Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2013-12-19

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

All times shown according to UTC.

Time Nick Message
23:00 e0ne joined #fuel
23:03 albionandrew joined #fuel
23:24 e0ne joined #fuel
00:00 e0ne joined #fuel
00:12 IlyaE joined #fuel
00:25 e0ne joined #fuel
01:25 e0ne joined #fuel
02:12 rongze joined #fuel
02:25 e0ne joined #fuel
03:07 vkozhukalov joined #fuel
03:15 h6w joined #fuel
03:16 h6w Does anyone else get an "invalid hostname" warning (due to the underlines) when installing new nodes?
03:25 e0ne joined #fuel
03:35 ArminderS joined #fuel
03:38 h6w MiroslavAnashkin: Timezone tag!  I took your suggestions and double-checked the DHCP.  There are no other DHCP servers running.  So I created a version in a VirtualBox "host-only" environment, disabled the DHCP server, reinstalled the master, checked in "Fuel Setup" that no other servers were running, and the error still occurs. :-(
03:44 test__ joined #fuel
03:53 ArminderS- joined #fuel
04:09 e0ne joined #fuel
04:25 e0ne joined #fuel
05:18 rongze joined #fuel
05:21 Shmeeny joined #fuel
05:25 e0ne joined #fuel
05:48 mihgen joined #fuel
06:20 rongze joined #fuel
06:25 e0ne joined #fuel
06:27 rongze joined #fuel
06:33 rvyalov joined #fuel
06:37 rongze joined #fuel
07:00 Arminder whats the current status for fuel 4.0?
07:00 mrasskazov joined #fuel
07:01 Arminder mihgen: ^
07:03 mihgen Arminder: it's Ok except a few criticals we are still struggling with… the most severe one is with new Swift - affects Ubuntu HA deployment
07:05 teran joined #fuel
07:10 ArminderS- anything severe for centos+ha+neutron w/gre+ceph for volume/glance/object/ephemeral
07:11 xarses joined #fuel
07:15 ArminderS- xarses: would you like to comment on the current status for fuel 4.0?
07:15 ArminderS- anything severe for centos+ha+neutron w/gre+ceph for volume/glance/object/ephemeral
07:15 anotchenko joined #fuel
07:25 e0ne joined #fuel
07:26 mihgen ArminderS-: this one is under investigation: https://bugs.launchpad.net/fuel/+bug/1260911
07:27 mihgen this one is fixed: https://bugs.launchpad.net/fuel/+bug/1256289
07:28 ArminderS- thanks mihgen
07:30 ArminderS- how do i see what build i'm on?
07:32 e0ne joined #fuel
07:34 anotchenko joined #fuel
07:44 SteAle joined #fuel
07:49 rongze joined #fuel
07:55 xarses .   /api/version
07:59 jkirnosova joined #fuel
08:01 jkirnosova left #fuel
08:02 ArminderS- i used the iso building doc to create the iso, where would I find the api/version file?
08:02 ArminderS- under clone of fuel-main repo?
08:02 mihgen in rest api
08:02 mihgen :8000/api/version
08:06 ArminderS- thanks
08:07 ArminderS- so when some bug says its from iso 139 and its fixed, how does one conclude whether their build is newer or old?
08:07 ArminderS- i matched that with 1, all except the fuellib_sha matches
08:07 ArminderS- with the version I'm running
08:10 xarses there are about two to four builds a day this late in the release, so 139 is over a week old
08:11 xarses 146 was from last thursday i think
08:17 alex_didenko joined #fuel
08:17 mattymo xarses! you're on late
08:20 alex_didenko joined #fuel
08:21 xarses just going to bed
08:22 mihgen ArminderS-: since yesterday you can also see build number in /api/version, and timestamp when it was built
08:22 mihgen if it is being build on our Jenkins
08:29 vkozhukalov joined #fuel
08:32 ArminderS- thanks mihgen, xarses
08:46 anotchenko joined #fuel
08:52 anotchenko joined #fuel
09:06 SteAle joined #fuel
09:13 ArminderS joined #fuel
09:18 evgeniyl joined #fuel
09:26 e0ne joined #fuel
09:30 miguitas joined #fuel
09:47 ruhe joined #fuel
09:53 anotchenko joined #fuel
09:58 ruhe joined #fuel
10:03 SergeyLukjanov joined #fuel
10:23 rongze joined #fuel
10:38 anotchenko joined #fuel
10:42 teran joined #fuel
10:51 SteAle joined #fuel
10:51 xdeller joined #fuel
10:59 e0ne joined #fuel
11:04 ruhe joined #fuel
11:08 e0ne joined #fuel
11:12 SteAle joined #fuel
11:32 ruhe joined #fuel
11:41 vk joined #fuel
11:47 anotchenko joined #fuel
11:50 mihgen joined #fuel
11:54 anotchenko joined #fuel
12:04 rongze joined #fuel
12:13 e0ne_ joined #fuel
12:20 ruhe joined #fuel
12:24 rongze_ joined #fuel
12:40 anotchenko joined #fuel
12:45 Vidalinux joined #fuel
13:06 e0ne joined #fuel
13:12 anotchenko joined #fuel
13:16 e0ne joined #fuel
13:17 e0ne_ joined #fuel
13:18 e0ne joined #fuel
13:59 e0ne_ joined #fuel
14:11 Nikolay joined #fuel
14:21 SteAle joined #fuel
14:32 mattymo joined #fuel
14:52 anotchenko joined #fuel
15:21 IlyaE joined #fuel
15:22 anotchenko joined #fuel
15:36 MiroslavAnashkin h6w: Can you share diagnostic shapshot? Please go to Fuel web UI, Support, click Generate diagnostic snapshot. Your issue may be related to initial Fuel Admin network settings mismatch with your network configuration.
