Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2015-01-12

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

All times shown according to UTC.

Time Nick Message
00:22 thumpba joined #fuel
00:30 jaypipes joined #fuel
01:43 teran_ joined #fuel
01:45 teran__ joined #fuel
02:21 Longgeek joined #fuel
04:02 aleksandr_null joined #fuel
04:21 Longgeek joined #fuel
05:16 HeOS joined #fuel
05:32 Longgeek joined #fuel
05:53 monester_laptop joined #fuel
06:28 saibarspeis joined #fuel
06:44 fandi joined #fuel
06:46 avorobiov joined #fuel
06:47 fandi joined #fuel
07:00 benrodrigue joined #fuel
07:08 Longgeek joined #fuel
07:24 fandi joined #fuel
07:26 sambork joined #fuel
07:35 Miouge joined #fuel
07:37 subscope joined #fuel
07:42 okosse joined #fuel
07:44 HeOS joined #fuel
08:05 sambork joined #fuel
08:06 dklepikov joined #fuel
08:07 e0ne joined #fuel
08:10 ChrisNBlum joined #fuel
08:11 avlasov joined #fuel
08:22 Philipp__ joined #fuel
08:25 saibarspeis joined #fuel
08:27 mescanef joined #fuel
08:47 devstok joined #fuel
08:48 devstok is possible to install the 3 controllers in HA enviroment on 3 separated proxmox?
08:48 kaliya hi devstok do you mean on different network segments?
08:49 devstok no
08:50 devstok I mean to use 3 machine
08:50 devstok install proxmox on these machine
08:50 devstok and then install 3 controller on the 3 proxmox
08:50 devstok in HA
08:51 kaliya install 3 HA controllers on 3 VMs?
08:51 devstok is possible or this try is not feasible?
08:51 devstok yes kaliya
08:52 kaliya the main consideration is that the controllers must be on the same network segments
08:52 devstok network segments is not a problem
08:52 kaliya 3 controllers on 3 VMs is a default setup, yes
08:53 devstok isn't a default setup for sure but is feasible? and the performance?
08:54 kaliya performances depends especially on I/O, and if you activate or not Ceilometer on controllers
08:54 devstok ceilometer is heavy for the performance?
08:55 kaliya yes
08:56 SS-Yuriy joined #fuel
08:57 devstok ceilometer can be installed on other server like computes?
08:59 kaliya yes
09:01 devstok can be a good solution to get better performance? also for the horizon ?
09:01 kaliya horizon must be on controllers
09:02 devstok yes but I see horizon with ceilometer is slower than grizzly horizon
09:10 sovsianikov joined #fuel
09:17 dkusidlo joined #fuel
09:19 sambork joined #fuel
09:21 hyperbaba joined #fuel
09:29 evg left #fuel
09:30 evg joined #fuel
09:41 xek joined #fuel
09:42 azemlyanov joined #fuel
09:45 kaliya joined #fuel
09:55 dkusidlo joined #fuel
10:03 stamak joined #fuel
10:15 dkusidlo joined #fuel
10:17 monester_laptop joined #fuel
10:27 e0ne joined #fuel
10:30 taj joined #fuel
10:30 stamak joined #fuel
11:20 sambork joined #fuel
11:21 teran joined #fuel
11:43 Miouge joined #fuel
11:55 ykotko joined #fuel
11:58 Miouge joined #fuel
12:03 omolchanov joined #fuel
12:03 teran joined #fuel
12:04 dkusidlo joined #fuel
12:04 alexdepalex joined #fuel
12:06 alexdepalex Howdy. I'm using Fuel 6.0 and I'm trying to configure bonding. I've got 2 dedicated interfaces for public, storage and management, but I don't get the option to bond these.
12:06 alexdepalex I can't place a tick next to the interface and the bond button at the top is also not there.
