Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-01-10

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

All times shown according to UTC.

Time Nick Message
23:07 syt joined #fuel
23:25 sparc joined #fuel
23:39 kpimenova_ joined #fuel
23:41 MiroslavAnashkin joined #fuel
01:08 jouston joined #fuel
01:35 rmoe joined #fuel
02:19 IlyaE joined #fuel
03:46 xarses joined #fuel
03:48 rongze joined #fuel
04:45 mihgen joined #fuel
04:50 dhblaz joined #fuel
05:34 SergeyLukjanov joined #fuel
06:22 rongze joined #fuel
06:22 rongze joined #fuel
07:15 IlyaE joined #fuel
07:22 syt joined #fuel
07:47 e0ne joined #fuel
08:14 rongze joined #fuel
08:24 e0ne joined #fuel
08:26 jouston joined #fuel
08:37 e0ne_ joined #fuel
08:41 rongze joined #fuel
09:01 mihgen joined #fuel
09:01 teran joined #fuel
09:04 bogdando joined #fuel
09:13 bogdando joined #fuel
09:14 besem9krispy joined #fuel
09:18 IlyaE joined #fuel
09:23 pvbill joined #fuel
09:23 miguitas joined #fuel
09:25 pvbill Hi all - when trying to set up Mirantis 4.1 environment with Neutron, I get "can't find suitable CIDR" - anything specific need to do with IP ranges, VLANs or....?
09:28 AndreyDanin joined #fuel
09:31 miguitas_ joined #fuel
09:32 mattymo joined #fuel
09:33 akasatkin joined #fuel
09:33 AndreyDanin pvbill, where did you get this error? Do you use webUI or CLI?
09:34 besem9krispy joined #fuel
09:36 pvbill This is the Fuel dashboard at port 8000 - fresh off installing the Fuel node
09:37 mattymo pvbill, you may have had errors with the initial Fuel node setup
09:37 teran joined #fuel
09:37 mattymo maybe due to disk or network issues
09:38 teran joined #fuel
09:39 mattymo you may want to check the log here: /var/log/puppet/bootstrap_admin_node.log
09:41 pvbill ok, will check the log, thanks
09:46 mrasskazov1 joined #fuel
09:49 AndreyDanin pvbill, can you provide a diagnostic snapshot?
09:50 pvbill I sent in a snapshot to fuel support, but didn't get a response - somewhere else I should send?
09:52 pvbill (understand it's been Russian newyear/xmas week)
09:53 mattymo pvbill, yes most of us just returned from vacation (or come back Monday)
09:55 mattymo pvbill, do you have a ZenDesk ticket #?
09:57 pvbill no, don't see a response
09:58 mattymo pvbill, what email did you use to open a ticket?
10:00 pvbill should have been bill with prozeta.cz
10:00 pvbill let me just enter it again
10:01 mattymo I'm sorry, but I can't find it.
10:02 syt joined #fuel
10:08 pvbill I just reported another one, including uploading the log tgz, and it says no open or closed requests
10:08 pvbill https://mirantis.zendesk.com/requests...
10:08 pvbill user bill eldridge
10:09 syt joined #fuel
10:10 SergeyLukjanov joined #fuel
10:13 mattymo pvbill, try to open without the log
10:13 mattymo I believe there is an upload file size limit
10:13 mattymo I'm checking with our support team (I'm on the dev team)
10:19 pvbill ok, wasn't the attachment - needed a platform type & network type. but when I had an attachment , it didn't give me an error that these weren't selected - just silently failed. without attach, gives error.
10:20 pvbill #1267
10:24 mrasskazov1 joined #fuel
10:27 syt Is there any whitepaper or guideline on how to add a new release to an existing fuel installation in the future?
10:35 teran joined #fuel
10:40 VictorDenisov joined #fuel
10:44 rongze joined #fuel
10:44 bogdando joined #fuel
11:03 mattymo syt, there should be a blog post about that coming very soon
11:03 mattymo syt, stick around and I should have an answer later today
11:03 syt I will ;) thanks..
