Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2014-10-16

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

All times shown according to UTC.

Time Nick Message
00:11 xarses I believe its in the juno code base, but it's still not configured by fuel
00:29 rmoe joined #fuel
00:42 Rajbir joined #fuel
01:16 alex_didenko joined #fuel
01:19 Longgeek joined #fuel
01:27 xarses joined #fuel
01:48 ilbot3 joined #fuel
01:48 Topic for #fuel is now Fuel 5.1 for Openstack: https://wiki.openstack.org/wiki/Fuel | Paste here http://paste.openstack.org/ | IRC logs http://irclog.perlgeek.de/fuel/
02:04 Longgeek joined #fuel
02:04 jpf_ joined #fuel
02:12 rmoe joined #fuel
03:18 dhblaz joined #fuel
03:22 jpf joined #fuel
03:43 dhblaz joined #fuel
04:10 ArminderS joined #fuel
04:10 fandi joined #fuel
04:16 alex_didenko joined #fuel
04:20 pasquier-s joined #fuel
04:40 anand_ts joined #fuel
04:53 jaypipes joined #fuel
05:22 ArminderS joined #fuel
05:35 ArminderS joined #fuel
05:35 ArminderS joined #fuel
05:41 ArminderS joined #fuel
05:46 anand_ts joined #fuel
05:46 anand_ts hello all, why I can't ping to gateway or the laptop connected to the switch  from Fuel master node. But I can ping discovered nodes ? There are two things in Fuel menu. 1. Pool of static nodes and 2. pool of node discovery? what is the difference between 1 & 2 ?
05:47 ArminderS joined #fuel
05:48 ArminderS joined #fuel
05:49 ArminderS joined #fuel
06:03 Longgeek joined #fuel
06:14 syt joined #fuel
06:28 adanin joined #fuel
06:31 pal_bth joined #fuel
06:46 pal_bth joined #fuel
06:49 akupko joined #fuel
06:51 anand_ts hello all, is it normal that Fuel master node cannot ping the administrative gateway. for eg, my fuel master node has 10.54.7.20 IP, but it can't ping 10.54.7.1 gateway and laptop connected at 10.54.7.30. but it can ping the nodes discovered. 10.54.7.145 & 10.54.7.205. What could be the reason?
06:52 adanin joined #fuel
06:55 xarses joined #fuel
07:02 akupko joined #fuel
07:05 e0ne joined #fuel
07:11 e0ne joined #fuel
07:24 dancn joined #fuel
07:27 baboune joined #fuel
07:28 kaliya anand_ts: hi. It depends on your network settings. If configured properly, the master should be able to reach the Internet
07:28 harybahh joined #fuel
07:28 e0ne joined #fuel
07:29 baboune hi, question about nova in Fuel 5.1: nova server-group-create --policy=anti-affinity hadoop_dn returns an error "invalid choice: 'server-group-create". "nova help" does not list server-group-xx options at all. But this is normally part of the nova CLI.  What am I doing wrong?
07:34 azemlyanov joined #fuel
07:35 alex_didenko joined #fuel
07:37 e0ne joined #fuel
07:37 baboune is this the reason for the pb: https://ask.openstack.org/en/question/30433/why-are-nova-server-group-apis-missing-in-rdo-icehouse-installation/ ?  Is that missing in the 5.1 release?
07:39 anand_ts kaliya: yes, node has internet access, I connected switch to firewall directly. I can ping all other IP's except the gateway
07:47 baboune according to this presentation from Mirantis http://cdn2.hubspot.net/hub/197500/file-695130377-pdf/pdf/Whats-New-In-Icehouse_Webcast.pdf server-group should be included as part of the nova cli.  Any fixes/ideas?
07:53 e0ne joined #fuel
07:53 kaliya baboune: a nova client just prior of icehouse release is included in fuel 5.1, for stability reasons. You can consider to install the latest client, but it's not tested
08:03 baboune kaliya: but it is "bug" in the existing 5.1 release right?
08:10 baboune will there be an upgrade to 5.1 containing the fix soon?
08:18 baboune hummm... If not in the CLI is it available in the REST API?
08:18 kaliya baboune: better, a feature not imported. I will query the devs
08:22 pasquier-s_ joined #fuel
08:24 pasquier-s__ joined #fuel
08:31 hyperbaba_ joined #fuel
08:35 HeOS joined #fuel
08:40 akupko joined #fuel
08:49 merdoc anand_ts: maybe there some restriction on gateway? firewall?
