Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2015-07-06

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

All times shown according to UTC.

Time Nick Message
01:46 Longgeek_ joined #fuel
02:08 youellet_ joined #fuel
02:47 Longgeek joined #fuel
02:51 hakimo joined #fuel
03:08 Samos123 network verification is failing with following error: http://paste.openstack.org/show/347116/ it says an error in mcollective
03:08 Samos123 im running controllers in kvm vms, could that be the problem?
03:09 Samos123 i have a feeling promisc mode may be blocked?
03:16 Longgeek_ joined #fuel
03:31 Samos123 only the nodes which are run as KVM VM fail with Reason: Execution expired, i already manually set my bridges to promisc on  at the host level
04:45 xarses joined #fuel
05:41 Samos123 seems I was hitting this bug: https://bugs.launchpad.net/fuel/+bug/1466020 killed stale python net_probe.py scripts manually and rerunning
05:59 Samos123 still failing, anybody other suggestions?
06:23 dancn joined #fuel
06:29 Samos123 i rebooted all bootstrapped nodes and it's working now
06:30 Samos123 verification succeeded and passed
06:59 stamak joined #fuel
07:16 saibarspeis joined #fuel
07:35 devvesa joined #fuel
07:36 samuelBartel joined #fuel
08:06 hyperbaba joined #fuel
08:10 HeOS joined #fuel
08:17 subscope joined #fuel
08:37 baboune joined #fuel
08:38 baboune Mirantis 6.1, compute node, Neutron VLan ubuntu, there are no /etc/init.d/ scripts for openstack services.  How to stop/start openstack-ceilometer-compute for example?
08:42 thumpba joined #fuel
08:56 evg Samos123: It was because the nodes are remained from your previous deployment. New mco password was generated and it's transfered during pxe boot.
09:16 subscope joined #fuel
09:36 strictlyb joined #fuel
10:21 saibarspeis joined #fuel
10:42 Longgeek joined #fuel
10:45 neophy joined #fuel
11:05 evg baboune: please use initctl utility. initctl start <service>. The config files are in /etc/init/
11:17 e0ne joined #fuel
11:36 subscope joined #fuel
11:56 e0ne joined #fuel
13:21 Samos123 evg: weird I remember those were clean nodes not from a previous deployment, but not 100% sure
13:23 Samos123 whats the recommended way of making changes to configuration after deployment. I would like to utlize the existing puppet orchestration and configuration framework for an already deployed environment
13:23 Samos123 such that new nodes added are configured automatically but also already deployed nodes should do a puppet apply again
13:24 Samos123 i remember in old releases there was a specific documentation part about manually adding puppet modules but couldn't find it in 6.1 docs
13:29 rbrooker joined #fuel
13:43 jobewan joined #fuel
13:49 championofcyrodi a few months ago someone gave me a link to the community downloads for fuel, however I can't seem to find it anymore and keep getting linked to the commerical version.
13:50 championofcyrodi anyone still have that link?
13:50 championofcyrodi looking to blow away my fuel server and redeploy w/ 6.1
13:54 xdeller joined #fuel
14:01 championofcyrodi nvm, found it: https://www.fuel-infra.org/
14:19 claflico joined #fuel
14:32 dklepikov joined #fuel
14:43 Longgeek joined #fuel
14:46 rmoe joined #fuel
15:26 justif joined #fuel
15:32 hammondr joined #fuel
15:32 hammondr good morning.
15:33 hammondr under fuel 6.1, a network and subnet are created in the admin tenant.  when creating a new tenant router, should the tenant member be able to add an interface to the ext04_net network without admin assistance?
15:42 Billias for Ceph rdb
15:42 Billias I also need 2 servers to act as metastore?
15:42 Billias Ceph RBD*
16:00 Longgeek joined #fuel
16:30 xarses joined #fuel
16:33 kaliya joined #fuel
16:50 mrasskaz- joined #fuel
16:55 darkhuy joined #fuel
16:57 darkhuy joined #fuel
17:01 Longgeek joined #fuel
17:04 barakme joined #fuel
17:05 barakme hello. Anyone here?
17:07 barakme So if I have questions about using mirantis express, specifically with the API, is this the place?
17:12 xarses Billias: Ceph is a distributed object store, by default It has replication factor of 2, so you need two nodes to meet it. You can change it from the setting page back to 1, but it will not have host redundancy
17:13 xarses barakme: around provisioning your MOX cloud, or with the Openstack API's?
17:14 xarses this channel is around using the fuel deployment engine. Which does sit inside your MOX cloud
17:14 xarses This is also an ad-hoc discussion channel, If you are a MOX customer you should be able to contact Mirantis Support for SLA based responses
17:15 barakme ok. Thanks.