11:05 VictorDenisov joined #fuel
11:10 tatyana joined #fuel
11:41 vk joined #fuel
11:49 sparc joined #fuel
11:56 tatyana joined #fuel
12:00 syt joined #fuel
12:09 getup- joined #fuel
12:20 miguitas joined #fuel
12:37 teran_ joined #fuel
12:39 teran__ joined #fuel
13:08 SergeyLukjanov joined #fuel
13:11 tatyana_ joined #fuel
14:12 syt joined #fuel
14:43 syt joined #fuel
14:45 syt1 joined #fuel
14:49 sparc joined #fuel
14:55 teran joined #fuel
15:10 besem9krispy joined #fuel
16:28 Bomfunk joined #fuel
16:31 Nikolay joined #fuel
16:49 teran_ joined #fuel
16:57 teran joined #fuel
17:06 rshon joined #fuel
17:11 rmoe joined #fuel
17:19 SergeyLukjanov joined #fuel
17:30 vk joined #fuel
17:37 e0ne joined #fuel
17:38 e0ne joined #fuel
17:46 IlyaE joined #fuel
17:49 teran joined #fuel
17:49 pvbill joined #fuel
17:50 besem9krispy joined #fuel
17:54 xarses joined #fuel
18:07 angdraug joined #fuel
18:07 pvbill More issues on installing Mirantis 4.0:
18:08 pvbill 1) tried Ubuntu PXE option, but the generated hostname was too long; rebooted, edited it down when it came up again, and PXE then proceeded
18:09 pvbill 2) then got an error of Anaconda kickstart configuration file missing Bootloader configuration - "Reboot".  what to do?
18:17 MiroslavAnashkin pvbill: Did you mean you selected a boot option from bootstrap menu on PXE boot?
18:17 pvbill yes - it popped up 4 options, so 1st time selected Ubuntu, 2nd time selected CentOS
18:20 MiroslavAnashkin pvbill: Simply don't touch options in PXE bootstrap boot menu on the nodes. Cobbler automatically chooses the necessary one.
18:20 miguitas joined #fuel
18:20 MiroslavAnashkin pvbill: Ubuntu vs CentOS is selected when you create new environment in Fuel web interface.
18:22 pvbill ok, didn't think to let it take default #1, with it mentioning the 2 options. Will try letting it run when I'm back Monday. Thanks
18:26 chicken joined #fuel
18:38 SergeyLukjanov joined #fuel
18:41 daMaestro joined #fuel
18:41 daMaestro so I've been unable to get cinder working in 4.0
18:41 daMaestro not with LVM backed or ceph backed
18:41 daMaestro i've got a test cluster setup right now with two cinder nodes, but for some reason the controller also thinks it's a cinder host
18:42 daMaestro which i believe is the issue, as the VG does not exist on the controller, all storage is used by glance
18:42 daMaestro i've yet to have a healthcheck all pass
18:42 daMaestro what am I doing wrong here?
18:44 daMaestro http://fpaste.org/67447/89379488/
18:45 daMaestro http://fpaste.org/67449/93795441/ is the health check
18:58 Vidalinux joined #fuel
19:02 xarses daMaestro: what roles did you deploy in your current cluster?
19:12 daMaestro 1 x compute 1x control 2 x cinder
19:12 daMaestro http://fpaste.org/67457/13893811/ is what i just saw when doing the health check
19:13 daMaestro i manually stopped cinder-volume on the control node
19:13 daMaestro as it should not be running
19:13 daMaestro and that traceback was from the compute node
19:13 daMaestro i can manually add/remove a volume via horizon
19:14 xarses daMaestro: it should be running if you deployed any of the ceph roles, or selected any of the ceph options from the setup wizard or settings page
19:15 daMaestro no ceph nodes or options selected
19:15 daMaestro only cinder lvm
19:15 xarses hmm odd
19:16 daMaestro http://fpaste.org/67458/38938138/ is the error on the control node
19:16 daMaestro and i verified that the VG does not exist
19:16 daMaestro only glance exists on the control node
19:17 besem9krispy joined #fuel
19:17 xarses hmm, thats strange
19:17 syt joined #fuel
19:17 daMaestro so.. any thoughts on this exception.SecurityGroupCannotBeApplied() error?
