces, loading kernel modules, etc.), it might be easier for me to watch | 17:10 |
puranamr | nothing out of the blue :) all as per laid out info only | 17:10 |
drwahl | ok, i don't really need to watch then | 17:10 |
puranamr | no i should not do those kind of customizing at all | 17:10 |
puranamr | did u create one, what is the IP | 17:10 |
drwahl | coming up now.... | 17:11 |
puranamr | ok | 17:11 |
drwahl | you have access to the controller node right now, right? | 17:12 |
davidlenwell | adam_g: ping | 17:13 |
puranamr | looking for the controller IP | 17:13 |
puranamr | is it 67.205.61.202? | 17:14 |
drwahl | 67.205.62.65 | 17:14 |
puranamr | controller right | 17:15 |
drwahl | compute node is 10.10.10.168 | 17:15 |
puranamr | permission denied? | 17:15 |
drwahl | sigh... one sec | 17:15 |
puranamr | sure | 17:15 |
drwahl | doh! stupid ssh! | 17:16 |
drwahl | i think it's using upstart, but i'm using the sysVinit scripts to restart it | 17:16 |
puranamr | ok | 17:16 |
drwahl | ok, try now | 17:17 |
puranamr | nope | 17:18 |
drwahl | hmm... working for me | 17:19 |
puranamr | y-day was able to get in, sure | 17:19 |
drwahl | what IP are you trying to ssh to? | 17:19 |
drwahl | oh | 17:20 |
drwahl | you are trying to get into the controller, huh? | 17:20 |
drwahl | ok, try now... | 17:20 |
drwahl | (i thought you were already on the controller) | 17:20 |
puranamr | it was closed, this is controller ip i am using: 67.205.62.65 | 17:20 |
drwahl | ya, try logging into that as dhc-user again | 17:21 |
puranamr | sure | 17:21 |
drwahl | it *should* work now | 17:21 |
* drwahl crosses his fingers... | 17:21 | |
puranamr | yup | 17:21 |
puranamr | :) | 17:21 |
drwahl | \o/ | 17:21 |
drwahl | it's almost like i know how computers work! | 17:22 |
drwahl | heh | 17:22 |
puranamr | :) | 17:22 |
openstackgerrit | David Lenwell proposed stackforge/akanda-rug: Added blessed to the requirements file as it is needed by rug-ctl browse https://review.openstack.org/205184 | 17:24 |
adam_g | davidlenwell, here now' | 17:27 |
davidlenwell | I've been trying to figure out why this is failing.. https://review.openstack.org/#/c/203834/ | 17:28 |
puranamr | drwahl: controller is .156? | 17:28 |
drwahl | 157 | 17:28 |
puranamr | ok | 17:29 |
adam_g | davidlenwell, http://logs.openstack.org/34/203834/4/check/gate-functional-dsvm-akanda/e5288c5/logs/screen-n-cpu.txt.gz?level=ERROR | 17:31 |
adam_g | davidlenwell, its failing on https://launchpad.net/bugs/1476402 because with that permissions patch, the test actually gets run | 17:31 |
openstack | Launchpad bug 1476402 in OpenStack Compute (nova) "boot fails on network /w only an ipv6 subnet, AttributeError: 'NoneType' object has no attribute 'get'" [Undecided,In progress] | 17:31 |
adam_g | davidlenwell, im hoping we can use https://review.openstack.org/#/c/203875/ to get around it and be able to land code while we wait on nova | 17:32 |
*** cleverdevil has joined #akanda | 17:44 | |
*** cleverdevil has quit IRC | 17:48 | |
*** cleverdevil has joined #akanda | 17:48 | |
davidlenwell | I manually built on https://review.openstack.org/#/c/203875/ and verrified that the router comes up and becomes active | 17:50 |
davidlenwell | adam_g: ^ | 17:50 |
davidlenwell | So why is it still failing | 17:50 |
adam_g | let me set that permissions patch as dependent on the nova fix thing | 17:51 |
puranamr | drwahl: ping | 17:51 |
drwahl | pong | 17:52 |
davidlenwell | ahh .. yeah .. that would make sense | 17:52 |
puranamr | was it plain trusty image that u created as compute vm? | 17:52 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Remove 0755 permissions from files https://review.openstack.org/203834 | 17:52 |
drwahl | more or less. just the trusty image that we provide for dreamcompute | 17:52 |
drwahl | only slight modifications for things like dhc-user. just cloud-init stuff | 17:53 |
drwahl | otherwise just a plain trusty image | 17:53 |
puranamr | what is in the cloud-init? | 17:53 |
adam_g | davidlenwell, doh, actually the nova workaround is working yet http://logs.openstack.org/75/203875/2/check/gate-functional-dsvm-akanda/13b987e/logs/devstacklog.txt.gz | 17:53 |
drwahl | puranamr: /etc/cloud/cloud.cfg.d/99_dreamcompute_common.cfg has the details | 17:53 |
puranamr | i created a trusty image from dream compute and ran the same steps as i just did, and there is difference i see | 17:53 |
puranamr | ok let me check | 17:54 |
drwahl | i literally just selected the trusty image via horizon and then set the password for the dhc-user | 17:54 |
drwahl | so it should be essentially identical to any other trusty image you would spin up in DHC | 17:55 |
puranamr | yes, it is the same as i see in my compute node too | 17:55 |
puranamr | the 99_.cfg file i mean | 17:55 |
drwahl | our image is slightly out of date, so you may need/want to update the system (apt-get upgrade) | 17:55 |
puranamr | yeah i have done all those steps before stacking | 17:56 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Remove 0755 permissions from files https://review.openstack.org/203834 | 17:56 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: devstack: set use_ipv6=True in nova.conf, workaround nova bug https://review.openstack.org/203875 | 17:56 |