17:18 Billias xarses: Do I need also meta system? (I saw that cpu can be low, and memory is calculated per GB)
17:19 Billias I need it only for block storage, I will have the VMs locally
17:19 championofcyrodi joined #fuel
17:23 xarses We don't install or support cephfs (metadata/mds) It's not needed for Open Stack, you have the option to configure ceph RBD for glance, cinder, and nova-ephemeral. and RadosGW for Swift/S3 API
17:29 championofcyrodi 6.1 fresh deployment failed, http://paste.openstack.org/show/349181/  attempting to troubleshoot.
17:30 championofcyrodi was near the end of the ubuntu install and about 54-56% completed overall.
17:32 championofcyrodi curl 10.20.0.2:8080/2014.2.2-6.1[/ubuntu/auxiliary] results in HTTP 301 (moved permanently)
17:34 championofcyrodi looks like almost all the endpoints fail to work unless you append a 'forward-slash' ( / ) to the end of the uri/
17:35 championofcyrodi looks like all the 'target' images are centos and it's looking for ubuntu... i guess i'll double-check the 'fuel-createmirror'
17:39 jaycee joined #fuel
17:50 mwhahaha championofcyrodi: that error is usually from the image building process for the image based deploy, if you look in the astute logs you can usually see what exactly is the issue
17:50 mwhahaha it's usually either a missing dependency or it took to long to download all the items
17:51 mwhahaha there should also be an image building log i believe as well
17:54 e0ne joined #fuel
17:59 championofcyrodi so the astute log has this every second from 17:20:55  - 17:27:46:
17:59 championofcyrodi debug: [535] Data received by ProvisiningProxyReporter to report it up: {"nodes"=>[{"uid"=>"1", "progress"=>80, "status"=>"provisioning"}, {"uid"=>"2", "progress"=>80, "status"=>"provisioning"}, {"uid"=>"3", "progress"=>80, "status"=>"provisioning"}, {"uid"=>"4", "progress"=>80, "status"=>"provisioning"}, {"uid"=>"5", "progress"=>80, "status"=>"provisioning"}]}
17:59 championofcyrodi , followed by the failure message pasted above.
17:59 championofcyrodi as if it got to "80%" and just timed out.
17:59 mwhahaha check the image building log /var/log/fuel-agent-env-1.log
18:03 championofcyrodi seems to be okay running the apt process of installing packages, but then seems to hit some issues w/ grub config: http://paste.openstack.org/show/349182/
18:32 claflico joined #fuel
18:54 mwhahaha championofcyrodi: no those error lines are normal, i'm not sure what happens during the 'Containerizing temporary image file' but that seems to be where it was just sitting for a while. on my laptop that only takes like 3 minutes
18:54 mwhahaha when did you start the process? perhaps it just took too long
19:02 Longgeek joined #fuel
19:06 championofcyrodi started the deploy about 12:30ish and quickly got to 47%.. went to lunch, came back ~40 minutes later, and it was just slowly churning at 53%, then 54... then around 56% it failed with the earlier exception.
19:07 championofcyrodi i just deleted the environment and tried again, since my environment config was already created before i ran 'fuel-createmirror' for the first time...
19:07 hammondr left #fuel
19:07 championofcyrodi but now for whatever reason, none of the vlan tags are working for only 1 of the several interfaces on my controller, so i can't pass the network verify, even though the config is the same as last time when it did pass.
19:09 championofcyrodi basically i have 4x1Gbps & 1x10Gbps NICs connected to the single controller... 3x1Gbps bonded for storage network, the 10Gbps for public, and the single 1Gbps for PXE,Management,Private
19:10 championofcyrodi but the single 1Gbps for pxe,management and private is not getting the VLAN response for ANY of the VLANs configured... but all other interfaces are.
19:10 mwhahaha since you deleted the environment rather than reset, are you sure you reconfigured the network for the new environment correctly?
19:10 championofcyrodi (the other interfaces on the compute nodes are using one shared NIC w/ VLANs for everything except storage, which has a dedicated NIC
19:11 championofcyrodi yea, printed the config
19:11 championofcyrodi and double checked...
19:12 championofcyrodi fuel just hates me
19:16 championofcyrodi i have 2 x HP Proliant D380 with the integrated NIC PXE booting... The one configured as a compute node is using eth6 w/ VLANs and not getting the error.  The other is my controller, using eth6 w/ VLANs and failing...
19:17 championofcyrodi guess i'll try a different port on the switch, again...
19:25 championofcyrodi okay... just moved the private and management VLANs off the PXE ethernet adapter and on to the bond0 w/ the storage, and it worked.
