Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-01-27

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

All times shown according to UTC.

Time Nick Message
23:38 e0ne joined #fuel
00:41 IlyaE joined #fuel
00:41 jouston joined #fuel
01:03 IlyaE joined #fuel
02:10 xarses joined #fuel
02:14 teran joined #fuel
02:36 jouston joined #fuel
02:47 ilbot3 joined #fuel
02:47 Topic for #fuel is now Fuel for Openstack: http://fuel.mirantis.com/ | Paste here http://paste.openstack.org/ | IRC logs http://irclog.perlgeek.de/fuel/
03:21 teran joined #fuel
03:50 jouston joined #fuel
04:20 teran joined #fuel
04:24 frankbutt joined #fuel
04:24 frankbutt left #fuel
04:51 teran joined #fuel
05:05 MACscr joined #fuel
05:05 MACscr been a couple months since i checked in, but hows the ubuntu/ceph support like in the latest fuel release?
05:22 teran joined #fuel
05:53 teran joined #fuel
06:24 teran joined #fuel
06:35 IlyaE joined #fuel
06:45 mrasskazov joined #fuel
07:09 e0ne joined #fuel
07:21 mihgen joined #fuel
07:49 teran joined #fuel
08:00 jouston joined #fuel
08:43 book` joined #fuel
08:43 miguitas joined #fuel
08:57 Bomfunk joined #fuel
09:01 bookwar joined #fuel
09:05 vk joined #fuel
09:09 mrasskazov joined #fuel
09:15 e0ne joined #fuel
09:15 alex_didenko joined #fuel
09:34 korn__ joined #fuel
09:42 teran joined #fuel
09:55 anotchenko joined #fuel
09:56 tatyana joined #fuel
10:01 mihgen joined #fuel
10:10 anotchenko joined #fuel
10:11 miguitas joined #fuel
10:13 e0ne_ joined #fuel
10:29 vk joined #fuel
10:32 teran joined #fuel
10:49 teran joined #fuel
10:54 anotchenko joined #fuel
11:18 anotchenko joined #fuel
11:39 e0ne joined #fuel
12:10 jouston joined #fuel
12:16 e0ne_ joined #fuel
13:00 anotchenko joined #fuel
13:06 Dr_drache joined #fuel
13:06 Drachev2 joined #fuel
13:06 Dr_drache hmmm
13:20 vkozhukalov joined #fuel
13:21 Dr_drache is there a reletivly simple way to pull down a newer release?
13:36 miroslav_ Dr_drache: Current version is available in Github. You may build your own ISO. However, it require Ubuntu with certain package versions
13:38 mattymo or Debian
13:41 Dr_drache well, that sucks.
13:41 Dr_drache LOL
13:42 Dr_drache miroslav_, I understand it's on github, but 90% of projects on github don't compile/build without "special sauce" type situations.
13:42 Dr_drache just trying to fully evaluate the software, and I cannot.
13:43 miroslav_ Btw, if you have HP DL360 G5 hardware under the hand, you may assist vkozhukalov in this chat to verify the /disk/by-id and /disk/by-path differences between CentOS and Ubuntu. It should greatly speed-up the patch creation.
13:44 Dr_drache miroslav_, I do, that's one of my issues, really slows down testing, that and the CEPH issue. I'm more than willing give more info, if the info I gave friday wasn't enough.
13:49 Dr_drache sorry if i come off abrasive. I'm just frustrated with the software thus far. no negtives ment to anyone.
14:04 vkozhukalov Dr_drache, could you please add some commands into ubuntu preseed early and late sections on master node?  Those commands are necessary to find out if there is a difference in naming hard drives under centos and ubuntu kernels. Please change /var/lib/cobbler/kickstarts/ubuntu-amd64.preseed and restart cobbler and then try to install ubuntu cluster and give us diagnostic snapshot.
14:05 vkozhukalov Dr_drache, patch for changing preseed file here https://drive.google.com/file/d/0BybDDjx4oqkYdWkyX3llMUtnNVk/edit?usp=sharing
14:07 Dr_drache vkozhukalov, give me a few min, and I'll get that done
14:10 vkozhukalov Dr_drache, thanx very much. Waiting for result. Of course, at least one of the baremetal servers needs to be HP with smart array raid controller.
14:11 anotchenko joined #fuel
14:14 Dr_drache no problem
14:29 teran_ joined #fuel
14:40 Dr_drache sadly patch isn't included in fuel.
14:40 Dr_drache gotta manually do it.
14:41 MiroslavAnashkin Dr_drache: Please run `yum install patch' on master node
14:41 xdeller joined #fuel
14:42 MiroslavAnashkin it is included into local repository but not installed by default.
14:45 Dr_drache MiroslavAnashkin, love too, but It has to be segragated, because I have a live DHCP server on my network, and fuel won't start with one.
14:47 anotchenko joined #fuel
14:50 MiroslavAnashkin Dr_drache: Hmm, yes, Fuel master node is DHCP for admin network segment. Is it possible to deal with existing DHCP?
