Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2015-01-21

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

All times shown according to UTC.

Time Nick Message
00:31 benrodrigue joined #fuel
00:50 e0ne joined #fuel
01:25 rmoe joined #fuel
01:52 mattgriffin joined #fuel
02:01 emagana joined #fuel
02:55 Longgeek joined #fuel
03:01 xarses joined #fuel
03:51 fandi joined #fuel
04:28 geekinutah joined #fuel
04:44 thumpba joined #fuel
04:50 thumpba_ joined #fuel
04:52 thumpba joined #fuel
05:03 thumpba joined #fuel
05:09 mattgriffin joined #fuel
05:30 emagana joined #fuel
05:32 mattgriffin joined #fuel
05:49 sressot joined #fuel
07:08 saibarspeis joined #fuel
07:10 sambork joined #fuel
07:15 corepb_ joined #fuel
07:22 dklepikov joined #fuel
07:52 stamak joined #fuel
08:05 sambork joined #fuel
08:17 benrodrigue joined #fuel
08:17 kupo24z joined #fuel
08:25 dkusidlo joined #fuel
08:27 ddmitriev joined #fuel
08:46 dkusidlo joined #fuel
08:50 ChrisNBlum joined #fuel
08:51 mescanef joined #fuel
08:51 dkusidlo joined #fuel
08:53 thumpba joined #fuel
08:53 hyperbaba joined #fuel
09:10 HeOS joined #fuel
09:14 avgoor joined #fuel
09:14 hyperbaba Hi, is fuel 5 going to be updated in the future? I'm reluctant to switch to juno , so i wanna stay on icehouse for now?
09:18 robklg joined #fuel
09:20 fandi joined #fuel
09:32 monester_laptop joined #fuel
09:33 evg hyperbaba: Hi, yes, 5.0 and 5.1 branches continue
09:34 dkusidlo joined #fuel
09:35 avgoor joined #fuel
09:38 tatyana joined #fuel
09:38 avgoor joined #fuel
09:57 samuelbartel_ joined #fuel
09:57 samuelbartel_ hello
09:58 samuelbartel_ i am facing this bug on my fuel master node
09:58 samuelbartel_ https://bugs.launchpad.net/fuel/+bug/1405686
09:59 samuelbartel_ how can I update nailgun config without  re install
10:03 e0ne joined #fuel
10:07 omolchanov joined #fuel
10:14 dmellado joined #fuel
10:20 DaveJ__ joined #fuel
10:22 DaveJ__ Hi
10:23 DaveJ__ I'd like to deploy multiple regions with Fuel - is that possible, or are there steps to follow after deploying to seperate fuel clouds to allow me to setup regions ?
10:47 baloney_ joined #fuel
10:55 evg samuelbartel_: hello
10:55 samuelbartel_ evg: Hello
10:55 baloney_ hi
10:55 samuelbartel_ it was an issue with an iso provided by support team
10:55 samuelbartel_ I am currently reinstalling a standard 6.0 iso
10:56 samuelbartel_ i thought it would be faster than correct the issu
10:56 evg samuelbartel_: I know, i'm exploying it now too
10:56 samuelbartel_ install is going on
10:56 evg samuelbartel_: sorry?
10:57 evg samuelbartel_: what do you mean? It doesn't work
10:57 samuelbartel_ evg: do you manage to solve the issue by modifying maybe nailgun configuration
10:57 samuelbartel_ with the iso
10:58 evg samuelbartel_: ebout the bug
10:58 samuelbartel_ with the iso i had nailgun was not able to connect to keystone
10:58 samuelbartel_ i didn't find out how to solve it
10:59 evg samuelbartel_: edit /etc/puppet/2014.2-6.0/modules/nailgun/templates/nginx_nailgun.conf.erb according to the fix
10:59 evg samuelbartel_: delete this line "proxy_set_header Host $host"
10:59 samuelbartel_ yes I saw the fix
10:59 evg samuelbartel_: then "dockerctl destroy nginx"
10:59 samuelbartel_ but it wont solve the problem on a deployed fuel master node isn't it
11:00 samuelbartel_ oh ok
11:00 samuelbartel_ by destroying it and re creating it
11:00 evg samuelbartel_: or more quickly
11:00 samuelbartel_ it is going to reexecute puppet part
11:00 samuelbartel_ i got it
11:00 evg samuelbartel_: dockerctl shell nginx
11:01 evg samuelbartel_: edit the file /etc/nginx/conf.d/nailgun.conf
11:01 evg samuelbartel_: kill -1 "nginx root process"
11:02 evg samuelbartel_: I've just done so
11:02 samuelbartel_ evg: ok thank you very much!
