Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2015-06-04

| Channels | #fuel index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:26 xarses joined #fuel
00:45 mattgriffin joined #fuel
00:57 LiJiansheng joined #fuel
01:05 rmoe joined #fuel
01:09 CheKoLyN joined #fuel
01:22 CheKoLyN joined #fuel
01:48 mattgriffin joined #fuel
02:02 mattgriffin joined #fuel
03:20 tatyana joined #fuel
03:36 lquince joined #fuel
04:14 Longgeek joined #fuel
04:25 wayneeseguin joined #fuel
04:35 lquince_ joined #fuel
04:37 lquince__ joined #fuel
04:53 Longgeek joined #fuel
04:57 Longgeek_ joined #fuel
05:09 lquince joined #fuel
06:11 monester_laptop joined #fuel
07:19 aarefiev joined #fuel
07:30 TiDjY35 joined #fuel
07:31 pradip_ joined #fuel
07:41 samuelBartel joined #fuel
07:51 artem_panchenko joined #fuel
08:38 baboune joined #fuel
08:40 baboune new Fuel controller, 6.0, "verify network" is hanging.  Been waiting for 20 min.  fuel task shows check_dhcp amd verify networks as "running" and "progress" 0.  What is the most likely issue? How to stop the tasks?
08:56 HeOS joined #fuel
08:58 lquince joined #fuel
09:00 lquince_ joined #fuel
09:08 HeOS joined #fuel
09:43 teran joined #fuel
09:44 e0ne joined #fuel
09:48 teran_ joined #fuel
10:37 LiJiansheng joined #fuel
10:54 Longgeek joined #fuel
11:08 monester_laptop joined #fuel
11:50 danwest joined #fuel
12:13 e0ne joined #fuel
12:15 Bomfunk joined #fuel
12:15 Hulda joined #fuel
12:34 rmoe joined #fuel
12:39 dklepikov joined #fuel
12:39 stamak joined #fuel
12:39 Hulda hey, got some issues with my deployment. i had to do a restore on docker on my fuel master and now the deployment fails for new nodes. ceph-deploy cant pull the config from the monitor node because it wont allow the pub key.
12:40 Hulda authorized_keys on the new node differs from the existing nodes. it's the second line that differs (root@<random text>). what is that?
12:43 kutija_ joined #fuel
12:46 DaveJ__ joined #fuel
13:02 kaliya Hulda: 6.0?
13:08 newbism_ joined #fuel
13:09 Miouge_ joined #fuel
13:10 Hulda kaliya: yes
13:11 Hulda i suspect it's a docker container? but the id doesnt match with any container
13:11 newbism_ OMFG
13:11 newbism_ well thats annoying
13:11 newbism_ whole thing failed because I had the nameservers twice... 2015-06-03 18:45:51.322 35961 TRACE neutron.api.v2.resource HTTPBadRequest: Invalid input for dns_nameservers. Reason: Duplicate nameserver ''.
13:12 newbism_ arg
13:12 newbism_ AAAAND because its 'deployed' I cant just change it
13:12 newbism_ I have to ... delete the entire environment, create a new one... reasssign all the interfaces
13:13 newbism_ reboot all the nodes, reinstall
13:21 lee__ joined #fuel
13:26 LiJiansheng joined #fuel
13:28 newbism_ Shoulden't there be a better resilancy after provision? Why reinstall the entire node/cluster why not just re-apply the puppet manifests?
13:37 kaliya newbism_: who prevented you to not reapply the manifests?
13:37 kaliya puppet apply /path/to/file.pp
13:37 newbism_ will that update the GUI too? I hate getting things out of sync
13:37 newbism_ like one time I deleted an LVM from under cinder.... bad bad bad
13:37 kaliya what is "the GUI"? Horizon? The Fuel UI? Other nice things?
13:38 newbism_ FUEl UI
13:38 newbism_ Fuel UI :)
13:38 kaliya that runs in the nginx container
13:38 newbism_ cool?
13:38 newbism_ it polls for status?
13:39 kaliya status of?
13:43 youellet_ joined #fuel
13:51 danwest joined #fuel
13:52 gsgs joined #fuel
14:06 e0ne joined #fuel
14:31 rmcadams joined #fuel
15:07 rmoe joined #fuel
15:09 championofcyrodi hi.  trying to increase DHCP agent's lease duration on fuel 6.0
15:09 championofcyrodi the default is a little too often for our instances.
15:10 championofcyrodi and generating excess traffic.  https://www.mirantis.com/blog/improving-dhcp-performance-openstack/ recommends such an action.
15:10 championofcyrodi but there is do dhcp_lease_duration option in /etc/neutron/dhcp_agent.ini
15:10 championofcyrodi is it just using the default of 120 since it is not specified?
15:11 championofcyrodi I'm thinking a slightly more sane value would be like... 300-600 (5-10 mins)
15:11 championofcyrodi generally i start a VM and leave it up for weeks
15:21 newbism_ joined #fuel
15:28 newbism_ Ok so I finally have an 'operational' cluster...
15:28 newbism_ try to go to horizon UI dash
15:28 newbism_ only thing on the page at port 8080 is ... {"versions": [{"status": "CURRENT", "id": "v1.0", "links": [{"href": "http://10.106.48.10:8000/v1/", "rel": "self"}]}]}
15:35 angdraug joined #fuel
15:36 newbism_ cleared cache.. worked
15:37 blahRus joined #fuel
15:37 championofcyrodi this seems to be helping me out: http://docs.openstack.org/havana/config-reference/content/demo_multiple_operation.html
15:38 newbism_ there is networks net04 and net04_ext ... can I use either of them depending if I want the VM to be nat or not?
