Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-01-03

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

All times shown according to UTC.

Time Nick Message
23:13 rongze joined #fuel
00:11 teran joined #fuel
00:14 rongze joined #fuel
00:52 vsonic joined #fuel
00:53 vsonic Hi all, I have recently installed openstack with fuel and I am really liking it, but when I try to create an environment using neutron I get an error saying "cannot find suitable cidr" does anyone know why that is?
01:01 xarses joined #fuel
01:01 xarses_ joined #fuel
01:05 mutex vsonic: cidr is a network subnet
01:05 mutex so 192.168.1.0/24 is a cidr
01:07 vsonic mutex: is there some way of setting the cidr in fuel?
01:08 mutex during the install yes
01:08 mutex otherwise you can set it in the network configuration section
01:11 vsonic mutex: the network settings for the environment is stuck with a loading bar, is there another way to set it?
01:15 rongze joined #fuel
01:15 mutex hm, seems like you've got problems
01:15 mutex I mean the network configuration cluster wide
01:15 mutex not for the node
01:16 vsonic It works with Nova network, its just when I set it to use Neutron that everything stops working
01:17 mutex are you using VLANS or GRE ?
01:18 vsonic Im using GRE but it happens with both VLANS and GRE
01:25 h6w Oh, hallo!  Thought people were asleep.
01:25 h6w Question from the OS channel.  Does fuel support bare VLANs?
01:27 h6w I notice that under Nova-Network, FlatDHCP Manager offers the "Use VLAN Tagging", but "VLAN Manager" doesn't.
01:34 rmoe joined #fuel
01:38 rongze joined #fuel
01:47 vsonic joined #fuel
01:52 vsonic joined #fuel
02:00 xarses joined #fuel
02:01 vsonic joined #fuel
02:19 rongze joined #fuel
02:20 rongze joined #fuel
02:32 rongze joined #fuel
02:38 vsonic joined #fuel
03:27 ArminderS joined #fuel
03:28 jouston Hello guys. Anyone know where I can find Mirantis hardened openstack source deb? I'd like to modify it a bit when doing deployment.
03:31 jouston joined #fuel
03:32 ArminderS- xarses:nope, couldn't get it working
03:33 ArminderS- the bigger problem is that even if i don't use nic bonding, still it won't complete the build, till i remove the extra dual NIC
03:34 ArminderS- with extra NIC inside, it fails to connect to management & storage networks
03:51 h6w ArminderS-: Are you talking about NIC binding?  I'm having issues where my NICs on new nodes are unpredictably ordered.
04:08 ArminderS h6w: aye, nic bonding
04:08 h6w I understood bonding was different to binding.  Bonding is when you join two NICs together to form one virtual NIC which can theoretically create faster bursts.
04:08 ArminderS i had success building cluster fine earlier on the same nodes, but adding another dual NIC to have 3 interfaces bonded for storage network, the deployment fails
04:09 ArminderS tried as explained at http://docs.mirantis.com/fuel/fuel-4.0/reference-architecture.html#advanced-network-configuration-using-open-vswitch
04:09 ArminderS but no luck
04:10 ArminderS tried different ways to provided interfaces parameter, but it errors out
04:10 ArminderS err:  Transformation bond 'bond0' have wrong 'interfaces' parameter. at /etc/puppet/manifests/site.pp:130 on node node-20
05:25 ArminderS joined #fuel
05:44 xarses ArminderS: can you paste the network section of your astute.yaml?
06:19 ArminderS xarses: you mean text under ADMIN_NETWORK?
06:21 ArminderS http://paste.openstack.org/show/59801/
06:24 xarses sorry, the network_scheme section
06:26 xarses ArminderS: like in the linked doc you noted above
06:44 ArminderS you mean for the node
06:46 ArminderS here it is for 1 of the ceph nodes-> http://paste.openstack.org/show/59809/
06:47 ArminderS xarses: above is the one I modified as per the document and uploaded back to fuel
06:49 ArminderS the default one for same is here -> http://paste.openstack.org/show/59810/
06:53 ArminderS is it possible to reset fuel master without rebuilding from ISO?
