Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-08-06

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

All times shown according to UTC.

Time Nick Message
00:34 rmoe joined #fuel
01:16 IlyaE joined #fuel
01:28 Samos123 tar: ./bin/gunzip: Cannot hard link to `./bin/uncompress': Operation not permitted
01:28 Samos123 seeing these kind of issues on make iso now
01:29 Samos123 and in the end E: Unmet dependencies. Try 'apt-get -f install' with no packages (or specify a solution).
01:32 Samos123 longer log of what's happening: http://paste.openstack.org/show/90830/
01:32 mattgriffin joined #fuel
01:34 Samos123 I am running as root even now so Operation not permitted shouldn't be an issue. And the Operating system is 12.04 and setup using prepare-build-env.sh script
01:47 ilbot3 joined #fuel
01:47 Topic for #fuel is now Fuel 5.0 for Openstack: https://wiki.openstack.org/wiki/Fuel | Paste here http://paste.openstack.org/ | IRC logs http://irclog.perlgeek.de/fuel/
01:53 IlyaE joined #fuel
02:33 IlyaE joined #fuel
02:58 KLonHJ joined #fuel
02:58 KLonHJ Hi
02:59 KLonHJ is there anyone online?
03:06 mattgriffin joined #fuel
03:09 KLonHJ Hi!I have added something to the fuel WebUI, but it seemed not work! So where can I find the .html file?
03:57 xarses joined #fuel
03:58 Longgeek joined #fuel
04:30 jaypipes joined #fuel
04:31 ArminderS joined #fuel
04:40 christopheraedo KLonHJ: you can look at the "Adding Zabbix Role" example here to see how the UI components are modified - http://docs.mirantis.com/fuel-dev/develop/addition_examples.html
04:42 christopheraedo Also, did you modify the UI in an already-deployed Fuel controller, or did you make changes in fuel-web and then create a new ISO with a pointer to your modified commit?
04:51 IlyaE joined #fuel
04:52 christopheraedo Samos123: two things I would suggest - one, check to be sure all the packages that were supposed to be installed by prepare-build-env were actually installed
04:56 christopheraedo Samos123: then try a make deep_clean; make iso
04:58 scroiset joined #fuel
05:00 KLonHJ I changed the UI in an already-deployed Fuel controller
05:14 Samos123 christopheraedo: Thanks for tips, I just ran prepare-build-env again but seems everything was installed correctly except this http://paste.openstack.org/show/90888/ but should be fine because it's documentation only?
05:14 Samos123 I think it may be because of virtualbox drive sharing
05:14 Samos123 let me try to not run it from the shared virtualbox folder
05:16 Samos123 https://www.virtualbox.org/ticket/4290 and https://www.virtualbox.org/ticket/818
05:35 christopheraedo Samos123: Good call, odds are pretty good that's the issue. Let us know what you run in to though here if you hit any more snags, someone should be able to help.
05:36 christopheraedo KLonHJ: if you're modifying an already deployed controller, you might have to re-build a docker container (if it's v5). There are a few folks on the channel who have more experience doing live mods though - anyone else have pointers on that?
05:39 IlyaE joined #fuel
05:39 KLonHJ I know that v5 is running on a docker contrainer, but our fuel version is v4.1.1, so how to solve the problem?
05:42 christopheraedo Sorry I'm not familiar enough with v4 to be of any help. Most of the US guys are offline now or going offline (like me :) ) but the folks in europe should be popping up soon, I'm betting it's pretty straightforward
05:42 christopheraedo and if you don't get a response soon, try sending a note to the fuel-dev mailing list with some details (version of fuel, which specific file you modified, etc.)
05:44 KLonHJ Ok,thanks anyway!
06:12 IlyaE joined #fuel
06:22 Samos123 christopheraedo: lol yea I got one step further, was indeed a problem with vboxfs and hard links.
