Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2013-11-12

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

All times shown according to UTC.

Time Nick Message
00:14 boldhawk joined #fuel
00:15 boldhawk guys - need some advice. My NIC's change order between discovery and  provision. I think it's to do with the fact I am trying to install an Ubuntu based stack. But currently when choosing the NIC layout, it seems the NIC name is the key rather than the MAC Address.
00:17 boldhawk so when I come to deploy, various problems occur from not installing Ubuntu to not being able to start the openstack installation after the ubuntu install completes.
01:33 xarses joined #fuel
01:45 boldhawk Sorry to ask the question again - guys - need some advice. My NIC's change order between discovery and  provision. I think it's to do with the fact I am trying to install an Ubuntu based stack. But currently when choosing the NIC layout, it seems the NIC name is the key rather than the MAC Address.
02:01 rmoe joined #fuel
02:48 boldhawk joined #fuel
02:48 boldhawk not sure if I got a response to my previous question - I got booted off by our internal IT firewall.
02:50 boldhawk_ joined #fuel
02:50 boldhawk_ any ideas?
02:53 boldhawk_ dfg
02:56 dhblaz joined #fuel
03:00 teran joined #fuel
03:03 boldhawk joined #fuel
03:03 boldhawk apologies - i keep getting dropped
03:04 boldhawk Two questions:
03:04 boldhawk 1) How do I stop the NIC order changing on Ubuntu? I have two NICS that keep swapping places and stopping the Ubuntu installation because it cannot find the DHCP server
03:05 boldhawk 2) I manage to successfully deploy a centos stack, (ubuntu still is causing problems) - but I chose my ceph journal to be on my second hdd however I found after some performance tests that the journal had been created on each osd drive instead of what I sepcific in Fuel.
03:06 boldhawk Any guidance would be appreciated as I'm not sure if there are bugs or self inflicted.
03:06 rmoe joined #fuel
03:25 teran joined #fuel
03:29 ArminderS joined #fuel
03:45 boldhawk ok - so the first workaround seems to be to change the kernel parameters in cobbler for the affected node from netcfg/choose_interface=eth05 to netcfg/choose_interface=eth0
03:51 jkyle joined #fuel
03:51 rmoe joined #fuel
03:56 teran joined #fuel
04:06 rmoe joined #fuel
04:32 mihgen joined #fuel
04:35 rmoe joined #fuel
04:59 mihgen joined #fuel
05:11 boldhawk hey - is anyone active on here?
05:27 xarses ya
05:31 boldhawk hey xarses
05:32 boldhawk Seeing a lot of differences between CentOS and Ubuntu. I rebuilt the fuel server and I'm getting further so I might have caused my own problems with the SSL certificates.
05:33 boldhawk NIC order in Ubuntu is problematic.
05:34 boldhawk And the Ceph Journal always seems to be added to the OSD disk regardless of how I setup the disk layout
05:37 boldhawk Additionally the journal size always appears to be 2048 in /etc/ceph/ceph.conf
05:53 mrasskazov joined #fuel
06:11 xarses a journal is required for ceph
06:11 xarses if you dont allocate dedicated space 2Gb will be used on the OSD partation
06:11 xarses let me clarify, a journal is required for each OSD device
06:30 ogzy joined #fuel
06:31 ogzy is it possible to choose xenserver at the Fuel web ui while creating compute nodes, in addition to KVM and QUEMU
06:41 xarses ogzy: you would need to add the relavant xenserver packages and modify the openstack/manifests/compute.pp manifests appropriately. Otherwise, its currently not supported
06:43 ogzy xarses: can you tell me more about the xenserver packages?
06:44 xarses ogzy: nope, the only thing i know is xen requires a special kernel
06:44 ArminderS xarses: since you have been working up with vmware, any chances to have support for vmware esx(i) in near future?
06:44 ArminderS vmware esx(i) for compute
06:44 * xarses grins
06:44 ArminderS sorry if i made you grin :p
06:45 xarses if i where to guess, something like that would make sense
06:46 ArminderS nicely put
06:49 ogzy xarses: where should i start to create compute node on xen via fuel ui?
06:49 ogzy xarses: i already installed fuel
06:51 xarses ogzy: like i said, you'd need to find packages for xenserver, and add them to the fuel server's yum repo, and then you would need to modify the puppet manifests for openstack compute so that it installs and configures it
06:51 xarses sorry, its bed time for me, good night
06:52 ArminderS good night xarses
06:53 ogzy may i change the default centos installation to ubuntu based for fuel ? the current iso installs centos, http://docs.mirantis.com/fuel-dev/develop/env.html says i can create my iso right?
06:53 ArminderS ogzy: why you need to run xenserver?