15:38 MiroslavAnashkin h6w: You may use Google drive
15:42 rongze joined #fuel
15:44 Shmeeny joined #fuel
15:59 ArminderS joined #fuel
16:14 ArminderS joined #fuel
16:16 rongze joined #fuel
16:21 xarses joined #fuel
16:22 rongze joined #fuel
16:24 teran joined #fuel
16:37 teran joined #fuel
16:38 SergeyLukjanov joined #fuel
16:44 ArminderS- joined #fuel
16:47 ruhe joined #fuel
16:55 mattymo joined #fuel
16:55 vkramskikh joined #fuel
16:57 Vidalinux joined #fuel
17:02 angdraug joined #fuel
17:02 ArminderS joined #fuel
17:04 mattymo joined #fuel
17:04 vkramskikh joined #fuel
17:07 SergeyLukjanov joined #fuel
17:15 mattymo joined #fuel
17:15 vkramskikh joined #fuel
17:40 rongze joined #fuel
17:46 vkozhukalov joined #fuel
18:00 e0ne joined #fuel
18:04 e0ne joined #fuel
18:06 e0ne joined #fuel
18:18 e0ne joined #fuel
18:19 ruhe joined #fuel
18:23 xarses joined #fuel
18:27 mihgen joined #fuel
18:31 IlyaE joined #fuel
18:35 e0ne joined #fuel
18:40 rongze joined #fuel
19:00 mihgen joined #fuel
19:03 e0ne joined #fuel
19:53 e0ne joined #fuel
19:59 e0ne joined #fuel
20:15 albionandrew joined #fuel
20:27 mihgen joined #fuel
20:28 IlyaE joined #fuel
20:33 rmoe joined #fuel
21:03 albionandrew hi xarses. We deployed last night passed all health checks etc. We are unable to get a dhcp address from openstack. With tcpdump the controller can see the requests but does not respond. What should I restart/start ? /etc/init.d/dnsmasq?
21:08 ArminderS- joined #fuel
21:19 xarses neutron VLAN still?
21:22 xarses albionandrew: ^
21:23 albionandrew xarses yes.
21:24 albionandrew assigns ip shown in horizon in neutron range. When I log onto console no ip address. dhclient eth0 eventually times out.
21:24 albionandrew controller shows request.
21:25 albionandrew nova console list shows lots of [  152.427364] cloud-init[422]: 2013-12-19 21:19:34,107 - url_helper.py[WARNING]: Calling 'http://169.254.169.254/2009-04-04/meta-data/instance-id' failed [87/120s]: request error [HTTPConnectionPool(host='169.254.169.254', port=80): Max retries exceeded with url: /2009-04-04/meta-data/instance-id (Caused by <class 'socket.error'>: [Errno 101] Network is unreachable)]
21:25 xarses ok, so the dhcp should be provided by dnsmasq on controller running quantum_dhcp_agent
21:25 albionandrew yes I think so.
21:25 xarses but that its from the metadata service
21:26 albionandrew just an error I found might be useful
21:26 albionandrew so restart ... ?
21:27 albionandrew quantum_dhcp_agent and dnsmaq?
21:27 albionandrew or is there something else I should try?
21:27 xarses one of the controllers should have qantum metadata agent also
21:30 xarses you will want to verify the net namespaces for qrouter and qdhcp and verify that they are up
21:32 mihgen albionandrew: when you create an instance, do you add net04 network?
21:33 albionandrew net-4-ext
21:33 mihgen this net doesn't provide DHCP
21:33 mihgen please try with network which has no "-ext"
21:33 mihgen your instance should get IP then
21:34 dhblaz joined #fuel
21:34 xarses mihgen: should we remove the public flag from -ext so it isn't seen so easily?
21:37 dhblaz If the cluster isn't going to hand out dhcp addresses on ext why is there a check for a dhcp server there in the network verification test?