06:31 Longgeek joined #fuel
06:38 Longgeek joined #fuel
06:47 e0ne joined #fuel
06:52 e0ne joined #fuel
07:01 e0ne joined #fuel
07:08 pasquier-s joined #fuel
07:12 Longgeek joined #fuel
07:15 Longgeek joined #fuel
07:16 ddmitriev joined #fuel
07:22 f13o_ joined #fuel
07:29 Samos123 maybe network issue(Chinese GFW) but make iso is stuck on Cleaning up list of fastest mirrors
07:38 lordd joined #fuel
07:38 lordd hi guys
07:38 lordd I'm checking an issue with an SDN + healthchecks
07:39 lordd so where can I see the healthcheck process, step-by-step
07:39 lordd I mean github, or a file
07:42 lordd ok, health.py on fuelclient/cli/actions/ right?
07:44 taj_ joined #fuel
07:45 lordd hum.. no
07:53 adanin joined #fuel
07:56 sc-rm joined #fuel
07:56 ArminderS- joined #fuel
08:19 Longgeek joined #fuel
08:24 lordd joined #fuel
08:25 Longgeek joined #fuel
08:30 ykotko joined #fuel
08:49 mattymo|conf joined #fuel
08:53 meow-nofer_ joined #fuel
08:54 decede joined #fuel
08:54 jaranovich_ joined #fuel
08:54 brain461 joined #fuel
08:54 Bomfunk joined #fuel
08:55 brain461 left #fuel
08:59 ykotko joined #fuel
09:00 ws joined #fuel
09:07 e0ne joined #fuel
09:10 teknomega joined #fuel
09:24 e0ne joined #fuel
09:27 brain461 joined #fuel
09:32 KLonHJ Hi guys! Is there anyone familiar with Fuel UI of v4.1.1.? I have modified a html file in an already-installed environment, then I refreshed the web-brower, but it seemed not work.
09:33 KLonHJ So I want to know where should I change the code ?
09:36 mihgen vkramskikh: ^^
09:37 KLonHJ =. =
09:38 vkramskikh Hi KLonHJ, on the master node all static files are minified and concatenated together (main.js file)
09:39 vkramskikh i suggest to check out 4.1 branch of https://github.com/stackforge/fuel-web, make changes to the files you need and copy the whole "static" dir to the master node
09:42 HeOS joined #fuel
10:30 e0ne joined #fuel
10:30 amz joined #fuel
10:30 amz hi
10:42 Longgeek joined #fuel
10:45 HeOS joined #fuel
11:03 lordd hi
11:07 tdubyk joined #fuel
11:09 lordd SSL Handshake failed :(
11:09 KLonHJ Hi vkramskikh!Thank u for asking my question! I have find the code in main.js but how is main.js generated from js/* and temlate/*? Is there any command?
11:10 Longgeek joined #fuel
11:26 vkramskikh KLonHJ: yep, the command is "grunt build"
11:26 vkramskikh https://github.com/stackforge/fuel-web/blob/master/docs/develop/nailgun/development/env.rst#running-nailgun-in-fake-mode
11:46 amz a
12:16 taj_ joined #fuel
12:28 mattgriffin joined #fuel
12:38 HeOS joined #fuel
12:58 HeOS joined #fuel
13:42 lordd joined #fuel
13:52 IlyaE joined #fuel
13:57 taj_ joined #fuel
13:57 HeOS joined #fuel
14:01 mattgriffin joined #fuel
14:32 e0ne joined #fuel
14:39 Longgeek_ joined #fuel
15:15 xarses joined #fuel
15:22 lordd joined #fuel
15:27 blahRus joined #fuel
15:40 HeOS joined #fuel
15:58 ArminderS joined #fuel
16:21 angdraug joined #fuel
16:35 teknomega if i go from 2 controllers to 1 controller in Fuel when i remove a controller... do i have to do some clean up work to get the controller to function properly ?
17:19 IlyaE joined #fuel
17:36 BillTheKat joined #fuel
17:47 tatyana joined #fuel
17:54 BillTheKat I have a fuel 50 server on a bare metal server and I created VMs on vbox. The fuel server sees them and when deploying loads the OS as start the puppet deployment - but soon into the puppet deployment they go off-line and never come back - The servers are up - but I can not ping them all of a sudden.
