Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2013-12-11

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

All times shown according to UTC.

Time Nick Message
23:06 rmoe joined #fuel
23:07 teran joined #fuel
01:24 xarses joined #fuel
02:01 rongze joined #fuel
02:06 fandikurnia01 joined #fuel
02:46 rongze joined #fuel
03:01 Vidalinux joined #fuel
03:26 ArminderS- joined #fuel
03:38 wputra joined #fuel
03:39 archon1st joined #fuel
03:41 wputra hi all
03:42 wputra can somebody tell me why fuel need untagged vlan 0?
03:42 wputra in my case, this vlan need in privat interface
03:44 teran joined #fuel
04:01 teran joined #fuel
04:25 vk joined #fuel
04:47 teran joined #fuel
05:20 rongze joined #fuel
05:59 mihgen joined #fuel
06:02 IlyaE joined #fuel
06:03 rongze joined #fuel
06:15 rongze joined #fuel
07:02 mihgen joined #fuel
07:08 xdeller joined #fuel
07:16 IlyaE joined #fuel
07:18 miguitas joined #fuel
07:41 teran joined #fuel
07:58 vkozhukalov joined #fuel
08:09 SteAle joined #fuel
08:16 ruhe joined #fuel
08:22 ruhe joined #fuel
08:47 rongze joined #fuel
08:59 e0ne joined #fuel
09:19 e0ne_ joined #fuel
09:28 tatyana joined #fuel
09:30 fandikurnia01 joined #fuel
09:33 xdeller joined #fuel
09:35 fandikurnia01 joined #fuel
09:36 rvyalov joined #fuel
09:36 miguitas joined #fuel
09:38 miguitas Hi guys, I have a problem with fuel 3.2.1, quantum and rabbitmq cannot connect to rabbitmq, in rabbit log apperars "AMQPLAIN login refused: user 'nova' - invalid credentials"
09:40 miguitas I deployed ubuntu  release
09:44 anotchenko joined #fuel
09:54 alex_didenko joined #fuel
09:56 SergeyLukjanov joined #fuel
10:00 anotchenko joined #fuel
10:08 anotchenko joined #fuel
10:17 nurla joined #fuel
10:17 xdeller miguitas: HA or non-HA mode?
10:19 xdeller and which type of isolation? can you please file a bug on the https://bugs.launchpad.net/fuel with diagnostic snapshot attached?
10:19 miguitas multi node non-HA
10:19 miguitas I fix the problem by the moment executing puppet agent
10:19 miguitas on controller
10:20 xdeller you meant second run?
10:20 miguitas yes
10:21 xdeller anyway can you please create a bug? it looks very unusual
10:21 miguitas It's weird because I did not touch any config file
10:21 miguitas ofcourse
10:22 miguitas thanks for the help
10:23 xdeller thanks for report :)
10:32 Fecn joined #fuel
10:39 anotchenko joined #fuel
10:55 ArminderS- joined #fuel
10:55 Arminder joined #fuel
11:20 evgeniyl joined #fuel
11:24 anotchenko joined #fuel
11:27 evgeniyl joined #fuel
11:29 Vidalinux joined #fuel
11:33 bogdando joined #fuel
11:44 rongze joined #fuel
11:47 anotchenko joined #fuel
12:00 e0ne joined #fuel
12:10 SergeyLukjanov joined #fuel
12:38 SteAle joined #fuel
12:50 fandikurnia01 hello
12:51 fandikurnia01 can someone help how to implement this case using multiple vlan trunk as this reference  http://libvirt.org/formatnetwork.html#elementVlanTag
12:57 tonyha joined #fuel
12:57 manashkin_ fandikurnia01, what exactly you are going to implement? Are you going to use Nova-Network or Neutron with VLAN segmentation?