06:53 ArminderS i know the pxe config remains the same, but can everything else be reset?
06:54 ArminderS like the node number start afresh from node-1 and so does environment from 1
06:54 xarses ArminderS: you could drop the db and re-set it up
06:54 ArminderS little more details could help
06:56 xarses hmm your changes look like they should work
06:56 xarses I'll have to test it in the morning
06:56 xarses did you re-upload the whole yaml file, or just these changes?
06:57 xarses and did you change the yaml for each node role? controller-22, cephosd-22 etc..
06:57 ArminderS tried both, first whole of deployment
06:57 ArminderS and then just single node
06:58 xarses ok, to note, you have to re-put the whole file, not just this section
06:58 ArminderS actually had just 2 dual-nics, so added 1 each on 2 ceph nodes only
06:58 ArminderS the rest of controllers and compute got 4 nics only
06:58 ArminderS just 2 ceph nodes got 6
06:58 ArminderS i had put in the whole file
06:59 ArminderS editing the network_scheme section only
06:59 xarses ok, in /opt/lib/python/site-packages/nailgun there should be a manage.py somewhere
06:59 xarses probably in bin
07:00 ArminderS its here -> /opt/nailgun/bin/manage.py
07:00 xarses you can use that to drop the db, and then syncdb, and there's one other command that is in there that you need
07:02 ArminderS so first dropdb, then syncdb and then
07:10 xarses loaddefault
07:10 xarses sorry, had to look it up
07:12 rongze joined #fuel
07:12 ArminderS ok so -> manage.py dropdb && manage.py syncdb && manage.py loaddefault
07:12 ArminderS right
07:20 xarses yep
07:21 ArminderS thanks andrew
07:22 xarses yw
07:22 xarses g'night
07:22 ArminderS good night
07:30 teran joined #fuel
08:31 teran joined #fuel
10:04 rongze joined #fuel
10:14 jouston joined #fuel
10:21 archon__ joined #fuel
10:25 archon__ hello, i have problems with cinder in fuel 3.2.1 : http://paste.openstack.org/show/59858/
10:26 archon__ been doing that on all 3 controllers, same result, always dead but pid file exists
10:42 teran joined #fuel
10:42 besem9krispy joined #fuel
10:45 besem9krispy I'm getting blocked by "Some networks are assigned to the same physical interface as admin (PXE) network." -- is it possible to tell fuel that I do not care about this right at the moment?
10:45 besem9krispy And hello!
12:40 tatyana joined #fuel
12:58 ArminderS joined #fuel
13:03 teran joined #fuel
13:41 rongze joined #fuel
14:01 besem9krispy joined #fuel
14:45 rongze joined #fuel
14:56 rongze joined #fuel
15:12 ruhe joined #fuel
16:11 xarses besem9krispy: in neutron mode, no its not currently possible
16:18 besem9krispy I get that in Neutron AND flat dhcp modes :(
16:25 xarses besem9krispy: in flat dhcp it should share as long as the other roles are tagged interfaces
16:26 xarses but it wont if they don't have a vlan tagged
16:26 besem9krispy Do you think it matters that I used a network cable instead of a switch?
16:33 xarses oh, I think i'm thinking of a seperate issue.
16:33 xarses you can't share the admin network range with any network range in the cluster you are creating
16:34 xarses i was thinking of the interface role mapping, but i realize you are probably on the network screen
16:46 besem9krispy So if I want to deploy a rack of openstack servers, I need two 40 port switches, one of which is exclusively for the use of the admin network?
16:46 besem9krispy 80 cables, 40 of which carry very little actual data ?
