stupidnic | Is this a recent change? | 00:00 |
---|---|---|
fzylogic | we put config drive in ceph and it works | 00:00 |
fzylogic | and yeah, it's only recently supported | 00:00 |
stupidnic | Okay. I had this working a couple of months ago | 00:00 |
fzylogic | that's also why you didn't see that file there. when using ceph, it's only put on the local FS temporarily | 00:01 |
adam_g | stupidnic, what's in /proc/partitions within the appliance? | 00:01 |
stupidnic | hold on... it rebooted out from under me so I didn't get back in... checking now | 00:01 |
adam_g | if its in the libvirt domain xml, it *should* be showing up there | 00:01 |
stupidnic | adam_g: I see sr0 | 00:02 |
stupidnic | and it rebooted out from under me again... ugh | 00:03 |
adam_g | stupidnic, once you have it booted just killed the astara-orchestrator | 00:04 |
stupidnic | k | 00:04 |
adam_g | stupidnic, are you on liberty or master? | 00:04 |
stupidnic | liberty | 00:04 |
adam_g | stupidnic, what are the detalis of the management network you're using? | 00:04 |
stupidnic | What sort of information are you looking for? | 00:05 |
stupidnic | It's a vxlan and I put it in the service project | 00:05 |
adam_g | management_prefix, mainly | 00:05 |
stupidnic | fdd6:a1fa:cfa8:748e::/64 | 00:05 |
stupidnic | I can see packets coming from the controller to the compute node that the instance is on... and we have verified that the tap and vxlan interface in the same bridge | 00:06 |
adam_g | stupidnic, you still have the appliance up? | 00:07 |
stupidnic | yes | 00:07 |
adam_g | $ sudo mount /dev/sr0 /mnt | 00:07 |
stupidnic | k | 00:07 |
adam_g | $ cat /mnt/openstack/latest/user_data | grep manag | 00:08 |
adam_g | thats the cmd that should be running at boot to bring up the initial management interface | 00:09 |
adam_g | (assuming you grep something out of user_data) | 00:09 |
stupidnic | I can confirm that is there | 00:09 |
adam_g | running it manually *should* bring up whatever address is passed on the NIC /w the specified mac addr | 00:10 |
adam_g | (presumably eth0) | 00:10 |
stupidnic | okay... tried typing in that command | 00:12 |
stupidnic | /usr/localbin/astara-configure-management | 00:12 |
stupidnic | not found | 00:12 |
stupidnic | I am using the image downloaded from http://tarballs.openstack.org/akanda-appliance/images/ | 00:12 |
stupidnic | I am guessing that is way too old | 00:13 |
adam_g | thats the issue | 00:13 |
stupidnic | OKAY | 00:14 |
adam_g | ill build you a disk image for liberty | 00:14 |
stupidnic | so... rebuild my own | 00:14 |
adam_g | we should have liberty images being published there soonish i hope | 00:14 |
*** prithiv has joined #openstack-astara | 00:14 | |
stupidnic | adam_g: I am glad we got to the bottom of that... I was following your etherpad instructions ;) | 00:14 |
adam_g | stupidnic, do you need me to build an image for you or are you familiar enough with the process ? | 00:15 |
adam_g | we really need a good story for logging in the appliance, where we can debug this type of thing from the outside | 00:15 |
stupidnic | As long as the astara-appliance repo has the necessary changes... I should be able to handle it | 00:15 |
stupidnic | log to the console... that then gets dumped into main console log | 00:16 |
stupidnic | which you can easily access | 00:16 |
adam_g | stupidnic, yeah, that was the idea i had too | 00:17 |
adam_g | stupidnic, i think checking out stable/liberty of astara-appliance and running 'tox -e build_image' should be all you need | 00:21 |
stupidnic | hmmm okay... I wasn't aware that there was a liberty branch | 00:21 |
stupidnic | I might need to update my config | 00:21 |
adam_g | doh actually if you're using ceph you'll need to build the image as raw | 00:24 |
adam_g | or convert the qcow to raw after | 00:24 |
adam_g | before uploading | 00:24 |
stupidnic | Yeah I was aware of that | 00:24 |
stupidnic | Okay. Have the image building now | 00:26 |
stupidnic | adam_g: so I can make a note of it, what is the command to build the image as raw instead of qcow2? | 00:32 |
*** shashank_hegde has quit IRC | 00:33 | |
adam_g | stupidnic, you need to just pass '-t raw' to disk-image-create, you can pass it through tox | 00:33 |
adam_g | tox -e build_image -- -t raw | 00:33 |
stupidnic | got it thanks. | 00:42 |
adam_g | ok i gotta run, soups on. have a good weekend | 00:48 |
stupidnic | sigh | 00:49 |
stupidnic | still not working | 00:49 |
stupidnic | but the instance at least has the right hostname now | 00:49 |
stupidnic | Freshly built astara appliance from stable/liberty does not include the /usr/local/bin/astara-configure-management | 01:04 |
stupidnic | once again... astara versus akanda bites me in the ass | 01:05 |
stupidnic | the confdrive that is generated references astara-configure-management and the image has akanada-configure-management | 01:06 |
stupidnic | I am sure that I have the wrong version of astara installed on the controller | 01:07 |
elo | this is from the liberty branch for both | 01:24 |
stupidnic | ugh | 01:28 |
stupidnic | Still not working... | 01:29 |
stupidnic | the confdrive references astara-configure-management and the image only has akanda-configure-management | 01:29 |
stupidnic | would be nice to have a symlink between the two | 01:29 |
elo | yes. we should have backwards compatibility for the orchestrator to manager older appliance images | 01:31 |
elo | this is most likely a bug as I remember talking with Adam about this type of issue several days ago | 01:32 |
