Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-08-12

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

All times shown according to UTC.

Time Nick Message
00:03 KodiakF2 joined #fuel
00:10 e0ne joined #fuel
00:17 e0ne joined #fuel
01:04 xarses joined #fuel
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/
02:17 Samos123 MiroslavAnashkin: great thanks!! Will try it later
02:56 Samos123 MiroslavAnashkin: I meant a public jenkins job that can be triggered by public users giving specific gerrit reviews / github branches, the job would generate an ISO this way not every user has to create his own make iso environment. I already saw the master daily build.
03:41 mattgriffin joined #fuel
04:21 ArminderS joined #fuel
05:07 Longgeek joined #fuel
05:14 jaypipes joined #fuel
05:21 aleksandr_null joined #fuel
05:58 e0ne joined #fuel
06:01 adanin joined #fuel
06:02 e0ne joined #fuel
06:16 e0ne joined #fuel
06:21 Longgeek joined #fuel
06:35 e0ne joined #fuel
06:39 lordd_ joined #fuel
06:44 e0ne joined #fuel
06:51 e0ne joined #fuel
06:54 kaliya joined #fuel
06:57 pasquier-s joined #fuel
06:59 saibarspeis joined #fuel
07:00 e0ne joined #fuel
07:01 lordd_ joined #fuel
07:08 e0ne joined #fuel
07:14 e0ne joined #fuel
07:19 e0ne joined #fuel
07:26 e0ne joined #fuel
07:34 e0ne joined #fuel
07:37 HeOS joined #fuel
07:40 e0ne joined #fuel
07:46 e0ne joined #fuel
07:52 e0ne joined #fuel
08:02 e0ne joined #fuel
08:06 Krast joined #fuel
08:10 e0ne joined #fuel
08:16 saibarspeis joined #fuel
08:17 e0ne joined #fuel
08:39 e0ne joined #fuel
08:40 Krast joined #fuel
08:49 sikor_sxe joined #fuel
08:57 e0ne joined #fuel
09:03 e0ne joined #fuel
09:13 e0ne joined #fuel
09:19 e0ne joined #fuel
09:26 e0ne joined #fuel
09:30 geekinut1h joined #fuel
09:34 e0ne joined #fuel
09:44 e0ne joined #fuel
09:50 wayneeseguin joined #fuel
10:01 wayneeseguin joined #fuel
10:15 adanin joined #fuel
10:29 akupko joined #fuel
10:35 DaveJ__ joined #fuel
10:43 odyssey4me joined #fuel
10:49 odyssey4me mattymo - you around?
10:49 mattymo odyssey4me, you are in luck
10:49 odyssey4me :) good afternoon then - I wonder if you can help me understand something, or point me to someone who can assist
10:49 mattymo sure
10:50 odyssey4me In a standard Fuel 5.x deployment, I see that there are several floating IP's allocated to the admin project - but not assigned.
10:50 odyssey4me The number appears to match the number of fuel nodes registered with my fuel master (not those allocated to the environment).
10:51 odyssey4me Any idea why that is?
10:52 mattymo you mean within nova/neutron on public range?
10:53 odyssey4me Sorry - Neutron/GRE deployment. And they're Floating/External addresses which in a standard Fuel deployment is in the Public CIDR, yes.
10:53 odyssey4me However, these are not addresses that are assigned to the nodes as far as I can see.
10:54 mattymo there is one for every node, I believe
10:54 mattymo in env
10:54 mattymo but I'm not an expert on that area
10:54 mattymo better to file a bug
10:54 odyssey4me alright, will do
10:55 odyssey4me Another question - why is is that each controller gets an IP address in the 'public' range? (ie not a floating address, but a 'Public' address)
10:56 mattymo again no(
10:57 odyssey4me alright - I'll file bugs to query this
10:57 odyssey4me The floating IP association to the admin project makes absolutely no sense to me.
10:58 odyssey4me The public address assignment to each controllers make a little sense - but it's actually a waste of public addresses and an increase in security risk.