12:09 teran joined #fuel
12:12 evg_mobile joined #fuel
12:14 seeg alexdepalex, bond button will show when at least 2 interfaces are checked, I think Admin network cannot be moved nor bound to another interface
12:16 sambork joined #fuel
12:18 ddmitriev joined #fuel
12:22 alsadi joined #fuel
12:23 alsadi hi every body, this is my first time in your channel
12:23 alsadi most likely we have met in Paris in last openstack summit
12:24 teran_ joined #fuel
12:24 alsadi I'm testing fuel 6.0 on kvm virtual machines, and I noticed some high-load (compared to packstack)
12:24 alexdepalex Thanks seeg. Unfortunately, I can't check any interfaces.
12:25 alsadi for example hear-engine appears in top while nothing is trying to use it I guess
12:25 alexdepalex For 1 node in particular, I have 4 interfaces, 1 for admin, 1 unusded, and 2 10Gb.
12:25 seeg alexdepalex, also, interfaces can be configured only for undeployed nodes
12:25 alexdepalex I can drag vlans to the interfaces, but not check them.
12:25 alexdepalex All nodes are undeployed.
12:26 MiroslavAnashkin alexdepalex: Please select 2 Nics on the configuration page.
12:27 alsadi I've raised the log level to see what is going on, here is a strange log message http://pastebin.com/tsUJdHd6
12:28 MiroslavAnashkin http://docs.mirantis.com/fuel/fuel-6.0/user-guide.html#setting-nic-bonding-nic-aggregation
12:28 seeg alexdepalex, also from what I see in the code, the network provider must be neutron for bonding to be available
12:29 alsadi it says it's connected to MQ still it says it "Timed out waiting for RPC response"
12:29 alexdepalex http://i61.tinypic.com/ogkjue.png
12:29 alexdepalex I saw that as well, but do I need to specify that explicitly somewhere? Should be neutron and OVS.
12:29 alsadi here is glances/top screenshot http://i.imgur.com/Jbyli2w.png
12:30 alexdepalex I've selected Neutron with GRE.
12:34 seeg alexdepalex, when you go to nodes list screen, is the node you're modifying in 'pending addition' state? it cannot be in 'error'
12:35 alexdepalex It's in pending addition.
12:35 seeg ok, also mellanox and iser need to be disabled
12:36 alexdepalex They are
12:36 alexdepalex I've run through the planning and user guid already. Just to make sure I wasn't doing anything funny.
12:36 seeg hm, then it's strange you don't have checkboxes
12:36 seeg what browser are you using? are there errors in console log?
12:37 alexdepalex Safari. I can try chomre and ff.
12:37 seeg please try chrome
12:37 seeg though should work in safari too
12:37 alexdepalex No difference.
12:38 seeg when you right-click on page, choose 'inspect element' and go to console in chrome -- do you have any errors there?
12:38 alexdepalex No errors
12:40 alexdepalex There's a div physical-network-checkbox, but it's empty.
12:41 seeg hm, ok, no i have it also on my iso
12:43 alexdepalex How do you mean?
12:44 seeg I also don't have the checkboxes, i'll try to see why
12:44 jkirnosova seeg bonding is experimental feature, do you have 'experimental' feature group on your iso?
12:45 seeg aah, yes, true, it is experimental in 6.0
12:46 alexdepalex Ah. I read the guides, not the release notes ;)
12:47 alexdepalex Ok. I found how the enable the experimental features. I'll give that a try. Thanks for your help.
12:48 seeg yeah, thanks jkirnosova I forgot to add this when rewriting to React :)
13:00 moizarif joined #fuel
13:02 Longgeek joined #fuel
13:02 Miouge joined #fuel
13:06 dkusidlo joined #fuel
13:13 alsadi anybody please! would you take a look this this paste and tell me if it's normal or not  http://pastebin.com/tsUJdHd6
13:14 alsadi it says it's connected to MQ but it then says it "Timed out waiting for RPC response"
13:19 seeg alsadi, time out probably means that the other end didn't respond in time it's not MQ issue
13:19 agordeev joined #fuel
13:19 seeg i don't know what can be wrong though from the logs you sent
13:23 Samos123 does fuel support deploying a software raid?