11:03 baloney_ Hi everybody. I have an issue with the Fuel 6.0 iso setup (minimal virtualbox stack). Basically, the vbox nodes all have a GenuineIntel CPU vendor, but the qemu instances have an AuthenticAMD. This is causing my OS (Android) to fail the boot. I tried to specify model/vendor in libvirt/instance*.xml but it had no effect
11:04 evg samuelbartel_: making complete iso you need will take some time, we've got some problems here
11:05 baloney_ I have an other cluster in my enterprise not using fuel, which correctly sets the vendor as GenuineIntel with the same images
11:05 evg samuelbartel_: I'm going to just build new bootstrap and ubuntu install image for you
11:06 samuelbartel_ ok
11:06 samuelbartel_ so I will be able to update the boostrap image and ubuntu image on my fuel master node
11:07 evg samuelbartel_: yes
11:07 samuelbartel_ and then reboot my DL360 servers in order to reboostrap them with the RAID controller?
11:07 samuelbartel_ seems ok for me
11:08 samuelbartel_ we will be able to test the driver without needing to wait for the iso
11:08 samuelbartel_ wll be faster on that way
11:10 evg samuelbartel_: yes, but if i'm success. And, btw, we haven't tested install process yet
11:10 samuelbartel_ evg: ok. let me know when you will have these image avaialable
11:11 evg samuelbartel_: yes, I hope today
11:22 sambork joined #fuel
11:43 baloney_ joined #fuel
12:16 dkusidlo joined #fuel
12:16 saibarspeis joined #fuel
12:24 dkusidlo_ joined #fuel
12:27 azemlyanov joined #fuel
13:07 sambork joined #fuel
13:07 Longgeek joined #fuel
13:09 dkusidlo joined #fuel
13:25 benrodrigue joined #fuel
13:49 Miouge joined #fuel
13:58 ykotko joined #fuel
14:07 corepb_ joined #fuel
14:09 Longgeek joined #fuel
14:22 saibarspeis Did anyone tried to add a compute node to an existing environment on a running Fuel 5.1 with ubuntu icehouse openstack cluster ?
14:22 saibarspeis I am going to do this in a datacenter to an openstack cluster deployed with Fuel and I am afraid some things will be missing (like network components)
14:26 benrodrigue joined #fuel
14:26 MiroslavAnashkin If you did not added custom components in scope of that existing environment installation - then should not be any issues. BTW, failure of a compute node deployment does not affect working environment at all.
14:29 mattgriffin joined #fuel
14:35 e0ne joined #fuel
14:39 geekinutah joined #fuel
14:55 akupko joined #fuel
14:56 e0ne joined #fuel
15:10 Miouge joined #fuel
15:22 amalinow_ joined #fuel
15:28 amalinow_ In Fuel 6.0 Release Notes we can find "Image base provisioning" as new, experimental feature. Does it mean that using Fuel I can deploy node using my custom image? Or is it posible only for images certified/delivered by Mirantis?
15:34 dkusidlo joined #fuel
15:36 samuelBartel joined #fuel
15:39 blahRus joined #fuel
15:43 mpetason joined #fuel
15:43 corepb_ joined #fuel
15:53 baloney_ joined #fuel
15:53 samuelBartel Hello
15:54 samuelBartel I saw an openstack mailing list there is now a community website
15:54 samuelBartel https://www.fuel-infra.org/
15:54 samuelBartel is there any difference between the Fuel-mirantis 6.0 iso and the fuel community 6.0 iso?