18:05 KodiakF BillTheKat:  I suspect fuel is changing network ifaces which is knocking them offline.  You may want to look at locally cached puppet logs on the dead hosts to confirm this.
18:06 BillTheKat KodiakF - I am retrying it - I may be mistaking - about not being able to ping them once they go off line.  Will know in a bit :)
18:07 KodiakF Jump on the local console and tail -f /var/log/messages to see the puppet apply
18:07 KodiakF the local console of one of the dying VMs that is
18:09 BillTheKat KodiakF: will do
18:10 KodiakF Lemme know how it turns out; I'm mixing a rational hypothesis not from my own experience w/ Fuel, along with prior experience w/ Puppet.
18:18 dwangoAC joined #fuel
18:18 HeOS joined #fuel
18:19 dwangoAC Greetings! I am trying to determine if Fuel can do an initial "All-in-one" install (controller and compute on the same physical server) and add compute nodes later.  Is this possible?
18:23 BillTheKat KodiakF: The tail showed no puppet at all.... Hmmm  - one of my co-workers suggested setting the interface type to PCnet-FAST III in vbox
18:23 dwangoAC KodiakF: By the way, thank you for directing me toward Fuel
18:24 KodiakF Hey no worries it looks pretty rad can't wait to start using it at work - gotta finish cluster v1 first though
18:24 dwangoAC KodiakF: I still don't know if I can make it do what I want as I can't replace our existing DHCP or PXE servers but it seems very interesting indeed
18:24 KodiakF maybe puppet isn't logging to /var/log/messages - maybe it's putting stuff in /var/log/puppet
18:37 kupo24z joined #fuel
18:38 kupo24z angdraug: xarses: is there a somewhat painless way to get fuel to deploy on ubuntu 14LTS?
18:40 MiroslavAnashkin kupo24z: No, there are only ways full of pain. You have to re-build all the local Ubuntu repository packages against 14.04 for the start.
18:42 MiroslavAnashkin dwangoAC:  "All-in-one" install is not supported by Fuel starting from Fuel 3.0. This setup is incompatible with HA Neutron networking scheme we use in Fuel.
18:43 dwangoAC MiroslavAnashkin: I see - my need is for internal servers (not public facing) and the ability to add compute nodes as we add more instances of the thing we're hosting internally
18:43 dwangoAC And for proof-of-concept I only have one server at the moment, although I could put two others to use if I had to
18:43 dwangoAC But that still only gives me 3, and the official Ubuntu documentation seems to insist on a minimum of 10 (!) servers.
18:45 dwangoAC I looked at StackGeek and tried it out, it seemed to do the job of getting OpenStack itself installed but without any network configuration and it ultimately caused Horizon to fail to load so I gave up on that route (and KodiakF discouraged its use for reasonable reasons)
18:45 MiroslavAnashkin For the proof-of-concept you may use virtual machines. Virtualbox>=4.2.2, VMWare, HyperV and KVM work fine with Fuel.
18:46 dwangoAC Err.. not quite what I mean - I have a bare metal server that I want to make a controller + compute setup and fire off, for starters, three instances (i.e. 3 KVM guests)
18:46 dwangoAC And in the future, as we add more instances, be able to have a controlled method to add another compute node
18:47 xarses kupo24z: You can change your ubuntu repo to point to an external 14.04 it will go out and try to grab packages, however in some cases we ask for specific versions so you would have to fix any of those issues. You may find that our Icehouse packages might further cause this, and lastly you might not get a working deployment anyways, we usually end up with a two to three weeks of bugs when we change OS versions. So there will be pain, but it depends on your t
18:47 xarses ollerance.
18:47 dwangoAC My research so far suggests that what I'm looking for doesn't exist - there's DevStack (completely unacceptable for anything except testing) and then there's High Availability, but not much "in between"
18:47 MiroslavAnashkin VirtualBox is the easiest way. KVM has the best networking part to integrate with external hardware.