13:26 alsadi seeg, thank you. but the MQ and heat-engine are on the same node (controller node) and both are running. is it normal for MQ to timeout?
13:27 seeg it's not MQ timeouting but rather the oslo messaging that didn't get RPC response in time, please see first answer to https://openstack.redhat.com/forum/discussion/404/timeout-timeout-while-waiting-on-rpc-response-with-qpid/p1
13:29 MiroslavAnashkin alsadi: Sometimes it happens with RabbitMQ when it stalls. I would restart RabbitMQ cluster completely - please wait 10 mins, I'll check how to do it in 6.0
13:32 alsadi seeg, MiroslavAnashkin, thank you, I'll read that and get back to you. would you please take a look on this screenshot http://i.imgur.com/Jbyli2w.png
13:44 moizarif joined #fuel
13:46 MiroslavAnashkin alsadi: BTW, single loaded CPU core is normal for controllers. There are DB sync, Corosync heartbeats, HA services syncs etc. It may take up to whole single CPU core.
13:48 alsadi MiroslavAnashkin, but I choose no HA in fuel when I installed the controller. does it do all this in this case too?
13:50 MiroslavAnashkin alsadi: Ah, then please try to restart RabbitMQ server with `service rabbitmq-server restart` first.
13:50 MiroslavAnashkin This high load may be result of "timed out _error_callback" RabbitMQ errors
13:53 MiroslavAnashkin After the RabbitMQ server restart - you may have to wait up to 1 hour and monitor queues with rabbitmqctl list_queues. Doubled queues with accumulating messages possible for 1-2 hours due to newer oslo messaging attempts to re-use existing queues.
13:54 MiroslavAnashkin These duplicated queues should disappear after 1-2 hours.
14:02 Miouge joined #fuel
14:06 alsadi MiroslavAnashkin, it's 2-core VM, I've restart rabbitmq and still the same, let's wait more
14:07 alsadi "rabbitmqctl list_queues | wc -l" gives 83
14:08 alsadi and nothing in " rabbitmqctl list_queues | grep -v '\s0$' " all queues are empty
14:13 MiroslavAnashkin That's good.
14:16 MiroslavAnashkin Are there new errors like http://pastebin.com/tsUJdHd6 after RabbitMQ restart?
14:19 alsadi MiroslavAnashkin, in which log ?
14:19 alsadi heat ?
14:20 benrodrigue joined #fuel
14:20 MiroslavAnashkin In the same log where these messages appeared previously
14:21 alsadi MiroslavAnashkin, Oh! that's my original pastebin, hmmm
14:22 MiroslavAnashkin yes)
14:22 alsadi I see connection refused (when I restarted) then the same old things (timeouts)
14:30 Miouge joined #fuel
14:41 Longgeek joined #fuel
14:52 dkusidlo joined #fuel
15:16 merdoc joined #fuel
15:22 merdoc hi all
15:30 teran joined #fuel
15:30 blahRus joined #fuel
15:31 jobewan joined #fuel
15:33 miroslav_ joined #fuel
15:36 dkusidlo joined #fuel
15:46 aarefiev_ joined #fuel
15:51 e0ne joined #fuel
15:59 e0ne_ joined #fuel
16:05 moizarif joined #fuel
16:05 Miouge joined #fuel
16:13 mpetason joined #fuel
16:23 taj joined #fuel
16:25 taj joined #fuel
16:27 taj joined #fuel
16:29 Longgeek_ joined #fuel
16:37 e0ne_ joined #fuel
17:06 taj_ joined #fuel
17:10 mattgriffin joined #fuel
17:17 rmoe joined #fuel
17:18 kupo24z joined #fuel
17:21 xarses joined #fuel
17:21 emagana joined #fuel
17:32 benrodrigue joined #fuel
17:40 aarefiev joined #fuel
17:49 thumpba joined #fuel
18:11 thumpba joined #fuel
18:12 mpetason_ joined #fuel
18:14 alsadi joined #fuel
18:15 thumpba_ joined #fuel
18:15 alsadi joined #fuel
18:16 mpetason joined #fuel
18:29 alexdepalex_ joined #fuel
18:34 alsadi arp takes too much time while arp -n does not even need a single millisecond < http://paste.openstack.org/show/156510/
18:35 alsadi MiroslavAnashkin, seeg given that my fuel node is down (10.20.0.2) what do you think?