19:26 e0ne joined #fuel
19:27 bluun joined #fuel
19:27 championofcyrodi seems to be taking a while on 2015-07-06 19:25:14.407 22815 DEBUG fuel_agent.utils.utils [-] Trying to execute command: debootstrap --verbose --no-check-gpg --arch=amd64 trusty /tmp/tmpBrEEZ1.fuel-agent-image http://10.20.0.2:8080/ubuntu-part
19:32 championofcyrodi progress => 12
19:33 mwhahaha sounds like you're running into io issues like it's just taking too long
19:34 mwhahaha there's a bug for that somewhere
19:35 championofcyrodi progress => 16 ; is there a workaround for increasing the timeout via ENV Var or something?
19:36 championofcyrodi perhaps i can just configure everything w/ the UI, and the deploy via CLI
19:36 mwhahaha https://bugs.launchpad.net/fuel/+bug/1453039
19:36 mwhahaha you could pull the tasks and update the timeout and upload the tasks as a work around to get by
19:37 championofcyrodi so the manual fix is to 'redeploy' after debootstrap is finished...
19:37 mwhahaha yea if it ever finishes
19:38 championofcyrodi but i'm not sure how to monitor the debootstrap process outside of the astute.log
19:38 championofcyrodi oh... the info in the bug made me think it *would* finish, just not in the 3600 second limit.
19:38 mwhahaha it might, but i'm not sure if it keeps running once the 3600 timeout hits
19:39 championofcyrodi yea, seemed like mine stopped at 80ish
19:39 mwhahaha eh i don't think those numbers mean anything
19:39 mwhahaha i think for that part it's just incrementing to show that it's still working
19:39 championofcyrodi i guess i could swap the fuel server out with a more powerful desktop
19:40 mwhahaha are you directly on the box or are you doing some virtualization?
19:40 championofcyrodi fuel is a dedicated box
19:41 mwhahaha and you created a mirror?
19:41 championofcyrodi holy moly... i think i see the problem!
19:41 championofcyrodi fuel machine only has 4GB of RAM... and is swapping like hell
19:41 championofcyrodi maybe not 'like hell' but it is using swap
19:42 mwhahaha 4g should be enough
19:43 mwhahaha in my dev env, i think my fuel master is only 2.5g
19:43 mwhahaha of course i've got an ssd
19:43 championofcyrodi showing ~ 100MB free with ~180MB swapping.
19:43 mwhahaha eh that's not bad
19:43 angdraug joined #fuel
19:43 championofcyrodi just ran 'swapoff -a'
19:44 championofcyrodi doesnt seem to be helping much though.
19:44 championofcyrodi seems like it's taking about 37 seconds for every 1% progress...
19:47 championofcyrodi http://manpages.ubuntu.com/manpages/natty/man1/eatmydata.1.html
19:47 championofcyrodi perhaps that would help...
20:21 HeOS joined #fuel
20:21 darkhuy do you guys typically run fuel off ubuntu or centos OS?
20:24 mwhahaha i think most people deploy ubuntu environments
20:24 mwhahaha but i like centos because it deploys faster since we ship the initial image on the iso
20:24 Longgeek joined #fuel
20:25 darkhuy my original setup was centos6.5 + fuel 6.0, then centos7 + 6.1
20:25 darkhuy since moving to ubuntu I've yet to be able to pass network verification
20:25 darkhuy don't know why
20:26 mwhahaha what do you mean centos7 + 6.1
20:26 darkhuy centos 7 with fuel 6.1 running inside virt-manager
20:26 darkhuy with bridged network interfaces
20:26 mwhahaha ah
20:27 mwhahaha ubuntu should work fine, that's what most people dev on either ubuntu or debian
20:28 darkhuy network setup is such a PITA
20:28 mwhahaha https://docs.fuel-infra.org/fuel-dev/devops.html
20:28 mwhahaha we've got scripts to managing environments -^
20:28 stamak joined #fuel
20:28 mwhahaha then you can use that and just bridge your fuel node to your actual admin interface
20:30 darkhuy interesting
20:30 championofcyrodi i just switched out the fuel server w/ a core i7 desktop, SATA II disk, 16GB RAM.
20:30 championofcyrodi installing and building the docker images now...
20:31 darkhuy motherfk*
20:31 championofcyrodi we went w/ centos for the first few deployments, decided to try ubuntu since it has a newer kernel
20:31 darkhuy when i reverted to 6.0 from 6.1 i forgot to hardboot the machines back to 6.0 bootstrap
20:31 darkhuy seems like ipmi resetting all the compute nodes back to the 6.0 bootstrap fixed it
21:13 jeh joined #fuel
21:47 Longgeek joined #fuel
22:59 rmoe joined #fuel
23:16 dmellado_ joined #fuel
23:32 Longgeek joined #fuel
23:57 CTWill joined #fuel

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