15:57 fandi joined #fuel
15:57 mattgriffin joined #fuel
16:01 e0ne joined #fuel
16:02 jaypipes joined #fuel
16:17 bookwar samuelBartel: the build is different by the value of FEATURE_GROUPS parameter
16:19 bookwar samuelBartel: in practice it means that community build have no mirantis branding
16:24 samuelBartel bookwar : ok thank you
16:41 youellet joined #fuel
16:54 championofcyrodi joined #fuel
16:54 championofcyrodi So, the time has come that the 1Gbps VLAN'd switch we are using for our fuel deployment (icehouse) is too slow for our ceph reads.
16:55 championofcyrodi and we are looking to buy a 10Gbps switch, and put 10Gbps ethernet cards in all ceph nodes.
16:55 championofcyrodi Is it just the 'storage' network? or the management as well that should be moved?
16:56 championofcyrodi would like to keep the 1Gbps switch and cards currently allocated to PXE, Private, Public, etc.
17:02 emagana joined #fuel
17:09 DaveJ__ Hi guys
17:09 DaveJ__ I was trying to test out fuel, with a controller running on a vmware virtual machine, and a real compute node
17:09 DaveJ__ I'm having difficulty with the networking - does anyone know if there are issues with the vmware network drivers ?
17:18 thumpba joined #fuel
17:19 thumpba_ joined #fuel
17:19 geekinutah joined #fuel
17:38 rmoe joined #fuel
17:44 xarses joined #fuel
17:45 MiroslavAnashkin Hmm - there are different issues with different VMWare network drivers. What is your network type, Nova or Neutron, GRE or VLAN?
17:46 MiroslavAnashkin Do you use VMWare workstation or ESXi, with or without vSwitch?
17:50 evg amalinow: yes, you can use your own images as well
17:57 thumpba joined #fuel
17:57 MiroslavAnashkin championofcyrodi: Ceph traffic goes by 2 networks - management (VM to Cinder and than to Ceph) and storage - for OSD replica synchronization.
17:58 championofcyrodi so i'll need to ensure those two networks are on a 10Gbps switch.
17:58 thumpba__ joined #fuel
17:59 championofcyrodi i'm curious as to upgrading though.  if I just add the new card to the node and reboot, will it be available to configure in the fuel dashboard.
18:00 championofcyrodi do I need to stop all the services in openstack first?  compute nodes as well?
18:00 championofcyrodi seems as though a live migration would be pretty much impossible.
18:06 championofcyrodi xarses: i'm still having trouble understanding the stricthostkey checking issue with my nova user.  on my controllers the user has the login /bin/false, so I am not able to replicate the ssh WARNING via command line like, sudo su - nova -c "ssh ..., it just exits (/bin/false) and nothing runs. commands like ls as well.
18:07 championofcyrodi also that user (nova) has no .ssh folder.  so i can't verify the config you pasted here: http://paste.openstack.org/show/158981/
18:07 championofcyrodi hmmm
18:08 championofcyrodi maybe UserKnownHostsFile=/dev/null is the problem... in that there is no equal sign in your paste, but there is in my config
18:08 championofcyrodi for root on the controllers
18:08 championofcyrodi per this conversation: http://irclog.perlgeek.de/fuel/2015-01-19#i_9968775
18:12 xarses championofcyrodi: you are using resize with ceph ephemeral storage?
18:12 championofcyrodi yes
18:13 championofcyrodi well, the volume was created with the nova boot command, with preserve after shutdown set to true.
18:13 championofcyrodi so the volume is visible in the 'volumes' table in the horizon ui.
18:13 xarses it just won't work, according to a co-worker, the code path is wrong for ceph
18:14 championofcyrodi hmm..  what if i just terminate the instance, and reboot another instance with more RAM from the same volume?
18:14 rmoe resize will work as long as the resize is scheduled where the host is currently running
18:14 MiroslavAnashkin If you change the network card - it will appear in master node, but you probably have to re-deploy your nodes with changed NICs
18:14 rmoe if it moves it to another compute host it will fail because the nova user has /bin/false as its shell
18:16 championofcyrodi rmoe: ^ thinking about that for a bit.