13:00 anotchenko joined #fuel
13:01 anotchenko joined #fuel
13:02 fandikurnia01 hi manashkin.. sorry for short information, exactly our network team using so many vlan : vlan 6,7,8,4090,4091,4092.  we are modify  patch.pp using trunk in br-ex. But after finish deploy, our subnet ip allocate to default vlan (vlan0)
13:04 fandikurnia01 so when we cannot  reach network subnet (vlan8),
13:05 fandikurnia01 we are using neutron network
13:09 manashkin_ fandikurnia01, are you use Fuel GUI or CLI to create your network settings? What Fuel version do you use?
13:11 alex_didenko joined #fuel
13:11 ArminderS- joined #fuel
13:11 fandikurnia01 manashkin, to discover and modify we use GUi, and then we modify network setting using cli
13:12 fandikurnia01 manashkin, we use fuel ver. 3.2.1
13:15 mihgen joined #fuel
13:43 manashkin_ May we have your network settings from Fuel GUI Networks tab?
14:42 anotchenko joined #fuel
14:56 fandikurnia01 joined #fuel
14:59 fandikurnia01 joined #fuel
15:03 SergeyLukjanov_ joined #fuel
15:08 vkozhukalov joined #fuel
15:10 SergeyLukjanov joined #fuel
15:26 Shmeeny joined #fuel
15:31 IlyaE joined #fuel
15:45 Shmeeny joined #fuel
15:50 albionandrew joined #fuel
15:54 rmoe joined #fuel
16:03 albionandrew Morning xarses - The fix you did last night that was on launchpad means I can just run nail gun again? Or do I need to do a new installation?
16:18 xarses you can just apply it to the discovery node and see if it works, if it does, you should re-pack the initrd image that cobbler uses so you don't have to totally rebuild
16:18 xarses otherwise, the image is rebuilt during the iso build process
16:20 albionandrew I'll wait for dhblaz see what he wants to do. Thanks
16:26 e0ne joined #fuel
16:47 teran joined #fuel
16:50 teran_ joined #fuel
16:51 rongze_ joined #fuel
16:52 rongze__ joined #fuel
16:53 nurla1 joined #fuel
16:54 ArminderS- joined #fuel
17:17 rongze joined #fuel
17:17 nurla1 joined #fuel
17:20 SergeyLukjanov joined #fuel
17:23 angdraug joined #fuel
17:27 rmoe joined #fuel
17:31 Vidalinux joined #fuel
17:44 xarses joined #fuel
17:50 Fecn joined #fuel
18:15 albionandrew xarses - We only have a build env for 4.0
18:29 xarses albionandrew: Isn't that what you are deploying?
18:30 albionandrew xarses - no we are using the 3.2.1 img
18:30 xarses albionandrew: ahh
18:32 albionandrew xarses - will that fix still apply?
18:32 xarses albionandrew: let me check, but i think so
18:33 albionandrew xarses - Thanks
18:40 IlyaE joined #fuel
18:43 xarses yes, it applies to the 3.2.1 image
18:44 albionandrew xarses - Tried on one node - http://pastebin.com/X0mzdYL7
18:44 albionandrew do I need to do something else?
18:50 albionandrew xarses it doesn't seem to change anything.
18:53 albionandrew xarses - dhblaz had me run - http://pastebin.com/mVJU9uc6
18:59 xarses I'll have to wait to discuss this with my co-woker when he has a moment
19:03 albionandrew xarses - Thanks
19:11 Shmeeny We just deployed Fuel 3.2.1 and are trying to get Ceph configured properly
19:11 Shmeeny we're having some issues specifying a dedicated journal drive
19:12 Shmeeny In the Fuel Web UI we configure with 1 OS drive, 1 Journal drive, and 4 OSD drives per Ceph node
19:12 Shmeeny The UI says it works, but when you SSH in to each node, the journal drive isn't mounted
19:13 Shmeeny and when you look in the running Ceph config it is pointing the journal to a directory that is mounted on one of the dedicated OSD drives
19:14 Shmeeny this is problematic because we'd like to use the SSDs for journal drives :)
19:15 Shmeeny when I grep through all the puppet files on the fuel node, I don't see anything that actually sets the configuration value "osd journal", but I do see a hard coded osd_journal_size of 2048 (which seems less than ideal for any production deployment with reasonable hardware)
19:15 rongze joined #fuel
19:15 Shmeeny Am I missing something that would allow Fuel to properly allocate a dedicated journal drive?
