Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2016-07-11

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

All times shown according to UTC.

Time Nick Message
00:59 Julien-zte joined #fuel
01:31 Julien-zte joined #fuel
01:51 code-R joined #fuel
03:53 code-R joined #fuel
04:06 code-R_ joined #fuel
04:20 code-R joined #fuel
04:21 Julien-zte joined #fuel
05:59 athinkingmeat joined #fuel
06:12 javeriak joined #fuel
06:25 dkaiharodsev joined #fuel
06:48 vkulanov joined #fuel
07:04 DavidRama joined #fuel
07:05 dkaiharodsev joined #fuel
07:14 yassine joined #fuel
07:25 dkaiharodsev joined #fuel
07:45 dkaiharodsev joined #fuel
07:46 dkaiharodsev joined #fuel
07:47 dkaiharodsev joined #fuel
07:48 dkaiharodsev joined #fuel
08:08 bgaifullin joined #fuel
08:31 noshankus joined #fuel
08:39 kaliya joined #fuel
08:59 Ravikiran_K joined #fuel
09:04 toMeloos joined #fuel
09:08 javeriak joined #fuel
09:13 noshankus joined #fuel
09:19 ekosareva joined #fuel
09:21 code-R joined #fuel
09:24 toMeloos We have provisioned compute nodes rebooting at random times and then PXE boot hanging on "Boot sector signature not found" with Fuel 8. Anybody know what's going on?
09:25 vkulanov joined #fuel
09:29 ikalnitsky toMeloos: did you managed to boot at least once from HDD?
09:30 ikalnitsky kozhukalov: agordeev: did you hear something about toMeloos' problem?
09:32 toMeloos ikalnitsky: yes it was provisioned and came up in the Fuel dashboard as an operational node
09:32 toMeloos we ran some VM's on them
09:33 toMeloos This test setup included 3 controllers and 2 computes. They were all deployed last Thursday and the 2 compute nodes were rebooted somehow (not by us) sometime this weekend
09:38 ikalnitsky Any chance they were deleted from Fuel env (i mean using fuel api) ?
09:39 fatdragon joined #fuel
09:41 tatyana joined #fuel
09:48 bgaifullin joined #fuel
10:00 yassine joined #fuel
10:19 ekosareva joined #fuel
10:22 toMeloos ikalnitsky: nope. We've reset this environment and redeployed it once or twice though, if that makes a difference
10:23 toMeloos ikalnitsky: so turns out we were also unable to SSH into one of the compute nodes and many commands on the other one failed on "annot execute binary file: Exec format error"
10:33 javeriak joined #fuel
10:36 vkulanov joined #fuel
10:56 agordeev toMeloos: hi, "boot sector signature not found" could be the result of several issues. could you provide more information?
10:57 toMeloos hi
10:59 toMeloos agordeev: looks like the real issue is that fuel has broken Ubuntu. We get "boot medium not found" when forcing to boot from disk and the PXE message "boot sector signature not found" seems like PXE-speak for the same problem.
11:02 toMeloos Nodes reboot after Fuel has installed Ubuntu, right?
11:02 agordeev toMeloos: yes, they'll be rebooted
11:03 agordeev toMeloos: perhaps, it's better to file a bug. it will definitely help fuel-team to go further with more precise root cause analysis. can you share the diagnostic snapshot with us?
11:04 toMeloos Initially the node reached "Ready" state. So the image is deployed correctly and it reboots. Then during the puppet run something breaks the OS
11:04 toMeloos Does it run apt upgrade / apt dist-upgrade? or grub update? For one it looks like the MBR is missing
11:07 agordeev i could only second your doubts about missing MBR on those nodes.
11:08 AlexAvadanii joined #fuel
11:08 toMeloos the 3 controller nodes are doing just fine...really weird
11:09 toMeloos The MBR is there initially.... since the first boot is succesful. Something is broken causing A) the machines to reboot themselves and B) not boot anymore
11:11 reddydodda joined #fuel
11:12 agordeev toMeloos: the log files could shed more light on that
11:13 reddydodda hi. i am using fuel8.0 and i had 3 controller nodes. 2 nodes went down and i was able to manage with one controller node. but after sometime . even it was not responding. i rebooted that one controller node but nothing is coming up.
11:13 reddydodda here is status of pcs resource
11:13 reddydodda http://paste.openstack.org/show/529628/
11:13 reddydodda can some one help me on this issue. ??
11:14 reddydodda can i add 2 more nodes and deploy changes ? will it brings up all services with out any data loss ??
11:22 code-R joined #fuel
11:27 Vijayendra joined #fuel
11:30 vkulanov joined #fuel
11:37 reddydodda Hi .. Please can someone help me in this error http://paste.openstack.org/show/529628/
11:48 Aurelgadjo reddydodda: https://docs.mirantis.com/openstack/fuel/fuel-8.0/operations.html#troubleshooting
12:10 reddydodda Hi here is pacemaker log file. https://drive.google.com/file/d/0Bwl2aRK5fBrWckZUT0JpcFhsMWc/view?usp=sharing
12:11 reddydodda Aurelgadjo: thanks for your reply. in troubleshooting it is not covered.