08:50 baboune kaliya: thx
08:51 e0ne joined #fuel
08:55 anand_ts merdoc: no restrictions as such. But I wonder why a laptop I connected in that network can't able to ping from master node. But from my laptop I can ping Fuel master node and gateway
08:56 merdoc anand_ts: looks like problem somewhere in routing
08:57 anand_ts merdoc: Okay checking this
09:31 adanin joined #fuel
09:38 teran joined #fuel
10:02 pasquier-s joined #fuel
10:04 teran joined #fuel
10:10 adanin joined #fuel
10:12 vtzan joined #fuel
10:27 alexz joined #fuel
10:33 saibarspeis joined #fuel
10:46 dkaigarodsev joined #fuel
10:53 teran_ joined #fuel
11:06 teran joined #fuel
11:35 pasquier-s joined #fuel
11:36 saibarspeis joined #fuel
11:44 teran joined #fuel
11:50 anand_ts joined #fuel
11:52 anand_ts hello all, error while installing compute node from dashboard. I got error deployment has failed, Check these nodes
11:57 teran joined #fuel
12:02 anand_ts_ joined #fuel
12:03 anand_ts_ Can we install 2 nodes (1 controller and 1 compute ) at a time from master node. I got error that compute node installation failed and controller ready
12:04 merdoc what logs says?
12:05 anand_ts_ merdoc: where can I find that? I mean from dashboard itself?
12:06 anand_ts_ merdoc: which logs to see. I can see puppet,syslog,anaconda logs. from where I can start troubleshooting
12:08 merdoc anand_ts_: I think you need too look at all of it.
12:15 harybahh joined #fuel
12:42 jpf joined #fuel
13:13 BillTheKat joined #fuel
13:26 sc-rm On which blueprint should I write a request for including the sysstat package in the repository, for debugging and tracking down performance problems on nodes?
13:26 kaliya sc-rm: just open a new one
13:27 sc-rm kaliya: here https://blueprints.launchpad.net/fuel/+addspec is good enough?
13:30 kaliya sc-rm: right place
13:43 BillTheKat joined #fuel
13:44 teran_ joined #fuel
13:46 tdubyk joined #fuel
13:54 akupko joined #fuel
13:55 hyperbaba_ Hello,
13:57 hyperbaba_ Couple of days ago we discused the problem with token expiration for glance service when creating snapshot (for more than 1 hour). The fix I get (new oslo packages and glance_api_version=2) is not working. Some of the errors are gone from the log but resulting snapshot stays in "Saving" state (but it's already finished on ceph storage)
14:01 mattgriffin joined #fuel
14:24 geekinutah joined #fuel
14:30 jpf joined #fuel
14:32 mpetason joined #fuel
14:33 BillTheKat joined #fuel
14:33 hyperbaba joined #fuel
14:48 thehybridtech joined #fuel
14:48 dhblaz joined #fuel
14:52 pasquier-s joined #fuel
14:52 xarses joined #fuel
14:56 adanin joined #fuel
14:56 nullme joined #fuel
14:57 nullme hello everyone, i was wondering if any has seen this issue during deployment or upgrade (/Stage[main]/Keystone::Roles::Admin/Keystone_user xxxx Could not evaluate: Execution of '/usr/bin/keystone --os-token 18mcvjf9 --os-endpoint http://127.0.0.1:35357/v2.0/ tenant-get ec1cc37e0f6f4c02b2d3e66adeb3484d' returned 1: No tenant with a name or ID of 'ec1cc37e0f6f4c02b2d3e66adeb3484d' exists.
14:58 MiroslavAnashkin nullme: Have deployment finished successfully after all?
14:59 nullme MiroslavAnashkin: no i get errors; Error occurred while running method 'deploy'. Inspect Astute logs for the details
15:00 championofcyrodi joined #fuel
15:03 MiroslavAnashkin nullme: Have this error happened in scope of the first controller deployment or you were adding one more controller?
15:05 nullme this happened on two occasions, new first controller, and upgrade of 5.0 to 5.1, im more focused on upgrade
15:05 blahRus joined #fuel
15:07 MiroslavAnashkin Has this error happened on Fuel master node?
15:08 MiroslavAnashkin Or do you upgrading already existing OpenStack nodes from 5.0 to 5.1?