19:17 daMaestro that seems to be what is blocking the healthcheck
19:18 daMaestro nova-nova.compute.manager ERROR: Error: Network requires port_security_enabled and subnet associated in order to apply security groups. specifically
19:18 daMaestro i'm using neutron with gre
19:18 xarses it shouldn't be running cinder volume unless you added the 'storage - cinder' on the controller, or ceph is running in the cluster
19:19 xarses on centos?
19:19 daMaestro where would i have added that? all i did was assign a node as a controller role
19:19 daMaestro yes, this is the centos undercloud
19:19 daMaestro i did do a test with the lts undercloud and observed the same errors in the healthchecks
19:19 daMaestro though i did not check the logs
19:24 xarses it would have been on the add roles page after you selected to add the node to a cluster
19:25 daMaestro ah, yeah that node is exclusively a controller
19:25 xarses are you using vlan tags on any of the management storage or public networks?
19:26 daMaestro http://imagebin.org/286057
19:26 daMaestro yes
19:27 daMaestro http://imagebin.org/286058
19:29 xarses its possible under centos to have issues with OVS when vlan tags are used that would cause network issues
19:29 daMaestro we are aware that physical nic layout is not optimal, but before i deploy 10g to these c6100s i'd like to get something proven using dual 1g ports
19:29 xarses daMaestro: its supportable the way you have it
19:30 daMaestro so do i need to define a non-default vlan range for the neutron l2 config?
19:30 daMaestro yeah, i just mean don't judge me putting guest, management and storage on the same nic
19:30 daMaestro we know we have risk there ;-)
19:31 daMaestro we've actually already ran into issues with only being able to reserve one gateway IP
19:31 xarses daMaestro: you're running a single controller, all other best practice griping left there
19:31 daMaestro we use vrrpe and each router has an active IP on the network
19:31 daMaestro so we have to reserve more then just one gateway IP
19:32 daMaestro we've bought some time with that issue by moving our vrrpe IPs to the top of the cidr
19:32 daMaestro xarses, k, thanks for understanding ;-)
19:32 daMaestro i actually went back to a single controller to debug not being able to pass the healthcheck
19:32 xarses you can do this to enable vlan splinters if you see that you have problems talking to each nodes interfaces  ovs-vsctl set interface eth1 other-config:enable-vlan-splinters=true
19:32 daMaestro fewer nodes to have to watch
19:33 xarses daMaestro: first thing, i'd check that each assigned address can ping the other nodes on the network
19:33 daMaestro that checks out
19:33 xarses make sure we dont have some issue with splinters or the physical network
19:34 daMaestro k, let me do a full audio of network connectivity
19:34 IlyaE joined #fuel
19:34 daMaestro so we should be using ovs vlan splinters hard trunks?
19:35 daMaestro if using tagged networking for inner host connectivity?
19:35 xarses most cases soft splinters works, which is the ovs-vsctl line above
19:37 xarses after you might be able to use this to help with the secuirty group error https://answers.launchpad.net/neutron/+question/229757
19:38 xarses but the health tests should have done that for you
19:42 daMaestro k, i just verified network connectivity using nmap between all the nodes, that is good to go
19:43 daMaestro xarses, that seems to say i need to go manually create all those network?
19:44 xarses you can give it a poke, it should be using net04
19:45 daMaestro okay, so i should redeploy with soft splinters first then i'll do some more health check runs
19:48 xarses_ joined #fuel
19:50 xarses joined #fuel
19:51 xarses sorry, kernel panic
20:00 daMaestro ouch
20:03 xarses that was the worse restore of chrome ever
20:19 vk joined #fuel
20:27 sharmanov joined #fuel
20:44 daMaestro well.. so controller, compute and cinder and cinder-volume and tgt is still being configured and started on the controller
20:51 daMaestro okay and ovs vlan splinters has caused networking to not work
20:59 daMaestro i'm trying again with hard vlan splintering
21:26 daMaestro hard vlan splintering got that working
21:43 xarses_ joined #fuel
21:45 rmoe joined #fuel
21:58 IlyaE joined #fuel
22:30 nurla joined #fuel
22:31 meow-nofer__ joined #fuel
22:31 zasimov joined #fuel
22:31 Nikolay joined #fuel
22:31 mattymo joined #fuel

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