*** cleverdevil has quit IRC | 17:58 | |
*** cleverdevil has joined #akanda | 17:58 | |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Fix debug and browse config issues https://review.openstack.org/204748 | 17:58 |
adam_g | davidlenwell, also, the absence of blessed from requirements.txt is intentional | 17:58 |
davidlenwell | yeah .. I saw your comment | 17:59 |
davidlenwell | I'm looking at how browse works right now | 17:59 |
drwahl | puranamr: i need to step away for 10-15 minutes. brb | 18:00 |
puranamr | sure | 18:00 |
davidlenwell | adam_g: I just have a hard time accepting that our project has a dependancy that we don't include in requirements.. I get the global requirements thing.. but isn't the correct behavior to change the software to not have the requirement rather than break our cli tool ? I get that its a minor issue to install a thing .. | 18:04 |
davidlenwell | I don't know .. it bugs me | 18:04 |
ryanpetrello | davidlenwell: you see the "extra dependencies" setuptools thing I linked | 18:05 |
ryanpetrello | seems like a reasonable way to do it | 18:05 |
davidlenwell | no | 18:05 |
davidlenwell | ahh | 18:05 |
davidlenwell | ryanpetrello: did you make a commit with it some place? | 18:06 |
ryanpetrello | nope, just a link to how it works | 18:06 |
ryanpetrello | a comment to the review you posted | 18:06 |
adam_g | davidlenwell, we discussed it a bit. it sounds like people are now using the horizon integratino now that its there, more than the ctl browse command. | 18:06 |
ryanpetrello | honestly, I don't really use rug-ctl browse at all | 18:06 |
davidlenwell | I use it | 18:06 |
ryanpetrello | because it can't keep up with a larger cluster | 18:06 |
ryanpetrello | too many OpenStack calls | 18:06 |
davidlenwell | ahh | 18:07 |
davidlenwell | well its handy for debugging devstack installs | 18:07 |
ryanpetrello | it's useful for dev debugging | 18:07 |
adam_g | davidlenwell, i wouldnt be against just dropping browse all together, or re-writing it to be prettytable based/non-interactive.. but i dont have cycles for the latter | 18:07 |
ryanpetrello | which is why I think it would make sense as an optional setuptools dependency | 18:07 |
davidlenwell | jinx.. I expect a coke delibery | 18:07 |
davidlenwell | adam_g: yep.. okay.. im gonna drop it for now | 18:07 |
ryanpetrello | davidlenwell: http://stackoverflow.com/questions/10572603/specifying-optional-dependencies-in-pypi-python-setup-py example | 18:07 |
davidlenwell | we have bigger fish to fry | 18:08 |
ryanpetrello | so I think you can do e.g., `pip install akanda-rug[dev]` | 18:08 |
ryanpetrello | (or whatever you name "dev") | 18:08 |
adam_g | davidlenwell, you can modify plugin.sh do 'pip_install blessed' at some point so something is working out of the box. if we end up distro packaging the rug, it can be tacked onto packaging | 18:08 |
davidlenwell | k | 18:08 |
davidlenwell | adam_g: works forme | 18:09 |
openstackgerrit | David Lenwell proposed stackforge/akanda-rug: added pip install blessed to plugin.sh https://review.openstack.org/205207 | 18:12 |
* drwahl is back | 18:17 | |
openstackgerrit | David Lenwell proposed stackforge/akanda-rug: added pip install blessed to plugin.sh https://review.openstack.org/205207 | 18:28 |
*** puranamr has quit IRC | 19:19 | |
*** puranamr has joined #akanda | 19:22 | |
drwahl | puranamr: any progress? | 19:24 |
puranamr | i am debugging the issue. i see a change of behavior (like what is in ur setup) in a newly created compute node. i shall keep you posted | 19:25 |
drwahl | thanks | 19:25 |
*** cleverdevil has quit IRC | 20:02 | |
openstackgerrit | Merged stackforge/akanda-rug: added pip install blessed to plugin.sh https://review.openstack.org/205207 | 20:05 |
*** puranamr has quit IRC | 20:09 | |
*** puranamr has joined #akanda | 20:57 | |
sarob | sounds like i should be using adam_g demo build using stable/kilo | 21:50 |
adam_g | huh? | 21:50 |
sarob | adam_g: you have stable/kilo running appliance with current patches? | 21:51 |
sarob | adam_g: or am i confused | 21:51 |
adam_g | sarob, i have a kilo environment snapshotted, yea | 21:51 |
adam_g | trying to get master going | 21:51 |
sarob | okay on the akanda tenant or yours | 21:52 |
sarob | we need to script | 21:53 |
sarob | break a route | 21:53 |
sarob | break an appliance | 21:54 |
sarob | the demo steps | 21:54 |
sarob | i will show horizon network | 21:54 |
sarob | then show icmp running against one of the interfaces | 21:54 |
sarob | i run the break the route script | 21:55 |
sarob | icmp fails | 21:55 |
sarob | then i manually update the router | 21:55 |
adam_g | sarob, that script is done | 21:55 |
sarob | cool in ops yet? | 21:55 |
adam_g | yup | 21:55 |
sarob | got it | 21:56 |
sarob | reviewing | 21:56 |
sarob | looks good | 21:58 |
adam_g | sarob, check your msgs | 22:00 |
*** cleverdevil has joined #akanda | 23:15 | |
*** cleverdevil has quit IRC | 23:50 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!