12:13 Aurelgadjo reddydodda: indeed, I meant to send you this link : https://docs.mirantis.com/openstack/fuel/fuel-8.0/operations.html#how-to-troubleshoot-amqp-issues
12:14 Aurelgadjo (especially for its crm_mon -fotAW -1 command)
12:23 reddydodda Aurelgadjo: Hi here is the output of crm_mon -fotAW -1 , can you help me is there anything wrong in rabbitmq server
12:23 reddydodda http://paste.openstack.org/show/529642/
12:27 Aurelgadjo reddydodda: this output says your corosync cluster is partitionned, you have to solve this first
12:28 vkulanov Folks, romcheg, ikalnitsky, bgaifullin, akasatkin, please, review two patches related to python-fuelclient (https://review.openstack.org/#/c/340030/, https://review.openstack.org/#/c/340219/)
12:29 reddydodda Aurelgadjo: Thank you for quick  reply. can you help me if we have any docs to resolve that issue ??
12:45 ekosareva akislitsky, apopovych, asvechnikov, bgaifullin, ikalnitsky, ikutukov please recheck https://review.openstack.org/#/c/282911 (resolved merge conflict)
12:52 Aurelgadjo reddydodda: that's some basic sysadmin skills there... connectivity check, reading logs etc
12:56 kaliya joined #fuel
12:56 zephcom joined #fuel
12:58 flerfb0rt joined #fuel
12:58 nokes joined #fuel
12:59 Julien-zte joined #fuel
13:00 nokes_ joined #fuel
13:18 bgaifullin joined #fuel
13:18 toMeloos Can anybody tell me what Ubuntu version is used in Fuel 9? Still 14.04 or is it 16.04?
13:20 athinkingmeat toMeloos: Ubuntu 14.04.4 LTS for nodes
13:20 toMeloos hmmm that's too bad
13:21 Sketch hmm, is network verification supposed to work on VM guest nodes?
13:22 Sketch "Repo availability verification using public network failed on following nodes Untitled (3f:bb)." ... yet if i log into it, i can ping archive.ubuntu.com just fine
13:23 code-R joined #fuel
13:28 pod2metra joined #fuel
13:30 pod2metra akislitsky, asaprykin, aroma, asvechnikov, ikutukov, zubchick, gkibardin review plz https://review.openstack.org/#/c/339005/
13:30 fritchie joined #fuel
13:39 Ravikiran_K joined #fuel
13:41 fatdragon joined #fuel
13:50 vkulanov bgaifullin, I've addressed your comment https://review.openstack.org/#/c/340030/
13:51 code-R joined #fuel
13:54 cartik joined #fuel
13:56 code-R_ joined #fuel
14:24 nokes joined #fuel
14:31 charz joined #fuel
14:54 AlexAvadanii Sketch: https://bugs.launchpad.net/fuel/+bug/1596981 ?
14:54 AlexAvadanii nevermind, that is only affecting fuel master
15:04 Sesso joined #fuel
15:24 toMeloos I've just created a bootstrap image with a newer kernel, and activated it, but the nodes keeps getting installed with the original fuel image. How can thate be?
15:26 toMeloos Can't find a way to check which bootstrap image was used, but the one deployed runs the 3.13 kernel and I'm pretty syre the image I created only has the 4.4 kernel
15:26 toMeloos *sure
15:30 mwhahaha toMeloos: might want to check the cobbler configuration on the master and verify which image is in use
15:30 toMeloos and how would i do that?
15:31 toMeloos 'fuel-bootstrap list' is saying my new image is active
15:31 mwhahaha you can use the cobbler commands on the master to list the configurations/systems
15:32 mwhahaha basically you'd be tracking down the cobbler configurations and file locations to make sure the image being used has the kernel you think it does
15:32 mwhahaha the locations depend on which version of fuel you're using
15:33 mwhahaha in the newer ones (9+) the image is in /var/www/cobbler/images/ubuntu_bootstrap
15:33 mwhahaha not sure where it sits in <=8 probably a similar location but it might be in a docker image
15:36 toMeloos we're on 8
15:36 toMeloos it is pointing to ubuntu_bootstrap
15:37 toMeloos and to /var/www/nailgun/bootstraps/active_bootstrap, which is in fact the right image
15:40 toMeloos I ran "fuel-bootstrap build --label 'with_xenial_kernel' --activate --package 'linux-image-generic-lts-xenial' --kernel-flavor linux-image-generic-lts-xenial" which completed succesfully (without the extra repo sources in /etc/fuel-bootstrap-cli/fuel_bootstrap_cli.yaml because that broke the build) as well... so how can this deployed node only contain a 3.13 image?!?