18:38 alsadi maybe the root to my problems is that my "public ip" is not set correctly
18:39 benrodrigue joined #fuel
18:42 MiroslavAnashkin alsadi: Fuel has bug - if fuel master node is down, NTP does not work at all. Depending on configuration,  some routing services may not work as well. Please start up the fuel master node for the greater good.
18:44 miroslav_ joined #fuel
18:46 alsadi MiroslavAnashkin, I set my public address to be "172.16.0.2" and alike but this was wrong my kvm does not have such IP should I re-install it with proper public network settings?
18:47 HeOS joined #fuel
18:49 mutex joined #fuel
18:50 MiroslavAnashkin You should configure your KVM virtual network to be routable from outside. Better if your public IP belongs to the already configured IP range you use for KVM virtual network.
18:51 MiroslavAnashkin Please check this docs section docs.mirantis.com/fuel/fuel-6.0/virtualbox.html
18:52 MiroslavAnashkin And this http://docs.mirantis.com/fuel/fuel-6.0/virtualbox.html#deployment-configuration-to-access-openstack-api-and-vms-from-host-machine
18:52 alsadi this this related to the high load?
18:53 MiroslavAnashkin No, it is related to the public address configuration...
18:54 e0ne joined #fuel
18:55 alsadi I'm searching for reasons why I have highload and noted that arp takes too much time which uses DNS which is not there .. this is how I figured out about the public network thing,
18:55 alsadi I'm not actually interested to fix public network issue by itself.
18:55 alsadi MiroslavAnashkin, another thing I've collected packets for a period of time using tcpdump and sorted them to see what is going on, here is the output
18:56 alsadi MiroslavAnashkin,  http://paste.openstack.org/show/156516/
18:58 alsadi MiroslavAnashkin, so they are mysql, beam.smp, memcached then neutron
19:01 alsadi beam.smp is MQ
19:01 MiroslavAnashkin All 4 services are OK
19:03 alsadi MiroslavAnashkin, thank you very much I'll try fix my public net and see if this fix it
19:06 alsadi have a good time, good bye
19:08 taj_ joined #fuel
19:14 alecv joined #fuel
19:16 ykotko joined #fuel
19:20 e0ne joined #fuel
19:31 teran joined #fuel
19:40 homegrown joined #fuel
19:47 mrasskazov joined #fuel
19:48 homegrown Playing around with Fuel6 I cannot seem to get Swift to show up in the dashboard, just when using ceph... am i missing something?
19:55 thumpba joined #fuel
19:57 alexdepalex_ joined #fuel
19:58 thumpba_ joined #fuel
20:03 e0ne joined #fuel
20:11 MiroslavAnashkin homegrown: Please check http://docs.mirantis.com/fuel/fuel-6.0/user-guide.html#storage-backend-for-cinder-and-glance
20:12 MiroslavAnashkin Swift is installed only if you have chosen Default backend for Glance and your env is Multi-node HA mode
20:19 ddmitriev joined #fuel
20:29 bdudko joined #fuel
20:29 aglarendil joined #fuel
20:29 apalkina joined #fuel
20:30 wayneeseguin joined #fuel
20:30 meow-nofer joined #fuel
20:32 obcecado joined #fuel
21:31 samuelbartel_ joined #fuel
21:38 aleksandr_null joined #fuel
21:41 angdraug joined #fuel
22:17 e0ne joined #fuel
23:02 aleksandr_null joined #fuel
23:05 mpetason joined #fuel
23:22 xarses joined #fuel

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