11:03 odyssey4me Floating IP association: https://bugs.launchpad.net/fuel/+bug/1355745
11:10 Longgeek joined #fuel
11:11 odyssey4me aha, just figured out the second question
11:11 odyssey4me each node has an ip address on the public network in order for the iptables rules used in the rabbit clustering to work
11:12 odyssey4me the haproxy floating ip is in a namespace on the controller - the controller then has an iptables rule to forward the appropriate traffic to the controller's ip
11:26 HeOS joined #fuel
11:34 Longgeek joined #fuel
11:34 tuvenen joined #fuel
11:37 brain461 joined #fuel
11:37 brain461 left #fuel
11:38 pasquier-s joined #fuel
11:41 odyssey4me correction - it routes from the host to the namespace :p
12:10 Eucke joined #fuel
12:18 Eucke left #fuel
12:19 Eucke joined #fuel
12:19 lordd_ joined #fuel
12:42 twforeman joined #fuel
12:44 getup- joined #fuel
12:44 brain461 joined #fuel
13:01 e0ne joined #fuel
13:10 CephFan1 joined #fuel
13:14 lordd_ joined #fuel
13:15 BillTheKat joined #fuel
13:21 Baboune joined #fuel
13:21 Baboune hello.. getting the following exception when "adding nodes" to an env: [7f4c6126d700] (logger) Traceback (most recent call last):   File "/usr/lib/python2.6/site-packages/web/application.py", line 239, in process     return self.handle()   File "/usr/lib/python2.6/site-packages/web/application.py", line 230, in handle     return self._delegate(fn, self.fvars, args)   File "/usr/lib/python2.6/site-packages/web/application.py", line 420,
13:21 Baboune any ideas?
13:22 Baboune http://pastebin.com/QwUMKxJi
13:22 Baboune File "/usr/lib/python2.6/site-packages/nailgun/network/manager.py", line 886, in _get_interface_by_network_name     '{1}'.format(network_name, node.full_name)) CanNotFindInterface: Cannot find interface by name "private" for node: Blade1 (2c:54) (id=28, mac=00:0a:f7:2b:2c:54)
13:24 getup- joined #fuel
13:33 teknomega joined #fuel
13:33 teknomega hi all
13:33 teknomega i have everthing up and working within openstack itself using mirantis fuel
13:33 teknomega now i have a dilema where ... internal / external networking doesn't work at all for the VM's
13:34 teknomega my external subnet is a WAN subnet routed to our Core Router from our ISP at the DC
13:34 teknomega 1.2.3.0/24 - WAN RANGE / and we then used a random 10.150.0.0/16 for our Private range
13:34 teknomega inside OS
13:35 teknomega i can ping every node that was provisioned with a WAN IP from the fuel OS configurations... so all of our compute nodes have a 1.2.3.x address and they can communicate with the real world
13:35 teknomega any idea why this may happen or what might be going on... we are using Neutron with VLAN's
13:36 odyssey4me tekomega - can your instance ping its gateway?
13:38 MiroslavAnashkin Baboune: Please try to reboot your Blade1. Then, please check if it boots from one interface only. There are servers, which attempts to boot from several network interface. Such settings will not work with Cobbler.
14:07 akupko joined #fuel
14:07 xarses joined #fuel
14:16 Baboune teknomega do u use VLAN? and is the tagging properly set in the router/switch?
14:22 e0ne joined #fuel
14:29 kaliya joined #fuel
14:39 xarses joined #fuel
14:49 odyssey4me_ joined #fuel
14:52 odyssey4me__ joined #fuel
14:54 akupko joined #fuel
14:59 angdraug joined #fuel
15:08 Eucke joined #fuel
15:23 lordd_ joined #fuel
15:26 blahRus joined #fuel
16:01 ArminderS joined #fuel
16:08 teknomega joined #fuel
16:09 teknoprep odyssey4me, i'll check
16:14 blahRus1 joined #fuel
16:21 teknoprep i thought the router was supplied by OpenStack ?
16:22 teknoprep the ip address of the router... in Neutron... is that not the IP address of the Virtaul Neutron Router that allows for NAT and what not
16:22 teknoprep or does that have to be a physical router for the private network
16:22 teknoprep i have a WAN IP router that works fine
16:30 teknoprep this doesn't make any sense
16:35 ArminderS- joined #fuel
16:36 odyssey4me teknoprep - yes, the virtual router is the project's router
16:36 teknoprep thats what i thought
16:36 odyssey4me assuming the router has an interface on the network
16:36 teknoprep yes
16:36 odyssey4me and assuming that the router has its gateway set to the public network
16:36 odyssey4me you should be able to get out
16:36 teknoprep i can't even ping between virtual server to virtual server
16:36 odyssey4me but the first start is to check whether your instance can reach the virtual router
16:36 teknoprep on the same network same everything
16:37 odyssey4me ok, then you have an l2 issue
16:37 odyssey4me you're using vlan networking, right?
16:39 teknoprep yes
16:39 jpf_ joined #fuel
16:39 teknoprep i am using vlan's 1000 - 1030
16:39 teknoprep for L2 config
16:39 teknoprep 60 for MGMT - 70 for Storage - 40 for WAN
16:39 teknoprep 10.150.0.0/16 is my Neutron L3 Config
16:40 teknoprep 10.150.1.2 is my L3 Internet network GW for Neutron
16:41 odyssey4me can you share the neutron network, router, subnet details in pastebin ?