15:09 nullme no i upgraded the fuel master node with no issues, then i did a new deployment and it crashed, on my other environment, i also did the fuel master node upgrade to 5.1 with no issues and so i did upgrade environment in experimental mode
15:13 MiroslavAnashkin Then, please generate diagnostic snapshot from master node with failed deploy and share it somewhere. Dropbox, Google drive - anywhere.
15:15 nullme will do, it takes a minute to generate it...
15:19 BillTheKat joined #fuel
15:19 ArminderS joined #fuel
15:20 jobewan joined #fuel
15:35 blahRus1 joined #fuel
15:48 angdraug joined #fuel
15:53 BillTheKat joined #fuel
16:01 kozhukalov joined #fuel
16:07 rmoe joined #fuel
16:14 HeOS joined #fuel
16:22 syt joined #fuel
16:30 nullme joined #fuel
16:50 adanin joined #fuel
16:58 Rajbir joined #fuel
16:59 ArminderS- joined #fuel
17:00 HeOS joined #fuel
17:09 pal_bth joined #fuel
17:58 syt joined #fuel
18:04 blahRus joined #fuel
18:10 kaliya_ joined #fuel
18:10 angdraug joined #fuel
18:17 emagana joined #fuel
18:22 syt joined #fuel
18:22 strictlyb joined #fuel
18:22 jseutter joined #fuel
18:31 xarses joined #fuel
18:41 nullme hi everyone... quick question, my diagnostic logs are very big, 3.4gb is there a way to make them small?
18:41 nullme MiroslavAnashkin: are you still here... i have a diagnostic link for you but its very big...
18:47 kaliya_ joined #fuel
18:52 syt joined #fuel
18:54 ilbot3 joined #fuel
18:54 Topic for #fuel is now Fuel 5.1 for Openstack: https://wiki.openstack.org/wiki/Fuel | Paste here http://paste.openstack.org/ | IRC logs http://irclog.perlgeek.de/fuel/
18:56 emagana joined #fuel
18:56 BillTheKat joined #fuel
18:56 nullme https://drive.google.com/file/d/0B7Wq17XKFYUmV1lKWDJ5SHV4R3c/view?usp=sharing
18:57 emagana joined #fuel
18:58 xarses joined #fuel
18:58 nullme anyone ever seen this error, (/Stage[main]/Glance::Keystone::Auth/Keystone_user[glance]) Could not evaluate: Execution of '/usr/bin/keystone --os-token 18mcvjf9 --os-endpoint http://127.0.0.1:35357/v2.0/ tenant-get ec1cc37e0f6f4c02b2d3e66adeb3484d' returned 1: No tenant with a name or ID of 'ec1cc37e0f6f4c02b2d3e66adeb3484d' exists.
18:59 ArminderS joined #fuel
19:02 boris-42 joined #fuel
19:05 dhblaz joined #fuel
19:14 teran joined #fuel
19:14 nullme anyone here?
19:20 BillTheKat joined #fuel
19:21 teran joined #fuel
19:23 MiroslavAnashkin Yes, still here
19:24 nullme Sorry i posted the link but its a big file...
19:24 nullme is that ok?
19:25 MiroslavAnashkin Yes, we have even 10G snapshots
19:26 MiroslavAnashkin Downloading.
19:27 MiroslavAnashkin I'll pass your snapshot to evgeniyl, he is in charge od upgrades
19:27 hyperbaba MiroslavAnashkin: Hi, regarding problem with snapshots on ceph. It's not working still after changing the glance api version to 2 in cinder.conf and updating the oslo packages you gave me. Now errors 32 are gone but still no closure of snapshot in glance. Stays in saving state regarding the time passed. Ceph did the job of saving it .
19:28 MiroslavAnashkin hyperbaba: What was the image format, raw or qcow or.. ?
19:28 hyperbaba MiroslavAnashkin: it was raw. Looks like it's still token expiration error
19:28 alex_didenko joined #fuel
19:28 hyperbaba MiroslavAnashkin: Or maybe I'm wrong
19:30 MiroslavAnashkin And what is your current snapshot format in /etc/nova/nova.conf? OpenStack still use curl to download snapshot then converts it to the target format locally, then uploads back to Glance
19:30 nullme MiroslavAnashkin: thank you for doing that!!
19:30 MiroslavAnashkin images_type and/or use_cow_images parameters.