18:48 dwangoAC MiroslavAnashkin: I'd be using KVM for certain - bad experiences with VirtualBox in the past (my nethack-tas-tools project on Gitorious abandoned it years ago)
18:48 xarses kupo24z: now that said, you might find it easier to deploy 14.04 your self and skip provisioning, or maybe hack on cobbler so that it works from the alternate source
18:49 IlyaE joined #fuel
18:53 decede Whats the point of having multiple OpenStack environments in fuel?
18:54 decede you need all the environments on the same L2 for deployment right so you wouldn't have multiple sites use the same fuel controller
18:55 xarses decede: correct, It's possible that you would have multiple seperate stacks in the same L2, for me it would be a dev, test, stage type deal
18:55 MiroslavAnashkin BillTheKat: PCnet-FAST III is recommended. And allow promiscuous mode for all VBox virtual adapters - it is mandatory.
18:56 decede xarses: ah ok
18:57 xarses decede: also when multiple-cluster-networks (https://blueprints.launchpad.net/fuel/+spec/multiple-cluster-networks) fuel will support multiple L3 segments and deploying across them
18:57 BillTheKat MiroslavAnashkin: I tried PCnet-FAST III but my external fuel server did not see the nodes
18:57 xarses its currently targeted to 6.0
18:57 xarses ~November
18:58 xarses also the current mode is necessary for our development cycle as it allows us to deploy different configurations from the same master node
18:59 xarses saves a lot of time
18:59 decede xarses: ah that looks interesting
19:01 xarses dwangoAC: fuel requires a minimum of 2 nodes (1 compute, 1 controller) however as you scale out you will not have an easy time converting the environment to HA, you would have a better experience deploying a second cluster
19:01 dwangoAC xarses: That's sensible
19:02 dwangoAC xarses: If I have two servers and one of them is beefier than the other which should I dedicate to compute?  I'm edging toward the less beefy, as it doesn't seem like the controller is hosting any actual instances.
19:02 dwangoAC And there will never be a need for HA - these are internal only, for development use, and if one goes down it's not the end of the world.
19:03 dwangoAC (Luckily - I don't take that freedom for granted! :)
19:03 xarses dwangoAC: the smaller node to the controller. for example we run all of our tests on 1.5 or 2GB nodes (computes or controllers)
19:04 dwangoAC I see
19:04 dwangoAC xarses: That's not much at all, truth be told...  Hmm... maybe I can make this work after all
19:04 xarses for 1 or 2 dozen machines its large enough
19:04 xarses (machines == instances)
19:04 dwangoAC xarses: Is there any documentation about how to get Fuel to behave when it isn't the DHCP server and PXE server?
19:04 dwangoAC xarses: That's perfect for our scale needs for the forseeable future
19:05 kupo24z xarses: This is for a newer libguestfs, i think it will be easier for me to just try to compile a 12.04 binary of the newer version
19:05 xarses an 8 gb quad core controller should be enough for 2-4 computes easily
19:05 dwangoAC (I value the time it takes to respond to someone personally on an IRC channel so I'm trying to do as much research as I can outside of pestering people)
19:06 xarses dwangoAC: so we use DHCP on the fuel Admin network. the Arch guide should discuss this
19:06 dwangoAC OK
19:06 xarses fuel must be the PXE server on that network, you could with some work use another DHCP server on the network
19:07 xarses this is because we use dhcp/pxe/cobbler to discover and provision the nodes
19:07 dwangoAC xarses: That's the sticking point - we have a PXE server with several purpose-built automated installers (some for CentOS, some for Ubuntu / Debian) - I can't touch that infrastructure under any circumstances, but I can add extra entries to that PXE server
19:08 xarses the network can be completely isolated from anywhere else, the fuel node does not need internet access
19:09 dwangoAC xarses: I could do some creative lab networking to isolate one network interface of the servers I plan on using but that'll get tricky when we spin up an instance in our other lab in the other building
19:09 dwangoAC Hmm.. concerns me that I won't be able to scale well... is there a way to put a wrapper around an existing PXE server to say "Hey, click on this entry to pass the PXE boot attempt off to another server"?