16:42 teknoprep sure
16:42 teknoprep what do you need ?
16:42 teknoprep from Fuel ?
16:42 odyssey4me nope, from neutron
16:42 odyssey4me paste.openstack.org
16:42 teknoprep where is the neutron config at that you need ?
16:43 odyssey4me the outputs of neutron net-list, neutron router-list to start with
16:43 teknoprep You must provide a username via either --os-username or env[OS_USERNAME]
16:43 odyssey4me so from the fuel-master, ssh to a controller, then enter 'source openrc'
16:43 odyssey4me then do those commands
16:43 teknoprep why is this not setup by default for openstack fuel mirantis
16:43 teknoprep i know how to do this...just wondering why its not setup
16:44 odyssey4me teknoprep - I've only ever used neutron/gre and nova-network myself... so not entirely familiar with what's going on with you
16:44 odyssey4me it's most likely that someone simple has been broken
16:44 teknoprep http://paste.openstack.org/show/93944/
16:44 odyssey4me it's either in your actual switch config, or in the openstack config
16:45 odyssey4me and the router-list ?
16:46 teknoprep http://paste.openstack.org/show/93946/
16:47 odyssey4me ok, now for the following
16:48 teknoprep ok
16:48 odyssey4me nova list; neutron net-show 6ef3bfec-505e-4424-983a-ff0bb12facaa; neutron net-show ba1c33c1-f8db-440f-add5-5394cfd6153d; neutron router-show 491abfdc-e8d8-4527-a5fd-350378efdb15
16:49 xdeller joined #fuel
16:51 teknoprep http://paste.openstack.org/show/93958/
16:52 xarses joined #fuel
16:53 odyssey4me teknoprep - so vlan1000 is where your instances are running, right?
16:53 teknoprep they should be
16:53 teknoprep yes
16:54 odyssey4me can they ping each other?
16:54 teknoprep no
16:55 odyssey4me are they all on one compute host?
16:55 odyssey4me or spread?
16:55 teknoprep spread
16:55 odyssey4me ok, can you find two on the same host and test a ping between them?
16:55 teknoprep we can't log into them
16:55 teknoprep to test
16:55 odyssey4me you can see the host they're on in the 'admin' view
16:56 teknoprep i need to get a box that can install from ISO
16:56 odyssey4me through the novnc console?
16:56 teknoprep or has a default user/pass and doesn't require a key
16:56 odyssey4me are any of them from the testvm image?
16:56 teknoprep i can fire that up if you want me to
16:56 teknoprep does it have a default root pass?
16:57 odyssey4me yeah, fire some of those up - yes it does, I think... otherwise I'll point you at two ways to do it
16:58 odyssey4me here's a small test image, which I think is the one used for the testvm: http://download.cirros-cloud.net/0.3.2/cirros-0.3.2-x86_64-disk.img
16:58 teknoprep i rebooted the  entire cluster
16:58 odyssey4me otherwise here's the ubuntu cloud image for trusty: https://cloud-images.ubuntu.com/trusty/current/trusty-server-cloudimg-amd64-disk1.img
16:58 teknoprep give me a few
16:59 odyssey4me if an image needs a key, then you can boot a server with a special config to set the root password - let me paste that up for you
16:59 teknoprep thanks
17:01 xarses the cirros image is easy to get into and you can make a 64m flavor and boot tons of them
17:01 teknoprep i am doing that now
17:01 teknoprep whats the password toget into it ?
17:02 odyssey4me it shows on the console
17:02 teknoprep thanks
17:02 xarses cirros : cubswin:) or something
17:02 odyssey4me as I recall it's root with the password cubswin:)
17:02 odyssey4me what xarses said :p
17:03 teknoprep botting 20
17:03 teknoprep that way i know some will be on save server
17:03 odyssey4me otherwise for something like the ubuntu image you can use the 'Customization Script' panel and enter this: http://paste.openstack.org/show/93959/
17:03 odyssey4me it will enable root login via ssh and set the password to the one shown there
17:04 odyssey4me more details about that trick here: http://docs.openstack.org/grizzly/openstack-compute/admin/content/user-data.html
17:15 xdeller joined #fuel
17:15 twforeman joined #fuel
17:25 odyssey4me teknoprep - have you determined whether instances on the same host can communicate yet?
