Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2015-08-05

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

All times shown according to UTC.

Time Nick Message
01:12 tzn joined #fuel
01:48 ilbot3 joined #fuel
01:48 Topic for #fuel is now Fuel 6.1 (Juno) https://software.mirantis.com | Paste here http://paste.openstack.org/ | IRC logs http://irclog.perlgeek.de/fuel/
02:03 ub joined #fuel
02:42 tzn joined #fuel
02:45 sbfox joined #fuel
02:52 hakimo_ joined #fuel
03:09 karume joined #fuel
03:16 xarses joined #fuel
03:45 tzn joined #fuel
04:19 monester joined #fuel
04:25 neophy joined #fuel
04:38 lucasPR joined #fuel
04:40 fedexo joined #fuel
04:46 tzn joined #fuel
05:13 ximepa joined #fuel
05:52 tzn joined #fuel
05:59 karume joined #fuel
06:10 hezhiqiang joined #fuel
06:37 mkwiek joined #fuel
06:41 ub joined #fuel
06:58 tzn joined #fuel
07:19 saibarspeis joined #fuel
07:19 Samos123 hi guys just wondering whether there is a roadmap for enabling Ansible based fuel plugins?
07:20 monester joined #fuel
07:23 youellet__ joined #fuel
07:25 ddmitriev1 joined #fuel
07:41 karume joined #fuel
07:47 e0ne joined #fuel
07:49 martineg_ would be convenient, but it works to bootstrap Ansible in a shell task and run playbooks packaged in the plugin?
07:50 martineg_ so far I've run Ansible from the fuel node post-deploy.
07:56 e0ne joined #fuel
07:59 tzn joined #fuel
08:19 karume joined #fuel
08:27 mquin joined #fuel
08:44 kutija_ joined #fuel
08:51 HeOS joined #fuel
09:04 tzn joined #fuel
09:22 tzn joined #fuel
09:31 e0ne joined #fuel
10:24 ximepa joined #fuel
10:30 e0ne joined #fuel
10:40 RedShift_ joined #fuel
10:40 RedShift_ joined #fuel
11:04 hezhiqiang joined #fuel
11:06 kutija_ joined #fuel
11:07 e0ne joined #fuel
11:11 sergmelikyan joined #fuel
11:11 sbog joined #fuel
11:24 jaypipes joined #fuel
11:28 RedShift joined #fuel
11:36 mquin joined #fuel
11:47 Kesali joined #fuel
11:50 Kesali left #fuel
12:06 e0ne joined #fuel
12:34 warpc__ joined #fuel
12:39 RedShift joined #fuel
12:53 karume joined #fuel
13:01 thumpba joined #fuel
13:11 ximepa left #fuel
13:32 RedShift joined #fuel
13:39 jdandrea joined #fuel
13:41 Rodrigo_BR joined #fuel
13:42 Rodrigo_BR hello
13:42 Rodrigo_BR Can i add new compute node in existing environment  using fuel 6.1 ?
13:46 sc68cal_ joined #fuel
13:53 kat_pimenova joined #fuel
14:07 thumpba joined #fuel
14:18 mkwiek joined #fuel
14:19 mudblur joined #fuel
14:27 azemlyanov joined #fuel
14:32 claflico joined #fuel
14:33 blahRus joined #fuel
14:37 RedShift joined #fuel
15:04 bdudko joined #fuel
15:05 aglarendil_ joined #fuel
15:06 gongysh_ joined #fuel
15:08 bildz So i attempted to delete an openstack instance because it was hanging on the install and it just shows that all the servers are pending deletion.  How do I get this to progress?
15:14 jobewan joined #fuel
15:17 bildz should i just use the dockerctl to destroy/rebuild?
15:35 RedShift joined #fuel
15:38 neophy joined #fuel
15:40 bdudko joined #fuel
15:40 aglarendil_ joined #fuel
15:41 gongysh_ joined #fuel
15:54 neophy joined #fuel
16:01 tzn joined #fuel
16:31 RedShift joined #fuel
16:43 kpimenova joined #fuel
16:58 DaveJ__ joined #fuel
16:58 xarses joined #fuel
17:04 mkwiek joined #fuel
17:06 tzn joined #fuel
17:18 skylerberg joined #fuel
17:25 neophy joined #fuel
17:28 jaypipes joined #fuel
17:31 mkwiek joined #fuel
17:53 e0ne joined #fuel
17:57 karume joined #fuel
18:00 claflico joined #fuel
18:11 bildz anyone back yet?