15:11 e0ne joined #fuel
15:31 e0ne joined #fuel
15:40 IlyaE joined #fuel
15:48 xarses joined #fuel
15:57 angdraug joined #fuel
16:00 mihgen joined #fuel
16:16 Dr_drache MiroslavAnashkin, running the deploy.
16:24 e0ne_ joined #fuel
16:25 teran joined #fuel
16:35 Dr_drache MiroslavAnashkin,  http://www.sendspace.com/file/ixclie
16:35 Dr_drache dianostic logging file.
16:40 MiroslavAnashkin Dr_drache: Thnx!
17:01 Dr_drache MiroslavAnashkin, any ideas on the ceph issue?
17:12 rmoe joined #fuel
17:13 teran joined #fuel
17:23 vt102 joined #fuel
17:24 teran_ joined #fuel
17:30 mihgen_ joined #fuel
17:40 vkozhukalov joined #fuel
17:57 vk joined #fuel
18:08 teran joined #fuel
18:25 TVR_ joined #fuel
18:26 TVR_ I have a really stupid question... I installed a small cluster.. it succeeded:
18:26 TVR_ Deployment of environment 'Lab_Cluster_01' is done. Access the OpenStack dashboard (Horizon) at http://172.16.0.2/ or via internal network at http://10.7.212.88/
18:26 TVR_ when I try to connect to 10.7.212.88 it fails....
18:27 TVR_ other openstack dashboards would be /dashboard.... so I also tried http://10.7.212.88/dashboard
18:27 TVR_ it also cannot find it...
18:27 TVR_ netstat -tulp shows it's listening... telnet 10.7.212.88 80 makes a socket
18:27 TVR_ am I missing something?
18:28 TVR_ 10.7.212.0/24 is the only routable from my laptop is why I chose that adress
18:32 Bomfunk joined #fuel
18:45 MiroslavAnashkin Dr_drache: What kind of CEPH issue? I've ran through chat history and could not find.
18:45 MiroslavAnashkin TVR_: Are you use bare-metal or virtual machine installation?
18:46 Dr_drache MiroslavAnashkin, https://bugs.launchpad.net/fuel/+bug/1253115 very simular to that.
18:46 Dr_drache i'll get more info next time I try to deploy
18:46 TVR_ bare metal... fuel provisioning node was from image file... can ssh into the horizon server 10.7.212.88 without issue
18:48 MiroslavAnashkin TVR_: Please check your routes then. What operating system do you use on your laptop?
18:49 TVR_ windows for laptop... using either IE or Chrome, or firefox... (no firewall or norton software blockages.... can route to the horizon box without issue
18:54 MiroslavAnashkin Dr_drache: Neutron is not stable in working with VLAN splinters. We hope to fix this issue in 4.1, which is currently sheduled to the end of February.
18:55 MiroslavAnashkin TVR_: Please check the 10.7.212.0 network is included to your current route table. You may run `route print` to check it.
18:56 TVR_ so, to be clear, if I go to http://10.7.212.88/ I should see the standard dashboard... correct? I don't need to add /dashboard and it should just work... yes?
18:56 MiroslavAnashkin TVR_: In most cases - yes
18:57 TVR_ ok.. so if it is a networking issue.. I can figure that out... I just needed to eliminate the  "you forgot the /whatever" type incident
18:57 TVR_ thanks
18:57 MiroslavAnashkin TVR_: At least dashboard links in Fuel web interface does not include /dashboard
18:58 TVR_ ok.. cool... because the creation of the fuel node, the deployment, etc went fairly smooth...
18:59 TVR_ thanks again.. I will figure out where I am dropping then
19:03 MiroslavAnashkin TVR_: So, please check the route to dashboard first.
19:05 MiroslavAnashkin TVR_: Then, if part of the route goes through Neutron - please check promiscous mode is enabled on every physical network adapter, which holds neutron on this route.
19:06 TVR_ ok.. yea, that makes sense..... I will check that shortly
19:06 TVR_ thanks...
19:16 Dr_drache MiroslavAnashkin, the same error shows up in nova with flatDHCP
19:17 Dr_drache seems the option for "Ceph RBD for ephemaral volumes (nova)" causes ceph to crash and burn
19:24 Dr_drache MiroslavAnashkin, OVS VLAN splinters hard trunk workaround setting seems to have it work
19:25 MiroslavAnashkin Dr_drache: Yes, and it depends on NIC drivers.
19:26 Dr_drache blah, guess it was a bad move to buy all these new servers for a fuel deployment :P
19:28 Dr_drache take that back "014-01-27 19:29:14ERR
19:28 Dr_drache (/Stage[main]/Ceph::Conf/Exec[ceph-deploy config pull]/returns) change from notrun to 0 failed: ceph-deploy --overwrite-conf config pull node-15 returned 1 instead of one of [0] at /etc/puppet/modules/ceph/manifests/conf.pp:44"
19:28 Dr_drache oh well.