19:17 Shmeeny by default Ceph will just use /var/lib/ceph/osd/$cluster-$id/journal if the osd journal setting is omitted from the config
19:17 Shmeeny which seems to be the behavior I'm seeing
19:18 jhurlbert joined #fuel
19:18 xarses Shmeeny: there is a fact in the ceph module that ceph::osd uses
19:22 xarses Shmeeny: we should not need to mount the journal with how it was implemented. You should find that $clusert-$id/journal is a symlink to the journal /dev/<ssd>
19:23 Shmeeny I can follow that fact now. I'll have to check in a minute when the nodes are deployed for the symlink
19:23 Shmeeny if that is the case, I feel dumb, but at least it is there!
19:23 Shmeeny thank you
19:24 xarses Shmeeny: It should work both ways, let us know if there are issues on it
19:25 Shmeeny definitely the running size of the journal was 2048 though
19:25 Shmeeny so that seems less than ideal
19:29 e0ne joined #fuel
19:30 xarses Shmeeny: Correct, the default journal size is 2G. For a number of implementations this is a good value. The recommended ceiling is is 10G
19:34 Shmeeny Gotcha so running 10Gbps DAC, we'd see a max of around 1000MB/sec
19:35 Shmeeny so the journal could be full within 2 seconds if not pushed to the disks, or am I looking at that right?
19:39 xarses Shmeeny: Sounds about right. The journal on a SSD is intented to catch random IO so that it can be serliazed better before sending to the spinning media since it performs much better in linear IO
19:40 Shmeeny Thanks!
19:48 xarses If you spooling a lot of data from a DAC during tests, you will likely want to experiment with your cluster design. you might want to reduce the replicas to 1 for the pool that you use to spool that data in
19:48 xarses and then move it to a two or three once it's received
19:48 xarses or maybe a pure ssd pool to start
19:49 xarses keep in mind, that in your example, you will have 1000MB/sec * num replicas * 2 (journal plus OSD) while writing
19:50 xarses so you need to be careful to have enough bandwidth between the osd devices on the storage network
19:50 e0ne joined #fuel
19:57 nurla1 left #fuel
20:05 Fecn joined #fuel
20:06 albionandrew xarses Any other suggestions for what I could try re the agent problem?
20:13 teran joined #fuel
20:30 Fecn joined #fuel
20:33 xarses albionandrew: can you try adding http://paste.openstack.org/show/54846/ on top of the other patch?
20:33 albionandrew xarses ok will put that on the top.
20:34 xarses hang on it cut the output
20:35 xarses albionandrew: http://paste.openstack.org/show/54847/ is correct
20:35 albionandrew xarses - Thanks
20:39 tatyana joined #fuel
20:40 albionandrew xarses - That looks like it fixed it on one node, I will try it on the other two I was having issues with
20:45 xarses albionandrew: yay!
20:52 teran joined #fuel
20:56 albionandrew xarses it looks like that fixes the 3 machines I was having trouble with I've put the amended file here just incase someone is looking back at the IRC logs at some point - http://pastebin.com/j27WTZqQ Thanks again!
20:58 xarses albionandrew: will add it to the gerrit review shortly
20:59 xarses Please comment on the ticket so I can push to get the review approved
21:02 albionandrew xarses - Comment made. Thanks again
21:04 Fecn joined #fuel
21:10 Shmeeny joined #fuel
21:21 e0ne joined #fuel
21:25 Fecn1 joined #fuel
21:35 teran joined #fuel
21:38 mrasskazov joined #fuel
21:40 fandikurnia01 joined #fuel
21:42 e0ne_ joined #fuel
21:44 mrasskazov joined #fuel
22:05 IlyaE joined #fuel
22:12 e0ne_ joined #fuel
22:13 Fecn joined #fuel
22:16 mrasskazov joined #fuel
22:19 Shmeeny joined #fuel
22:37 teran joined #fuel

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