stupidnic | I think my problem stems from not using the stable/liberty branches on the git repos I used to install from | 01:32 |
stupidnic | elo: so what should I be doing here? Should I be using the stable/liberty branch on the appliance? Or should I go with HEAD so that I can get around this issue? | 01:36 |
elo | I've not tested or deployed different versions. | 01:36 |
elo | I would say go with HEAD to get around for now | 01:36 |
stupidnic | Well I have all my repos checked out at stable/liberty | 01:36 |
stupidnic | and I reran pip install on them | 01:37 |
stupidnic | but it doesn't seem to have corrected the issues | 01:37 |
stupidnic | My other option is to tinker with the image itself and just make the symlink in the image I have already built | 01:37 |
elo | at least for the appliance to see if we can get around. Or you can tinker the image to create a symlink | 01:38 |
stupidnic | Let me try and rebuild on head | 01:38 |
stupidnic | I would rather not have to do any tinkering | 01:38 |
elo | understandable... I have to run... but will be back online later tonight and this weekend | 01:43 |
elo | ping me if you need me to help out more | 01:43 |
stupidnic | Alright. Thanks for your help | 01:43 |
stupidnic | I think I have the issue nailed down... just need to figure out how to correct it | 01:44 |
elo | I'll talk with Adam about backward compatibility issue | 01:44 |
stupidnic | is there any documentation about tweaking the build process? | 01:44 |
elo | we leverage diskimage-builder | 01:45 |
elo | http://docs.openstack.org/developer/diskimage-builder/ | 01:45 |
stupidnic | alright, thanks | 01:46 |
elo | I'm not aware of anything specific to astara-appliance images | 01:46 |
elo | I'll see what we can put together | 01:49 |
stupidnic | Progress is being made... | 02:01 |
stupidnic | I can actually ping the instance now | 02:02 |
stupidnic | but it still isn't be marked as online | 02:02 |
stupidnic | and the logs show that astara can't make the http connection to the instance so something else is wrong there | 02:02 |
*** prithiv has quit IRC | 02:17 | |
openstackgerrit | ting wang proposed openstack/astara: Py3: use dict.items() instead of dict.iteritems() https://review.openstack.org/268409 | 02:21 |
*** yanghy has joined #openstack-astara | 02:36 | |
openstackgerrit | ting wang proposed openstack/astara: Rewrite filter(lambada) in eventlet_backdoor.py https://review.openstack.org/268411 | 02:36 |
openstackgerrit | ting wang proposed openstack/astara: Py3: use dict.items() instead of dict.iteritems() https://review.openstack.org/268409 | 03:12 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/astara: Updated from global requirements https://review.openstack.org/268425 | 03:25 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/astara-appliance: Updated from global requirements https://review.openstack.org/268426 | 03:25 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/astara-horizon: Updated from global requirements https://review.openstack.org/268427 | 03:25 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/astara-neutron: Updated from global requirements https://review.openstack.org/268428 | 03:25 |
*** shashank_hegde has joined #openstack-astara | 06:06 | |
*** shashank_hegde has quit IRC | 06:35 | |
*** shashank_hegde has joined #openstack-astara | 06:48 | |
*** shashank_hegde has quit IRC | 07:10 | |
openstackgerrit | caoyue proposed openstack/astara: Remove unused logging import https://review.openstack.org/268614 | 07:36 |
openstackgerrit | caoyue proposed openstack/astara: Remove unused logging import https://review.openstack.org/268614 | 09:58 |
*** leonstack1 has quit IRC | 12:46 | |
*** leonstack has joined #openstack-astara | 12:48 | |
*** yanghy has quit IRC | 13:26 | |
*** davidlenwell has quit IRC | 15:49 | |
*** davidlenwell has joined #openstack-astara | 15:52 | |
*** shashank_hegde has joined #openstack-astara | 16:28 | |
*** shashank_hegde has quit IRC | 18:00 | |
*** shashank_hegde has joined #openstack-astara | 18:09 | |
*** shashank_hegde has quit IRC | 18:14 | |
*** shashank_hegde has joined #openstack-astara | 18:14 | |
*** shashank_hegde has quit IRC | 19:08 | |
*** shashank_hegde has joined #openstack-astara | 19:10 | |
stupidnic | sigh... so much progress and then another wall | 19:47 |
stupidnic | The router appliance is not responding to DHCP requests | 19:48 |
stupidnic | I can see the requests are making it all the way into the appliances tap interface, it just doesn't respond | 19:48 |
stupidnic | No my mistake... it is responding but with the wrong IP | 19:49 |
*** shashank_hegde has quit IRC | 20:07 | |
stupidnic | Okay. So the DHCP issue was related to security groups. So I opened up the security groups and the instance can get an IP now which is great. | 21:36 |
stupidnic | The next problem is that the metadata proxy is returning a 500 error | 21:37 |
*** shashank_hegde has joined #openstack-astara | 21:50 | |
*** shashank_hegde has quit IRC | 22:20 | |
*** shashank_hegde has joined #openstack-astara | 22:21 | |
*** shashank_hegde has quit IRC | 22:39 | |
*** shashank_hegde has joined #openstack-astara | 22:54 | |
*** shashank_hegde has quit IRC | 23:25 | |
*** shashank_hegde has joined #openstack-astara | 23:44 | |
*** shashank_hegde has quit IRC | 23:46 | |
*** shashank_hegde has joined #openstack-astara | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!