19:28 TVR_ I have seen this before
19:28 TVR_ log into one of the ceph nodes
19:29 TVR_ ping the other ceph node
19:29 TVR_ I bet you can't
19:29 TVR_ your networking VLAN tags are not good or the switch is not good for the management network
19:30 TVR_ the ceph setup has the nodes talk over that network
19:30 TVR_ :)
19:30 Dr_drache perhaps I need a better switch.
19:30 TVR_ I got arount that...
19:30 TVR_ I set the management network to NOT be tagged... and 'moved' it on another nic
19:31 TVR_ that solved my issue
19:31 Dr_drache I have a silent hate for vlan.
19:32 Dr_drache but, sadly it looks like I have to delete this cluster and "try again"
19:32 TVR_ so.. after you set up your network... add the nodes and configure the interfaces... drag the managemnet to a network by itself or with something tagged.. (it won't let you have two untagged on same phys nic)
19:32 TVR_ saddly.. yes..
19:32 Dr_drache been doing this every day for 2 weeks.
19:32 Dr_drache :P
19:33 TVR_ I am swimming in NICs and too lazy to do vlan switch configs, so I put everything on it's own nic
19:34 TVR_ I rolled my own writing all the puppet code having it connect to a ceph backend for the last 2 months, and neutron has killed me every time... I abandoned the RDO project and am now a fuel user.. this is a nice product I must say
19:34 Dr_drache I want to agree.
19:34 Dr_drache but havn't yet had a workable cluster.
19:35 TVR_ untag management.... then put it on a, or it's own nic.. you will be pleasantly surprised
19:35 Dr_drache nuetron?
19:36 TVR_ unfortunately... when you go to production, you will have to figure out what is wrong with your tags then.. but thats much later
19:36 Dr_drache sadly, not that MUCH later.
19:36 TVR_ neutron is now the basis for most of openstack SDN
19:36 Dr_drache we don't "require" VLANs
19:37 TVR_ these darn software people keep writing code... you know?
19:37 TVR_ heh
19:39 MiroslavAnashkin Dr_drache: Neutron, even in GRE mode may require VLANs. And starting from Havana - Neutron require promiscois mode everywhere.
19:40 angdraug joined #fuel
19:40 MiroslavAnashkin But for most cases you may simply click Verify network button on Networks tab in Fuel UI. But it works only before deployment.
19:40 TVR_ I am rebuilding mine (testing HA) so once complete, I will look to see if my nics are in promiscuous mode....
19:41 Dr_drache MiroslavAnashkin, that's fine.
19:41 Dr_drache I'm not against implmenting them.
19:42 Dr_drache now, if i can get ubuntu working :P
19:44 Dr_drache (we've been a ubuntu shop for going on 8 years)
19:44 Dr_drache i think
19:44 Dr_drache at least 5 i know for sure.
20:14 Dr_drache TVR_, looks like I need more than 3 NICS
20:15 MiroslavAnashkin 3 is minimum for certain configurations.
20:16 Dr_drache MiroslavAnashkin, If I choose not to vlan, i need like 5.
20:17 Dr_drache but, with ceph, it doesn't work with VLAN.
20:17 Dr_drache sooo, the only deployment choice I have is CentOS with no vlan. LOL
20:17 TVR_ I tagged the storage network... for now.. just to see it build.... but yes.. either you need 4 nics or VLANS that work
20:19 MiroslavAnashkin CEPH storage network and management network may be VLAN tagged, but your hardware switch should support and allow these VLANs
20:20 MiroslavAnashkin And your switch should not cut off the promiscous traffic
20:22 Dr_drache hmm, I got a crappy switch out of our storage room. maybe that's my entire issue
20:23 MiroslavAnashkin TVR_: Please try http://172.16.0.2/horizon/ or http://10.7.212.88/horizon/
20:24 TVR_ once back up.. will do...
20:24 MiroslavAnashkin looks like there is redirection bug under Ubuntu - it points to /dashboard instead of /horizon
20:24 Dr_drache TVR_, MiroslavAnashkin any advice for a trendNET switch?
20:25 TVR_ sorry.. none here..
20:27 Bomfunk joined #fuel
20:59 TVR_ apparently... if you create enough clusters, and then delete them... you will run out of IP space... It doesn't seem to recycle the IP's from deleted clusters... get the error :::   Not enough free IP addresses in pool
21:00 TVR_ rebuild off of USB stick time...
21:00 Dr_drache Yea, there is no easy way to start over for real
21:01 Dr_drache or edit the cluster once deployed.
21:01 Dr_drache or change to HA from non-HA.
21:01 Dr_drache lol
21:07 e0ne joined #fuel
21:45 Dr_drache seems proper vlan isn't the problem
21:58 xarses TVR_ you can reset the db if you need
21:58 xarses clean all the fluff out
21:58 xarses also, please create a pug for that
22:03 TVR_ joined #fuel
22:14 jouston joined #fuel

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