18:16 championofcyrodi (i am)
18:16 rmoe I should say it will fail on 5.1, this isn't an issue on 6.0 or release previous to 5.1 (I believe)
18:20 championofcyrodi weird.  just realized teh SSH error is connecting to server on 10.20.0.9.  but looking at fuel networks, i dont se a 10.20.x.x subnet.
18:21 championofcyrodi management and storage are 10.10.20 and 10.10.30...
18:24 championofcyrodi so the instance i am resizing does NOT have an ephemeral storage device. but instead is using an attached volume to boot from.
18:24 championofcyrodi but we are using ceph for both ephemeral storage and volumes.
18:26 rmoe yes, this issue will occur any time the resize moves the instance from one host to the other
18:26 rmoe as part of that process nova attempts to create the instance directory on the new copute host
18:26 rmoe that's what the ssh mkdir ... command is doing
18:26 rmoe that fails because the nova user on the new compute host doesn't have a real shell
18:26 rmoe if you change the nova user's shell on the compute hosts the resize will work
18:27 championofcyrodi any preference on the shell i should change it to?
18:27 championofcyrodi /bin/bash ?
18:27 rmoe that will work fine
18:28 championofcyrodi reviewing the settings for our installation, it looks like we are using CEPH RBD for volumes (cinder), CEPH RBD for images (glance), CEPH RBD for ephemeral volumes (nova), and Ceph RadosGW for objects. (swift)
18:28 jobewan joined #fuel
18:28 championofcyrodi looking at this i think i understand the diff between ephemeral storage and volume storage in the fuel context.
18:30 e0ne joined #fuel
18:32 championofcyrodi is there a puppet configuration i can modify to support updating the nova shell when adding new nodes?
18:33 angdraug joined #fuel
18:46 championofcyrodi rmoe:  you are 100% correct.  setting a shell for the nova user on all nodes fixed the issue
18:46 rmoe good deal
18:46 rmoe I'm testing a puppet change for this right now
18:47 rmoe should be a trivial fix
18:47 championofcyrodi should i create a bug? https://bugs.launchpad.net/fuel
18:47 championofcyrodi didnt see anything related to this when searching for 'nova shell'
18:47 championofcyrodi (in launchpad)
18:48 championofcyrodi this is the closest bug i could find https://bugs.launchpad.net/fuel/+bug/1316812
18:48 championofcyrodi related to upstart failing.
18:51 rmoe I'll create the bug for it
18:51 championofcyrodi can you link me on it when you do?
18:51 rmoe yep
18:51 championofcyrodi ty
18:55 blahRus1 joined #fuel
18:59 monester_laptop joined #fuel
19:00 thumpba joined #fuel
19:08 geekinutah joined #fuel
19:08 rmoe championofcyrodi: https://bugs.launchpad.net/fuel/+bug/1413339
19:08 monester_laptop joined #fuel
19:13 ni291187 joined #fuel
19:44 HeOS joined #fuel
19:53 xarses joined #fuel
20:01 thumpba_ joined #fuel
20:40 e0ne joined #fuel
20:47 mattgriffin joined #fuel
20:48 fandi joined #fuel
20:50 rbowen joined #fuel
20:53 fandi joined #fuel
21:03 ilbot3 joined #fuel
21:03 Topic for #fuel is now Fuel 5.1.1 (Icehouse) and Fuel 6.0 (Juno) https://software.mirantis.com | Fuel for Openstack: https://wiki.openstack.org/wiki/Fuel | Paste here http://paste.openstack.org/ | IRC logs http://irclog.perlgeek.de/fuel/
21:47 fandi joined #fuel
22:07 thumpba joined #fuel
22:11 thumpba_ joined #fuel
22:14 championofcyrodi if I update the instance XML on a hypervisor.. like /etc/libvirt/qemu/instance-0000001b.xml, for a specific instance.  will performing a hard/soft reboot re-read the configuration? or do i need to perfrom a rebuild?
22:36 Longgeek joined #fuel
22:49 championofcyrodi left #fuel
23:06 monester_laptop joined #fuel
23:12 benrodrigue joined #fuel
23:16 geekinutah joined #fuel

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