15:42 fatdragon joined #fuel
15:45 fatdragon joined #fuel
15:48 mwhahaha toMeloos: make sure the correct image is being used in the tftpboot folder
15:48 mwhahaha it's in /var/lib/tftpboot/images
15:48 mwhahaha you can also check the pxe configs for the hosts in there as well
15:51 toMeloos that just has bootstrap centos-x86_64 ubuntu_1404_x86_64 and ubuntu_bootstrap
15:52 mwhahaha right so check to see if the files in the ubuntu_bootstrap match your new image
16:22 Sketch what causes this: "Node is not ready for deployment: mcollective has not answered"
16:23 Sketch node is in bootstrap, it's been up over an hour, so everything should definitely be up by now...
16:23 Sesso joined #fuel
16:30 lanakars joined #fuel
16:33 xarses_ joined #fuel
16:35 DavidRama joined #fuel
16:44 code-R joined #fuel
16:50 code-R joined #fuel
16:53 code-R_ joined #fuel
16:59 DavidRama joined #fuel
16:59 DavidRama joined #fuel
17:26 javeriak joined #fuel
17:30 Egyptian joined #fuel
17:38 javeriak_ joined #fuel
17:38 raunak joined #fuel
17:46 fritchie joined #fuel
17:52 cr0wrx joined #fuel
17:54 cr0wrx anyone know the difficulty or whats involved in getting a second floating ip pool (still single interface for public network)? It sounds like it may be possible from looking online but haven't found any clear guides yet and doubt I will. I have a large public IP block exposed to openstack but want to break it down into several ranges that can be selected from when allocating (for firewall purposes outside of openstack)
18:02 xarses cr0wrx: if it's in the same subnet, fuel will do it for you during initial install, if your installed aready, you can simply perform the commands to add the pools in neutron
18:03 xarses if they are not part of the subnet, its a bit more complicated
18:03 cr0wrx what do you mean fuel will do it? I don't recall settings for multiple external networks
18:04 xarses also, if they are part of the subnet, and continuous, why would you create separate pools?
18:04 cr0wrx I have a /20 currently exposed (during fuel install) as the openstack public network, and one "external_net" network that you can allocate IPs from
18:05 cr0wrx However, I want some of that IP space to be for production, some dev, etc.
18:05 cr0wrx otherwise firewalls outside of openstack need to allow access to other production systems from the entire /20, rather than a smaller range
18:05 cr0wrx only need production VMs accessing external production services, not everything in openstack
18:08 xarses hmm, I was told that we added public ranges support in some way
18:09 xarses I don't see it in the ui
18:09 as0bu I was wondering if anything else needed to be done for this to be merged https://bugs.launchpad.net/fuel-plugins/+bug/1600400 thanks!
18:10 DavidRama joined #fuel
18:12 xarses as0bu: you need to poke the cores for the plugin https://review.openstack.org/#/admin/groups/1003,members
18:12 xarses probably via the openstack-dev ML
18:14 as0bu Ah thank you xarses
18:16 Sesso_ joined #fuel
18:16 cr0wrx maybe the ability to add multiple is through fuels 'node network group' function? idk, I haven't used it and just have the default network group. Or maybe it's a fuel 9.0 thing (I'm on 8.0)
18:16 cr0wrx in either case, I have an environment deployed already, so I think it would take some manual adjustments
18:17 xarses network groups don't do that, that's for having nodes in multiple L2 domains in the same openstack env
18:17 xarses I was on the understanding there was a (+) marker next to the field to create multiples
18:17 xarses which there is not
18:18 xarses so I need to chew on someone's ear when they get into the office
18:29 fritchie joined #fuel
18:37 bgaifullin joined #fuel
18:42 elo joined #fuel
19:10 DavidRama1 joined #fuel
19:10 raunak joined #fuel
19:12 DavidRama joined #fuel
19:13 thiagolib joined #fuel
19:15 DavidRama1 joined #fuel
19:15 DavidRama2 joined #fuel
19:19 ikutukov Please, see the update of Fuel plugins v5 spec https://review.openstack.org/#/c/271417/
19:19 obcecado joined #fuel
19:23 cr0wrx So I'm thinking maybe to break my /20 into multiple floating IP blocks, I need to 1) manually adjust existing subnet to be smaller cidr (to allow room for another network without overlapping cidr) 2) update qrouter/qdhcp to align with cidr changes 3) create new network / subnet with newly available IP block, bind network to existing router?
19:23 cr0wrx maybe?
19:34 DavidRama joined #fuel
19:40 DavidRama joined #fuel
19:45 raunak joined #fuel
19:54 DavidRama joined #fuel
20:41 cr0wrx joined #fuel
20:45 cr0wrx ok, so now I'm thinking it's still more involved. It seems you can only have one external network bound to one ovs label/interface. For the fuel install, it's deployed as a flat network bound to physnet1. I'm guessing I can't just add another bridge mapping with new label physnet2 pointing to same br-floating, so maybe I have to redo openvswitch configurations to spin up multiple bridges and labels and all the stuff that connects it toget
21:19 rans joined #fuel
21:19 rans Hello
21:20 rans need some help to deploy openstack ?
21:26 blahRus joined #fuel
21:35 Sesso joined #fuel
21:53 v1k0d3n joined #fuel
23:40 elopez joined #fuel

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