17:36 ArminderS joined #fuel
17:42 ruhe joined #fuel
17:47 angdraug joined #fuel
17:51 rmoe joined #fuel
18:14 xarses joined #fuel
19:04 jof_ left #fuel
19:16 rongze joined #fuel
19:16 xarses joined #fuel
19:29 xarses joined #fuel
19:36 Chocobo joined #fuel
19:36 Chocobo Hi all.  To use the virtualbox scripts does one of your network connectors need a name starting with "eth"?  When I launch the script everything seems to work expecpt networking.
19:48 xarses Chocobo: what os are you on?
19:50 Chocobo xarses: Arch Linux
19:51 Chocobo My interfaces are named eno1 (wired) and wlp2s0 (wireless)
19:51 xarses Chocobo: ya, the scripts should create networks for you
19:52 xarses it dosn't look for eth0 on the host
19:52 xarses it should have created vboxnet0 vboxnet1 vboxnet2
19:53 Chocobo xarses: It does create those but they are listed as "DOWN" and do not seem to have an IP associated with them.   The fuel vm seems to get an IP with DHCP but has not path out (hangs on ntp).
19:54 Chocobo When it gets to the step informing me that the fuel web ui is up at http://10.20.0.2:8000 ...  well, I can't access it
19:55 xarses fuel defaults to 10.20.0.2, its in the kernel args when the iso is loading which using the scripts should be correct
19:55 xarses hmm, thats odd
19:57 Chocobo xarses: Aye.  I am not running any firewall so that can't be it.  Maybe I should try again and check my routing table.   I can't even ping 10.20.0.2 once it is up though.
19:59 xarses you should be able to get to 10.20.0.1, which should be one of the vboxnet interfaces
19:59 xarses if not then the problem is with how the network was created in virtualbox
20:03 Chocobo I will try again.  :/   I loaded all of the virtualbox modules (vboxdrv, vboxpci, vboxnetflt, vboxnetadp)
20:04 xarses should just need vboxmanage
20:05 Chocobo Here is what I see when installing:  http://paste.openstack.org/show/59974/
20:06 Chocobo is vboxmanage a kernel module?  I don't seem to have that one.
20:07 Chocobo Yep, it is still hanging on ntpdate.   hrm.
20:12 xarses Chocobo: looks like its related to vboxnet0 not having an ip
20:12 xarses let me look at something
20:13 Chocobo Err.. solved it.  Arch doesn't come with ifconfig and other deprecated net tools.  I installed those and it seems to be working much better.
20:13 xarses hmm thats odd, everything is done in VBoxManage not sure how ifconfig was needed
20:15 Chocobo xarses: not sure... it must use one of these tools though:  http://paste.openstack.org/show/59975/
20:16 rongze joined #fuel
20:17 xarses ya, we are guessing that VBoxManage is using ifconfig behind the scene
20:18 Chocobo At least it was a pretty easy fix (just had to install net-tools).  Thanks for the help.
20:19 xarses Chocobo: https://bugs.launchpad.net/fuel/+bug/1265902
20:20 Chocobo Hey, cool.  Thanks.
20:27 Chocobo Wow, this is an amazingly cool tool.
20:27 Chocobo I set up a HA cluster by hand and it was a very painful process.
20:41 tatyana joined #fuel
20:43 Chocobo Hrmmm which AMQP does FUEL use?  I saw qpid and rabbit get installed.
20:44 xarses qpid is only used for RHOS, otherwise its rabbit
20:45 Chocobo xarses: thanks.   In my next cluster I was going to fool around with zeromq.  Any reason that Fuel doesn't use that one?
20:45 Chocobo I am interested to see how HA rabbit is handled (cluster or drbd)
20:46 xarses iirc we wrote some patches to make rabbit multi HA better
20:46 xarses so that's probably why we use it
20:48 Chocobo xarses: makes sense.  Thanks.
21:03 Chocobo So is there a clean way to shutdown the virtualbox cluster?
21:16 tatyana joined #fuel
21:19 rongze joined #fuel
21:19 xarses joined #fuel
22:09 xarses_ joined #fuel
22:10 SergeyLukjanov joined #fuel
22:20 rongze joined #fuel

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