06:54 ArminderS and do you intend to run xenserver on all the compute nodes
06:54 ArminderS or mix of kvm and xenserver?
06:54 ogzy ArminderS: we want to create our compute nodes on xen, yes all compute nodes on xen
06:54 ArminderS and any specific reasons?
06:56 ogzy ArminderS: xen has some already capabilties and we want to use them
06:56 ogzy ArminderS: like its desktop client
06:57 ArminderS i wouldn't care much about it if i'm using openstack for orchestrating the cloud
06:57 ArminderS you may end up doing something from xenclient & messing up your stack
06:57 ogzy ArminderS: you may be right, just searching whether it is possible and it deserves to be done
06:58 ogzy ArminderS: rackspace says it is compatible with xen, for ex
06:58 ArminderS if you intend to run xenserver only, you may wish to look for specific puppet module
06:59 ArminderS and since rackspace is deploying on xenserver
06:59 ArminderS you may find one in there library
07:06 ArminderS try this -> https://github.com/stackforge/puppet-nova
07:06 ArminderS it supports both libvirt & xenserver
07:06 ArminderS drivers
07:07 ArminderS https://github.com/stackforge/puppet-nova/blob/master/manifests/compute/xenserver.pp
07:15 ogzy ArminderS: ok checking thank you
07:18 ogzy ArminderS: and i can create ubuntu based installation iso via usb with this right? http://docs.mirantis.com/fuel-dev/develop/env.html
07:18 ArminderS you need ubuntu to create iso
07:18 ArminderS not sure you will end up with ubuntu as base os for fuel
07:19 ArminderS i have never been through iso creation
07:20 ogzy ArminderS: ok
07:40 mrasskazov joined #fuel
07:52 teran joined #fuel
08:20 iscsi moin
08:21 iscsi how to properly shutdown the openstack environment?
08:50 e0ne joined #fuel
08:58 ArminderS i think first compute nodes and then controllers with primary one going last
08:58 ArminderS and the primary one shall be the first to boot up
09:00 ArminderS wow, one of my nodes just dissappeared from fuel gui
09:00 ArminderS and not getting listed in fuel node -l too
09:00 ArminderS i wonder where it went
09:00 ArminderS i can ping it, but can't ssh it
09:00 mattymo it might be kernel panicked or its networking broke
09:01 mattymo if it doesn't check in for several minutes, it's marked offline
09:01 ArminderS can't see it in cobbler as well
09:01 ArminderS why the heck its still pinging
09:10 e0ne_ joined #fuel
09:48 e0ne joined #fuel
09:50 e0ne joined #fuel
10:59 AndreyGrebenniko hello there people
10:59 AndreyGrebenniko does anyone know how to implement quotes for glance?
10:59 AndreyGrebenniko does anyone know how to implement *quotas for glance?
11:02 e0ne joined #fuel
11:28 e0ne joined #fuel
11:54 mihgen joined #fuel
11:56 e0ne joined #fuel
11:58 e0ne_ joined #fuel
12:37 MiroslavAnashkin joined #fuel
12:39 aabashkin joined #fuel
12:41 MiroslavAnashkin joined #fuel
12:43 MiroslavAnashkin joined #fuel
12:49 MiroslavAnashkin joined #fuel
13:12 vk joined #fuel
13:12 vk_ joined #fuel
13:12 b-zone joined #fuel
15:24 rmoe joined #fuel
15:47 e0ne joined #fuel
15:55 ogzy while the compute node installaion i got error, when i check, it lost internet connetcion
15:58 ogzy and i can see it has many iptables installed
15:58 ogzy i just have two interfaces on the machine that will be installed as compute node
15:59 ogzy do i need to have three interfaces for compute node installation?
16:07 dhblaz joined #fuel
16:10 dhblaz crm status reports that quantum-l3-agent is down:
16:10 dhblaz http://paste.openstack.org/show/IepfGvHjpIZayLLqICJN/
16:11 dhblaz Can anyone report troubleshooting steps?
16:11 dhblaz crm resource cleanup p_quantum-l3-agent
16:11 dhblaz oops
16:11 dhblaz rather I meant to say I can normally resolve the error by issuing the command: crm resource cleanup p_quantum-l3-agent
16:12 dhblaz But that doesn't help me get to the root of the problem and keep this from happening in the future.
16:16 dhblaz It looks like the quantum-l3-agent is running though:
16:16 dhblaz [root@node-92 log]# ps ax | fgrep quantum-l3
16:16 dhblaz 21257 ?        S      0:00 python /usr/bin/quantum-l3-agent --config-file=/etc/quantum/quantum.conf --config-file=/etc/quantum/l3_agent.ini
16:17 e0ne joined #fuel
16:20 mrasskazov joined #fuel
16:39 IlyaE joined #fuel
16:42 xarses joined #fuel
16:53 SteAle joined #fuel
17:04 dhblaz xarses: Is there a troubleshooting guide for the quantum-l3-agent?