18:20 sergmelikyan joined #fuel
18:26 monester joined #fuel
18:27 claflico joined #fuel
18:34 bildz mwhahaha and xarses must be on vacation :)
18:35 xarses mwhahaha: is
18:35 xarses Rodrigo_BR: yes you can
18:36 bildz yeah it's been silent so I figured as much
18:36 xarses bildz: do yo mean that you've successfully deployed with fuel, and while spawning an instance in openstack it's stuck?
18:36 bildz at some point, i'll be able to assist people
18:36 bildz i setup my stack, added controllers/compute, verified, and pushed deploy
18:37 bildz it sat at installing centos for over a day, so then I decided to try removing a node, to which is says pending deletion
18:37 bildz then I tried to remove the instanance, and now it says "removing"
18:37 bildz does it ever complete the steps?
18:37 xarses Samos123: you can run ansible if you installed it in a prior task. The task executor will run whatever can be run in bash shell
18:37 bildz i've stopped/started dockerctl
18:38 bildz trying to figure out the fuel cli commands, but it's just showing "remove"
18:39 xarses ya, rabbit might have crashed
18:39 xarses you would be able to see this if none of the nodes where in a clean discovered state
18:40 RedShift joined #fuel
18:40 xarses you can `fuel task --task-id <> --delete --force`
18:40 xarses and then remove the node again
18:40 xarses or just delete the node from the db and restart it by hand
18:41 xarses `fuel node --env <> --delete-from-db --node <>
18:41 xarses syntax may not be 100%, check the commandline help
18:41 bildz 41 | error   | cluster_deletion        | 1       | 100      | 783be224-e2b6-4766-a7a5-799c54809832
18:43 bildz that worked let me take a look quick
18:44 bildz still looks like it's showing up in the web UI
18:46 bildz xarses: thanks man.  looks like i'm finally getting it to come back
18:47 bildz whew
18:47 deckardkain I'm trying to install openstack inside of vmware. I'm getting this error message. http://paste.openstack.org/show/409211/. The funny thing is i had this working yesterday. But had to redeploy because i fat fingred a CIDR for the fixed addresses.
18:47 thansen joined #fuel
18:48 deckardkain when i run the command on node-22 i get this http://paste.openstack.org/show/409212/
18:49 deckardkain not fully understanding how   Error in store configuration. Adding images to store is disabled.<br /><br />
18:52 monester joined #fuel
18:52 xarses deckardkain: did you enable the vcenter datastore on the settings page, and the configure it on the vmware tab?
18:53 thansen my understanding is that when ceph is enabled for nova that the machine will automatically use that...however when looking at the node it shows /var/lib/nova as a mounted lvm device?
18:54 neophy joined #fuel
18:55 xarses thansen: correct, fuel will still allocate a miminum of 5gb /var/lib/nova as a lvm volume. Its needed for the instance metadata, and is used sometimes as temporary storage by nova for some actions
18:55 thansen xarses, ok, how can I confirm my instances are also in the ceph storage?
18:55 xarses I know it used to be where snapshots where stored from ceph before being uploaded to glance due it not doing them inside ceph
18:56 xarses from a node with an admin or mon keyring (controllers) you can use the rbd or rados (I don't remember which one is correct at the moment to list the contents of the compute pool, you should find an object maching the uuid of the nova instance
18:56 thansen I've got some minimal experience with gluster but nothing with ceph so forgive my ignorance if I'm out there
18:57 xarses also, it will basicly boot the image instantly
18:58 thansen so an individual node will *additionally* need storage outside of ceph?
18:59 thansen lemme rephrase that, a given compute node will only be able to handle ephemeral storage up to the amount mounted at /var/lib/nova?
18:59 xarses thansen: in a perfect world its quite minimal, it would be limited to xml metadata about the node
18:59 thansen I see, is 5GB sufficient then?
19:00 thansen I see that now../var/lib/nova only has console.log and libvirt.xml in there on a per-instance basis
19:00 thansen so 5GB should do the trick for sure, any benefit to more?