19:10 dwangoAC This is *way* outside the context of Fuel at this point, so I'll start doing more research
19:10 xarses cobbler can't provision the nodes with out being able to control the PXE defaults, if your PXE is cobbler, then you could possibly use an external server but again it would be a bear
19:10 dwangoAC Hmm...  This is overall a sensible architecture, it's just we've been using PXE for years and I can't move that integrated infrastructure
19:10 xarses it could be made to live with out control of DHCP but again, cobbler expects to manage both, and fuel relies on cobbler
19:13 xarses dwangoAC: not to infer that you can, but if i where you I'd ask my network guys for a new network with out DHCP relay/forwarding to your other PXE server. Also as an aside and to the conversation above with decede, fuel can't deploy across separate L2 domains, so if the other building has a different IP range you would probably need another fuel node over there too.
19:13 dwangoAC xarses: Luckily, IP range is congruant
19:13 dwangoAC ...not sure that was the right word there
19:14 dwangoAC xarses: And, yes, we can manage ports off of a router to say that certain ports don't do DHCP and send those to particular top-of-rack switches - not my first choice but it could be done
19:14 xarses close enough. so it might be the same L2, you should still defier to the network guys
19:16 xarses kupo24z: ya, if thats all you want, rebuilding it is likely easier
19:21 turtlemansam joined #fuel
19:22 turtlemansam joined #fuel
19:22 IlyaE joined #fuel
19:23 decede should the fixed network have access to the internet?
19:27 taj joined #fuel
19:27 decede if it matters its a simple FlatDHCP network mode environment
19:28 xarses fixed no, thats your virtual network range for instances
19:28 decede I currently have internet access configured to the fuel network and the public net but im seeing issues pinging 8.8.8.8(configured DNS)
19:28 xarses floating ip ranges on the otherhand, yes. It needs to be a subset of the public range
19:28 decede xarses: ah ok cool must have an issue in my network setup then
19:29 turtlemansam joined #fuel
19:29 xarses make sure to evaluate the interfaces assigned to each node, it's likely that the public interface role is on the wrong physical interface
19:30 decede ye I forgot to change from the default but its deployed tagged access so it should be working
19:38 decede the network will be a little more contended then intended but should still work
19:39 e0ne joined #fuel
19:49 BillTheKat KodiakF:  MiroslavAnashkin: my Vbox issue was I needed to specify promiscuous mode - ugh ..... Should have known that :)
19:55 adanin joined #fuel
19:57 xarses_ joined #fuel
20:06 e0ne joined #fuel
20:08 nucccc joined #fuel
20:09 nucccc Any one have good experience with Fuel deployment and Murano ?
20:09 nucccc Deployment stuck on Murano_rabbitmq/exec executed successfully ...
20:09 nucccc Everytime.
20:21 decede how is the fuel network dhcp configured? my setup seems to be giving out the wrong default gateway ip
20:37 adanin joined #fuel
20:56 IlyaE joined #fuel
21:02 IlyaE joined #fuel
21:02 e0ne joined #fuel
21:07 KodiakF2 joined #fuel
21:14 e0ne joined #fuel
21:38 teknoprep is there a bug when adding a 3rd controller ?
21:38 teknoprep i had 2 controllers up and running no issue... added a 3rd and now all servers are error'd out
21:42 teknoprep the node number changed from 220 to 217 or vice versa
21:42 teknoprep i could see that being an issue
21:58 KodiakF2 Someone posted a dev environment setup guide for Fedora/KVM recently on here and I lost the link - anyone know what I'm talking about and feeling charitable?
21:59 KodiakF2 NM NM NM it's at the bottom of the dev guide under the heading devops
22:22 f13o_ joined #fuel
22:23 geekinutah joined #fuel
22:40 KodiakF3 joined #fuel
22:42 IlyaE joined #fuel

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