15:40 championofcyrodi if you use net04_ext, the VM should get an IP address from your 'public' network, aka the LAN you are using to connect to Horizon.
15:41 newbism_ cool!
15:41 championofcyrodi if you use net04, it will get a private openstack IP, and thus NAT will occur at the controller.  ( i think )
15:41 championofcyrodi i'm still learning the INs and OUTs of neutron.  pun intended.
15:42 championofcyrodi right now i'm trying to figure out how to properly/safely bounce my dhcp agent without destroying it
15:57 xarses joined #fuel
16:02 warpc__ joined #fuel
16:04 evkonst joined #fuel
16:08 jaypipes joined #fuel
16:11 evkonst joined #fuel
16:19 championofcyrodi this is frustrating.  all the openstack docs point to just restarting the neutron-dhcp-agent service... however mine is not running.  but if I start it, another neutron-dhcp-agent process starts...
16:20 championofcyrodi there is already a neutron-dhcp-agent process running, but no idea how it's being started
16:20 championofcyrodi is it just a passwordless ssh from one of the controllers?
16:20 deckardkain what are the credentials fuel sets for the openstack nodes?
16:21 championofcyrodi deckardkain: are you asking me?
16:21 deckardkain oh no
16:21 deckardkain just general question to the public
16:21 deckardkain s/public/community
16:22 championofcyrodi as far as i can tell, fuel uses 'root' w/ passwordless ssh to access everything.
16:22 deckardkain interesting
16:22 championofcyrodi there is also an openrc file that has openstack tenant credentials for using the openstack apis
16:22 championofcyrodi if that's what you mean
16:22 deckardkain both
16:22 deckardkain thank you
16:22 championofcyrodi so as root, i do:
16:23 championofcyrodi root@fuel ~]# ssh node-XXX
16:23 championofcyrodi root@node-XXX ~]# . openrc
16:23 championofcyrodi root@node-XXX ~] nova list
16:23 championofcyrodi which will list running instances in the 'admin' tenant, which is the default
16:26 deckardkain gotcha
16:26 deckardkain championofcyrodi: many thanks thats exactly what i was looking for
16:26 championofcyrodi your welcome.  you can also download the openrc for different tenants in the Horizon UI
16:26 championofcyrodi e.g. you create a new 'project'
16:27 championofcyrodi and want to manage that project's VM
16:27 championofcyrodi http://<horizon_host>/dashboard/project/access_and_security/
16:28 championofcyrodi API Access -> Download OpenStack RC File
16:28 ub joined #fuel
16:28 championofcyrodi which can then also be used on public side client to access swift. (provided you start the radosgw service on the controller serving requests)
16:29 ub2 joined #fuel
16:34 mattgriffin joined #fuel
16:36 championofcyrodi i can't seem to find any information on how the neutron-dhcp-agent is running w/o the init.d script being involved...
16:36 championofcyrodi openstack documentation is targeted for ubuntu distro where everything is configured manually via apt-get installs and editing configs.
16:39 championofcyrodi finally...
16:39 championofcyrodi https://bugs.launchpad.net/neutron/+bug/1257524
16:40 championofcyrodi the process is running but the service thinks its stopped...
16:40 newbism Trying to check out the VNC console... just stuck at 'loading' ...
16:40 championofcyrodi newbism: i've see that too.  not sure why that happens.
16:40 championofcyrodi at some point it just started working all the time.
16:40 newbism :|
16:41 championofcyrodi but i've been banging on this thing for almost a year now, and don't remember why
16:41 championofcyrodi https://www.softwareab.net/wordpress/openstack-debugging-neutron-dhcp/ section labeled "The Missing Link"
16:41 championofcyrodi explains why my dhcp agent is running...
16:42 championofcyrodi apparently dnsmasqs are left running when dhcp agent is killed?
17:07 ub2 joined #fuel
17:23 championofcyrodi okay i think i found it
17:23 championofcyrodi Started: [ node-46.ccri.com node-50.ccri.com node-54.ccri.com ]
17:23 championofcyrodi p_neutron-dhcp-agent(ocf::mirantis:neutron-agent-dhcp):Started node-50.ccri.com
17:23 championofcyrodi pcs status!
17:57 championofcyrodi blah... finally got it sorted... kind of
17:57 championofcyrodi documented here: http://championofcyrodiil.blogspot.com/2015/06/openstack-dhcp-lease-renewing-every.html
17:58 championofcyrodi however, if new compute nodes are added, I'm assuming they will get the default 120 second lease duration, and of course if the dhcp agent migrates to that node, the lease will go back to getting renewed every 60 seconds.
19:37 teran joined #fuel
19:48 HeOS joined #fuel
19:55 e0ne joined #fuel
21:53 angdraug joined #fuel
22:22 aleksandr_null joined #fuel
22:30 blahRus joined #fuel
22:49 jaycee joined #fuel
22:50 jaycee is it possible to mix kvm and lxc hypervisors in the same fuel cloud?
23:20 rmoe joined #fuel
23:34 angdraug no
23:35 angdraug at least Fuel won't set it up for you out of the box

| Channels | #fuel index | Today | | Search | Google Search | Plain-Text | summary