19:00 xarses like I said, snapshots used to wrongly end up there
19:01 xarses so test making a snapshot, I'm not sure it was fixed or not
19:01 thansen ok, this is a new deploy from 6.1
19:01 thansen ok, I'll do that
19:01 thansen where would it appear if it's *wrong*?
19:01 xarses with the other instance data
19:01 thansen ok
19:02 thansen lemme try real quick
19:03 deckardkain @xarses: If you mean the "VMWare vCenter/ESXi datastore for images(Glance)" yes. I also configured the Glance section of the VMWare tab. Pretty much copy pasted everything.
19:03 ub joined #fuel
19:06 thansen xarses, looks good, however the console.log changed ownership on the fs when I made the snapshot
19:06 thansen was nova:root and after it's root:root
19:06 thansen not sure if that makes anything else blow up but just an FYI
19:07 xarses thansen: apparently, part of it is still open
19:07 xarses https://review.openstack.org/#/c/188244/
19:10 thansen xarses, ok, so as I understand so far, the size of that storage would limit the size of a created snapshot
19:10 thansen just during creation time
19:10 xarses sounds that way to me aswell
19:10 thansen ie: snapshot gets written to /var/lib/nova then gets moved
19:11 thansen ok, so you'd want the size to be essentially <max image size> * <max concurrent snapshot creations>
19:13 angdraug joined #fuel
19:20 xarses probably
19:21 xarses per node
19:25 ub joined #fuel
19:47 mkwiek joined #fuel
19:56 thansen xarses, yeah, understood
19:57 thansen xarses, what's a good amount of 'base' storage speaking generally?
19:58 sergmelikyan joined #fuel
20:09 monester joined #fuel
20:11 xarses for the computes
20:12 xarses I'd probably keep around 500gb to a 1tb locally
20:14 thansen xarses, what fills that?
20:14 xarses hopefully nothing
20:14 mkwiek joined #fuel
20:15 thansen ok
20:16 xarses deckardkain: It seems like glance is having an issue with the config, can you find the portion of the glance-api.log that corosponds with that request, I'd guess some traceback was raised
20:18 HeOS joined #fuel
20:18 thansen does a compute node also acting as a ceph storage node act smart enough to do read/writes 'locally'?
20:19 JoeStack joined #fuel
20:19 xarses no, the placement algorythim doesn't allow for all of the data to be placed that way so It's likely that some of the placement groups will not be on the OSD devices on the local node
20:20 thansen I see
20:21 thansen in that case using a local ceph journal helps by 'cache'ing it locally then?
20:24 thansen or is a journal only fronting the actual storage?
20:34 monester joined #fuel
20:43 wayneeseguin joined #fuel
20:44 deckardkain @xarses: checking now
20:45 deckardkain @xarses: nailed it
20:46 sergmelikyan joined #fuel
20:47 deckardkain @xarses: so for the record. Hopefully this helps someone. If using fuel to deploy an openstack environment on esxi to be used as a hypervisor, you need to *not* use the same datastore for the instance AND for glance image storage. Also, if redeploying that environment onto the same datastore, you need to clear out any *previous* glance images.
20:48 deckardkain i just redeployed using two different datastores and cleared out my previous datastore that held old glance images.
20:52 deckardkain and it was a success
21:00 JoeStack <deckardkain> I use NFS as datastore. Don't think that I have a problem in deploying a 2nd openstack instance
21:06 JoeStack may found another issue regarding fuel 6.1 install. when using centos the deployment fails, but not when using ubuntu with same network settings
21:15 bildz xarses: "Too many nodes failed to provision"
21:16 bildz what's weird is that the network validation came clean and my controllers successfully installed
21:16 bildz just the compute nodes all failed
21:29 karume joined #fuel
21:31 deckardkain JoeStack: are you using VCenter to deploy openstack nodes and to deploy instances on?
21:39 JoeStack <deckardkain>: I use vCenter/vSphere as my baremetal environment but no vCenter integration within FUEL/openstack.
21:47 bildz im using openstack as baremetal
21:49 monester joined #fuel
22:03 RedShift joined #fuel
22:33 angdraug joined #fuel
23:04 NERvOus joined #fuel
23:18 karume joined #fuel

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