19:32 hyperbaba use_cow_images is true
19:32 hyperbaba Should i change it to False?
19:33 hyperbaba images_type is unset
19:33 Dr_Drache https://www.androidfilehost.com/?fid=23622183712468883
19:34 HeOS joined #fuel
19:34 Dr_Drache sorry guys
19:34 MiroslavAnashkin hyperbaba: yes. on every node and restart openstack-nova-compute service
19:34 Dr_Drache wrong window
19:34 hyperbaba MiroslavAnashkin: and then the rdb snap mechanism will be used?
19:35 MiroslavAnashkin Should. Currently it downlodas raw, converts to qcow, uploads back and duting upload converts it back to raw on the fly
19:36 Dr_Drache MiroslavAnashkin, holy cow, that sounds like a waste
19:38 hyperbaba MiroslavAnashkin: But i remember some time ago that this was patched in Icehouse. Maybe even by You.
19:38 MiroslavAnashkin Yes. And if were no format conversion - Ceph creates snapshot in 1-2 seconds
19:40 hyperbaba MiroslavAnashkin: Now it's stuck in Queued state
19:41 hyperbaba MiroslavAnashkin: Do I need to set images_type ? And, if yes to what value ?
19:41 MiroslavAnashkin http://docs.openstack.org/trunk/config-reference/content/list-of-compute-config-options.html
19:42 hyperbaba MiroslavAnashkin: Thanks, will look into it.
19:42 MiroslavAnashkin images_type = raw
19:43 teran joined #fuel
19:44 MiroslavAnashkin Juno also has force_raw_images = True, cant remember, has it Icehouse as well or not
19:48 hyperbaba MiroslavAnashkin: Tried with that images_type set to raw. Now it's just stuck in queued state
19:49 Dr_Drache images_type = raw; makes Use_cow_images invalid
19:49 Dr_Drache according to the documentation
19:53 teran_ joined #fuel
19:57 MiroslavAnashkin Yes, Use_cow_images is for compatibility with old releases
19:57 e0ne joined #fuel
20:00 mpetason_ joined #fuel
20:05 hyperbaba MiroslavAnashkin: The instance is in state "Image pending upload" and image is in state "queued". No Errors in nova-all.log. Now I am confused,
20:07 hyperbaba MiroslavAnashkin: and qemu-img process is running on the node. Looks like it's converting again. But from what to what? qemu-img convert -O raw rbd:compute/0f79cce4-1d59-412b-8275-c2def619bbeb_disk:id=compute /var/lib/nova/instances/snapshots/tmpE68qZt/65fa4682007f4a5093979cc0bd533741
20:07 kupo24z joined #fuel
20:08 kupo24z xarses: Do you know if Juno has the Ceph Ephemeral code?
20:08 kupo24z or was that not ported upstream
20:09 xarses angdraug: ^
20:10 xarses I believe it was merged upstream already
20:10 xarses angdraug: is the maintainer so he can answer best
20:11 kaliya_ joined #fuel
20:14 kaliya_ nullme: you can compress them
20:16 jetole_ joined #fuel
20:16 jetole_ Hey guys. I am running neutron/vlan. Is there a way for me to use the same network on multiple projects?
20:20 syt joined #fuel
20:31 hyperbaba Does anybody knows what images_type=rbd does in nova.conf ?
20:36 Dr_Drache hyperbaba : https://review.openstack.org/gitweb?p=stackforge%2Fcookbook-openstack-compute.git;a=commitdiff;h=41c3a71cc67dba3cd622deca7b2595e4a4cf7d2e
20:38 tatyana joined #fuel
20:40 hyperbaba Dr_Drache: Don't see what is gained by this?
20:41 emagana joined #fuel
20:41 Dr_Drache hyperbaba, it seems that's how you use ebd for emph volumes.
20:41 emagana joined #fuel
20:41 Dr_Drache rbd
20:42 Dr_Drache it seems ceph isn't being configured "properly"
20:42 hyperbaba I think there is a separate cephx user for "compute" pool . Right?
20:43 Dr_Drache now, don't take my word on that, because I don't know what "properly" is, but reading the documentation makes me think.
20:43 Dr_Drache I belive so yes
20:45 hyperbaba So , you think maybe it's a problem that "cephx" user for compute triggers a rbd snap and it's destined for cepx "user" for glance/cinder? And there is a conflict?