21:37 mihgen xarses: frankly speaking, I don't know that much on the topic at the moment. But these things confuse me too. I'm wondering why we keep hacky things like "net04", what 04 stands for, etc.
21:37 dhblaz 04 is the cluster id
21:38 mihgen dhblaz: dhcp check verifies that there will be no other DHCP in your private network
21:38 dhblaz it at least recently checked on the public/float network too
21:38 mihgen so it is ok in default deployments with Fuel, it is really required if you don't want to mess up DHCP leases to your VMs
21:40 mihgen ok, yes. It should warn actually, as Fuel uses public IP ranges - so you have to be aware, that your infrastructure DHCP may lease IP for some other node which is already statically set on one of Fuel nodes
21:40 mihgen so you may get two same IPs in same L2
21:41 dhblaz This is poorly communicated
21:42 dhblaz i.e. I haven't ever seen it in the documentation and the network verification checker (when it detects a dhcp_server which it doesn't usually) it shows up as an error.
21:43 dhblaz I don't think that this is a fuel thing, but if you put a nic in the ext network shows an IP address in the horizon interface, but whatever mechanism is designed to assign that IP address fails.
21:49 albionandrew mihgen xarses thanks
21:55 Vidalinux joined #fuel
22:00 dhblaz Yes, thank you both
22:01 mihgen woops a second late for dhblaz
22:01 Bomfunk joined #fuel
22:01 mihgen xarses: did you get the points dhblaz mentioned?
22:02 mihgen I can't clearly understand what we would need to add to docs about DHCP checker
22:03 mihgen and about ext network in Horizon - don't get suggestions
22:12 xarses I don't know why we check for dhcp on public, like when you say, we don't set it up anyway. I can say that the deployment topology intends that net04-ext should be non-public and you wouldn't normally assign a instance to it as instead you would assign a floating ip to it after it was built on net04
22:13 LGI joined #fuel
22:15 LGI Hi all, anyone able to point me to an option to use a 32bit arch for the master or controller node in an evaluation environment that I'm building?
22:17 xarses LGI: Fuel doesn't support deploying 32-bit cloud controllers, however once deployed you could create 32-bit guest instances.
22:22 LGI Thank you @xarses; I was trying a quick and dirty (32bit) alternative in place of a controller deployed on an old IBM x3950 that bootstraps fine but the hdd shows as 0?... so obviously when I attempt to allocate it as a controller it throws an error.  Anyone have any time to walk me through one or two things to figure the hdd identification isssue with me?
22:22 xarses LGI: it might be possible to deploy a 32-bit fuel controller, but you need 64-bit processors for computes to support virtual extensions that every hypervisor requires for reasonable performance of guest instances
22:23 LGI @xarses: yeah, my current master and compute nodes are both 64-bit, but the third one intended as the controller is having the ^ hdd issue.
22:25 albionandrew if LGI 's issue is the same one I had I put the fixed agent file here - http://pastebin.com/j27WTZqQ ... I don't know much but I had a similar problem and that fixed it.
22:25 albionandrew moved it to /opt/nailgun/bin/ then ran it it fixed the issue for me.
22:26 LGI ^ the error that I see on a 32-bit controller-node host during pxe bootstrapping is that the arch isn't support by the images available, but I can't for the life of me find the mirantis 32-bit images or where to put them if i did
22:26 xarses ahh OK, we can attempt to help trouble shoot the hdd issue. You can look at the code in this https://bugs.launchpad.net/fuel/+bug/1259276 bug as we recently fixed it for cciss controllers.
22:26 LGI thanks @xarses & albionandrew, I'll take a look at both of those links
22:27 xarses LGI same problem
22:37 h6w Morning all!
22:37 h6w MiroslavAnashkin: I've PMd you a link to my snapshot.
22:42 teran joined #fuel
22:43 LGI Okay @xarses and albionandrew, I believe that I need to update ks.template, but that file doesn't exist on my master...  would you suggest that I create the iso/ks.template file in /opt/nailgun/bin?
22:44 LGI or is it on the boostrapped node where the change goes?
22:44 LGI ^thinking out loud, going to look
22:44 albionandrew yes to the bootstrap.
22:45 albionandrew got op emerge got to go.
22:46 LGI perfect, thanks, found "agent" on bootstrap and will add that update.   ...
22:51 xarses LGI: so iso/ks.template is only for the controller. /opt/nailgun/bin/agent is for discovery (so you can map the disk in the UI when you add it to a cluster) and then there is a cobbler template generator pmanger.py in fuel-library deployment/puppet/cobbler that is used for deploying nodes
22:57 LGI @xarsus, so I've updated my agent; how do I get my master to re-provision the bootstrapped host, or do I re-bootstrap the host?
22:59 LGI ^ i'll just try allocating it and seeing if it fails again :0

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