17:11 xarses dhblaz: for the agent its self? or its function in crm?
17:12 dhblaz Any/All/Both?
17:13 dhblaz Here is my debugging output so far:
17:13 dhblaz http://paste.openstack.org/show/i7UotJ0fH9aiYxHkTF2P/
17:17 rmoe joined #fuel
17:19 dhblaz Interesting:
17:19 dhblaz [root@node-92 log]# crm resource status p_quantum-l3-agent
17:19 dhblaz resource p_quantum-l3-agent is NOT running
17:22 ArminderS joined #fuel
17:25 dhblaz After doing this I don't see the l3-agent in crm status after doing this:
17:25 dhblaz crm resource cleanup p_quantum-l3-agent; crm resource stop p_quantum-l3-agent; crm resource start p_quantum-l3-agent
17:34 e0ne joined #fuel
17:35 dhblaz This command seems to hang indefinitely: crm resource restart p_quantum-l3-agent
17:35 dhblaz [root@node-92 log]# crm resource restart p_quantum-l3-agent
17:35 dhblaz INFO: ordering p_quantum-l3-agent to stop
17:35 dhblaz waiting for stop to finish ..................................................
17:38 e0ne joined #fuel
17:38 dhblaz Eventually:
17:38 dhblaz ................... done
17:38 dhblaz INFO: ordering p_quantum-l3-agent to start
17:38 dhblaz But!
17:38 dhblaz [root@node-92 log]# crm resource status p_quantum-l3-agent
17:38 dhblaz resource p_quantum-l3-agent is NOT running
17:40 albionandrew joined #fuel
17:42 Arminder- joined #fuel
17:55 e0ne joined #fuel
17:59 Arminder joined #fuel
18:00 angdraug joined #fuel
18:20 Arminder joined #fuel
18:31 Arminder joined #fuel
19:06 e0ne joined #fuel
19:11 xarses joined #fuel
19:44 teran joined #fuel
19:45 dhblaz joined #fuel
20:01 dhblaz It seems pretty quiet in here.  Is something up?
20:02 xarses nope
20:03 xarses as to your crm / quantum-l3-agent, ts guides, i don't have anything for you
20:04 dhblaz Any pointers? Or any word on the maintenance release?
20:05 xarses nothing useful for either.
20:07 jsergent joined #fuel
20:13 mrasskazov joined #fuel
20:27 Bomfunk joined #fuel
21:02 mrasskazov joined #fuel
21:04 mrasskazov joined #fuel
21:13 mrasskazov joined #fuel
21:16 mrasskazov joined #fuel
21:25 mrasskazov joined #fuel
21:26 jsergent joined #fuel
21:34 e0ne joined #fuel
21:34 mrasskazov joined #fuel
21:37 IlyaE joined #fuel
21:45 dhblaz Any ideas why crm things l3_agent is down, but it is in the process table?  Or why those ports show down?
21:46 Bomfunk joined #fuel
22:05 jsergent joined #fuel
22:07 IlyaE joined #fuel
22:13 xarses the ports should show up if its working correctly. I don't know how CRM keeps tabs on it, you'd have to check what the health-check rules are
22:13 xarses dhblaz: it is sometimes, though not common to need to admin up and then down the port to get it to clear some weird state
22:14 dhblaz I didn't see a command to do this.
22:15 teran joined #fuel
22:17 xarses i cant remember if it was a nova or neutron command
22:17 xarses probably neutron
22:20 xarses see https://ask.openstack.org/en/question/1885/why-is-the-router-interface-down-and-how-do-i-bring-it-up/
22:20 xarses replace quantum with neutron
22:24 e0ne_ joined #fuel
22:29 teran_ joined #fuel
22:30 teran_ joined #fuel
22:31 teran joined #fuel
22:31 teran joined #fuel
22:35 dhblaz Thanks
22:36 teran joined #fuel
22:37 teran joined #fuel
22:41 xarses any use?
22:45 dhblaz I'm still fighting with crm
22:45 dhblaz p_quantum-l3-agent_start_0 (node=node-92.mumms.com, call=-1, rc=1, status=Timed Out): unknown error
22:46 dhblaz We still use quantum commands in 3.2
22:48 dhblaz The solution offered in that article didn't work for me (or the person that suggested the answer)
22:49 dhblaz Oh, wait - not true.  Of the 4 ports that were DOWN one of them is now ACTIVE, the one associated with network:router_gateway.
23:32 e0ne joined #fuel

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