20:48 hyperbaba As I remember this worked on 5.0 with andraug's patched code. raw images could be snapshoted to ceph instantly. Maybe the nova patch was reverted in 5.1?
20:49 Dr_Drache few things have reverted in 5.1 (api version for example) so, it's entirely possible
20:51 Dr_Drache I guess if we find that old patch
20:51 Dr_Drache we can look
20:51 hyperbaba Maybe angdraug knows the best
20:51 Dr_Drache he would
20:52 hyperbaba I remember that I took the patch and applied it manually to the nova code. Could not simply patch it because something else was changed. Of course I did that in a hurry and did not document it. Shame on me
20:55 * angdraug reads the channel backlog
20:56 angdraug no, snapshots *to* glance still don't use cow, even with all my patches to nova. problem is on the glance side
20:58 angdraug the opposite direction is fixed by my patches, and yes it was merged to juno
20:58 jetole_ Can I use the same openstack defined network with multiple projects? I'm using neutron/vlan
20:59 Dr_Drache angdraug, so this is something that is mostly fixed in juno, but icehouse is still cold?
20:59 angdraug we have all these patches in fuel/mos since 4.1, including 5.1
20:59 Dr_Drache ohhh
20:59 xarses joined #fuel
21:00 angdraug the regression with glance_api_version in cinder had nothing to do with openstack code, root cause was a merge problem with upstream puppet manifests for cinder
21:00 angdraug we didn't touch cephx stuff recently
21:00 angdraug what exactly is regressed?
21:00 hyperbaba angdraug: but i am trying to snapshot ephermal from raw image . Forced raw in nova.conf . And qemu-img is still called to "convert"
21:01 angdraug yes it is
21:01 hyperbaba angdraug: this worked in 5.0. I am certain of it
21:01 angdraug no, it didn't
21:01 angdraug the call to qemu-img convert is hardcoded in nova
21:02 hyperbaba angdraug: did you make a rbd snap call patch also?
21:02 hyperbaba angdraug: or someone else did. Was a few months back.
21:03 hyperbaba angdraug: I had that platform live upto month ago. Snapshoting a rdb raw ephermal was a breeze.
21:03 angdraug there was this bug: https://bugs.launchpad.net/fuel/+bug/1261590
21:03 kupo24z angdraug: do you know if standard juno can spawn instances from rbd ephemeral disks? or is it still only volumes
21:03 teran joined #fuel
21:03 angdraug but that was an outright crash, not lack of cow
21:05 angdraug hyperbaba: are you sure you were snapshotting an ephemeral and not volume backed
21:05 angdraug kupo24z: yes, juno can
21:06 angdraug there's couple of open bugs about snapshotting ephemeral on rbd:
21:06 angdraug https://bugs.launchpad.net/nova/+bug/1262914
21:06 hyperbaba angdraug: I never created a single volume on my test bed. Just ephermals. And i remember that i was patching something...But what....duno
21:06 angdraug https://bugs.launchpad.net/nova/+bug/1346525
21:07 hyperbaba Too much information for me today. Going to run naked over Skyrim now .
21:07 Dr_Drache ..........
21:08 Dr_Drache hyperbaba...
21:08 Dr_Drache good luck with that sir!
21:08 angdraug zoltan has recently proposed a patch for nova and josh added a patch for glance, tbh prospects of having these merged in juno are slim
21:08 angdraug we might pull it into 6.0 or 6.1 if these patches shape up fast enough
21:09 Dr_Drache angdraug, that would be awesome.
21:09 Dr_Drache :P
21:11 angdraug Dr_Drache: based on nick and reported real name, pretty sure hyperbaba expects a "mam", not "sir" :)
21:13 mpetason_ joined #fuel
21:27 adanin joined #fuel
21:28 emagana joined #fuel
21:33 nullme joined #fuel
21:33 emagana joined #fuel
21:35 nullme evgeniyl: just wanted to see if you got the snapshot?
21:39 blahRus joined #fuel
21:50 harybahh joined #fuel
22:04 xarses joined #fuel
22:23 nullme hi i uploaded a snapshot earlier today, just wondering if it was looked at?
22:38 xarses nullme: its in the middle of the night for evgeniyl__ now, so I don't think they would have looked at it.
22:39 nullme thank you for letting me know that
22:40 holser left #fuel
23:04 mpetason left #fuel
23:50 harybahh joined #fuel
23:54 adanin joined #fuel

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