17:26 teknoprep no a customer called
17:26 teknoprep working on this and thenback to that
17:26 odyssey4me fyi - a common issue we've found when working with vlan's is that the vlan doesn't exist on the physical switch
17:26 odyssey4me if it doesn't, then inter-host comms doesn't work
17:27 odyssey4me same-host commsn works as it all uses the internal switch on the host
17:33 teknoprep ok back
17:33 teknoprep yes we have already added 1000 - 1030 on the switch
17:33 teknoprep lol wtf
17:33 teknoprep the test VM is doint iPXE shit
17:37 odyssey4me to be honest I haven't tested the 'testvm' image other than to launch it to confirm working bits
17:37 teknoprep what image does work ?
17:37 odyssey4me the cirros image is a pretty standard openstack test image
17:38 odyssey4me it's very basic, small and has enough for net testing
17:38 Samos123 what is the default password and username for fuel web, according to /etc/fuel/astute.yaml i thought it may be admin/admin
17:38 odyssey4me the 'testvm' image, I think, has some customisations
17:38 odyssey4me Samos123 -yes it is
17:39 Samos123 it keeps on saying Unable to login
17:39 Samos123 guess something else is going on
17:39 teknoprep are these IMG files considdereted RAW ?
17:40 odyssey4me teknoprep - nope, qcow2
17:40 e0ne joined #fuel
17:41 angdraug joined #fuel
17:41 teknoprep thanks
17:44 xarses Samos123: are you using one of the community master builds? or is this a release Mirantis version?
17:45 teknoprep yay holy shit
17:45 teknoprep brb
17:45 teknoprep that cirros image doesn't have local console
17:45 teknoprep wow
17:46 xarses teknoprep: it should
17:46 teknoprep ok
17:47 teknoprep what good is associating a floating IP when using Nova ?
17:47 teknoprep i mean when using Neutron ?
17:48 odyssey4me the floating ip association is all about applying a nat
17:48 odyssey4me it's part of the general idea of having 'private' networks, but providing access to users to apply 'public' addresses to instances
17:48 teknoprep ahh... is there no NAT when a VM is directly connected to the External_Net ?
17:48 xarses it allows you to assign a floating ip which can pierce the nat to the instance
17:49 odyssey4me teknoprep - well, you shouldn't be building instances on the external network
17:49 xarses teknoprep: correct, but also note that fuel doesn't set up DHCP on the external_net
17:49 xarses so you won't be able to ping those instances
17:49 teknoprep gotcha
17:49 odyssey4me teknoprep - in fact, you shouldn't be building instances as admin
17:49 teknoprep unless i manually assign the ip address
17:50 teknoprep odyssey4me, this is only testing so far
17:50 teknoprep odyssey4me, just getting things up and running first
17:51 teknoprep so what is the external network for ?
17:51 teknoprep if i shouldn't assign that network to VM's
17:53 odyssey4me the external network is specially for the floating ip's ad the virtual router public interfaces
17:54 odyssey4me all projects setup their own internal networks, typically - depending on the topology you go with
17:54 teknoprep i get that
17:54 teknoprep i have an external network that can be applied to my VM's
17:54 teknoprep why would i want to do that ?
17:54 teknoprep or if i shnouldn't do that... why would it be allowed by default
17:55 odyssey4me teknoprep - are you creating vm's as the 'admin' user?
17:55 teknoprep yes
17:55 odyssey4me it's only available to that user because it's the owner of the network
17:55 odyssey4me it won't be available to any other users
17:55 teknoprep gotcha
17:56 odyssey4me the admin user should only be used to setup the infrastructure (external networks, etc)
17:57 odyssey4me you may wish to have a 'shared' network which all users can attach to... maybe for some sort of service network... there is that option
17:57 Samos123 xarses: I am using the following build fuel-community-master-62-2014-08-11_05-27-34.iso
18:02 odyssey4me teknoprep - I need to get going, I'll be online tomorrow if you're still struggling
18:09 teknoprep how do i remove a VM that is stuck in deleteing
18:09 teknoprep nvm
18:10 BillTheKat joined #fuel
18:23 rmoe joined #fuel
18:29 rmoe joined #fuel
18:35 e0ne joined #fuel
18:43 Samos123 im off to sleep, i noticed a postgres authentication issue in nailgun logs but didnt look carefully yet
19:46 e0ne joined #fuel
19:49 tatyana joined #fuel
21:42 kupo24z Any plans to up the ceph version in later fuel releases?
21:44 kupo24z Need something from 0.67.10
21:45 kupo24z Nvm, looks like its 0.80 on 5.1
22:17 angdraug joined #fuel
22:17 lordd_ joined #fuel
23:31 angdraug joined #fuel
23:33 lordd_ joined #fuel
23:35 lordd_ joined #fuel

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