| *** deep-book-gk_ has joined #openstack-sahara | 00:14 | |
| *** deep-book-gk_ has left #openstack-sahara | 00:16 | |
| *** tmckay has joined #openstack-sahara | 00:27 | |
| *** tuanluong has joined #openstack-sahara | 00:50 | |
| *** tmckay has quit IRC | 00:55 | |
| *** lucasxu has joined #openstack-sahara | 01:22 | |
| *** lucasxu has quit IRC | 01:24 | |
| *** zhuli has joined #openstack-sahara | 01:58 | |
| openstackgerrit | zhuli proposed openstack/sahara master: Fix TypeError when get resource list https://review.openstack.org/487264 | 03:24 |
|---|---|---|
| *** links has joined #openstack-sahara | 03:30 | |
| *** links has quit IRC | 03:30 | |
| *** rcernin has joined #openstack-sahara | 06:06 | |
| *** anshul has joined #openstack-sahara | 06:54 | |
| *** ukaynar has quit IRC | 07:06 | |
| *** ukaynar has joined #openstack-sahara | 07:07 | |
| *** ukaynar has quit IRC | 07:11 | |
| *** Poornima_K has joined #openstack-sahara | 07:37 | |
| *** Poornima_K has quit IRC | 07:45 | |
| *** ukaynar has joined #openstack-sahara | 08:15 | |
| *** esikachev has joined #openstack-sahara | 08:34 | |
| *** ukaynar has quit IRC | 08:37 | |
| *** ukaynar has joined #openstack-sahara | 08:38 | |
| *** tosky has joined #openstack-sahara | 08:42 | |
| *** ukaynar has quit IRC | 08:42 | |
| *** shuyingya has joined #openstack-sahara | 08:57 | |
| *** shuyingya has quit IRC | 08:59 | |
| *** shuyingya has joined #openstack-sahara | 09:05 | |
| *** shuyingya has quit IRC | 09:05 | |
| *** shuyingya has joined #openstack-sahara | 09:09 | |
| *** shuyingya has quit IRC | 09:10 | |
| *** shuyingya has joined #openstack-sahara | 09:27 | |
| *** mgoddard_ has joined #openstack-sahara | 09:30 | |
| *** tuanluong has quit IRC | 09:52 | |
| *** akuznetsov has joined #openstack-sahara | 09:56 | |
| *** shuyingya has quit IRC | 10:28 | |
| *** shuyingya has joined #openstack-sahara | 10:29 | |
| *** tosky has quit IRC | 10:45 | |
| *** akuznetsov has quit IRC | 10:48 | |
| *** esikachev has quit IRC | 11:05 | |
| openstackgerrit | ZhangHongtao proposed openstack/sahara master: Enable some off-by-default checks https://review.openstack.org/487379 | 11:45 |
| *** esikachev has joined #openstack-sahara | 12:01 | |
| *** esikachev has quit IRC | 12:05 | |
| iwonka | ping tellesnobrega | 12:13 |
| tellesnobrega | iwonka, pong | 12:13 |
| iwonka | i got weird problems when adding import | 12:15 |
| iwonka | but they burn down to "the JSON i'm exporting is not exactly a valid JSON" | 12:16 |
| *** chlong has joined #openstack-sahara | 12:16 | |
| iwonka | because it has single quotes instead of double | 12:16 |
| iwonka | and "False" instead of "false" | 12:16 |
| iwonka | and other stuff like that | 12:16 |
| iwonka | so, i tried to change export to use json.dumps | 12:16 |
| iwonka | it started escaping quotes | 12:17 |
| iwonka | so it's probably not a good idea | 12:17 |
| iwonka | and so, what is the correct approach to fix it? manually change the file so that json.loads worked? | 12:18 |
| iwonka | (by manually i mean str.replace in code…) | 12:18 |
| iwonka | also, one thing that just works is ast.literal_eval but it's not the safest way to do it… | 12:19 |
| tellesnobrega | iwonka, hmm, lets think. I believe that json.dumps creates a valid json so json can load it again | 12:25 |
| tellesnobrega | it doesn't make sense for a module to produce something that it doesn't understand | 12:26 |
| tellesnobrega | iwonka, can you try a test? Write a node_group_template, simple one, in a dict and use json.dumps | 12:27 |
| tellesnobrega | not using sahara, just a python shell | 12:27 |
| tellesnobrega | and after that try json.load | 12:27 |
| tellesnobrega | to see if that works | 12:27 |
| *** tmckay has joined #openstack-sahara | 12:28 | |
| iwonka | tellesnobrega: it works unless the dict is indexed with something else than strings (because then it converts that to str, so {1:"one} becomes {'1': 'one'} but probably it's not a big issue? | 12:32 |
| tellesnobrega | probably not, json.loads should bring stuff back | 12:33 |
| iwonka | it doesn't | 12:34 |
| iwonka | in this particular case it doesn't | 12:34 |
| tellesnobrega | it loads a dict of strings right? | 12:34 |
| iwonka | yes | 12:34 |
| iwonka | but in case of sahara stuff it's not an issue | 12:34 |
| iwonka | i think | 12:34 |
| tellesnobrega | that should be fine | 12:35 |
| tellesnobrega | yes | 12:35 |
| iwonka | so i should make json.dumps work? | 12:35 |
| tellesnobrega | I think json.dumps should work yes | 12:35 |
| tellesnobrega | this is on the sahara change, right? | 12:35 |
| iwonka | that's where i wanted to add it | 12:36 |
| tellesnobrega | wait | 12:36 |
| tellesnobrega | iwonka, what is the exact format we are returning from sahara to the CLI and Dashboard? | 12:39 |
| *** shuyingya has quit IRC | 12:40 | |
| iwonka | it's a response object | 12:40 |
| iwonka | (flask response) | 12:40 |
| tellesnobrega | ok, and it contains the dict, correct? | 12:41 |
| iwonka | yes, in res.response there's a one element list, which contains a string which contains a dict | 12:42 |
| tellesnobrega | cool | 12:42 |
| tellesnobrega | so I see two options: | 12:43 |
| tellesnobrega | 1) json.dumps on the sahara side and send that on the response | 12:44 |
| tellesnobrega | 2) json.dumps on the dashboard and CLI side before exporting | 12:44 |
| tellesnobrega | agree? | 12:44 |
| iwonka | mhmm | 12:45 |
| iwonka | probably in sahara it's beter | 12:45 |
| iwonka | because it neeeds to be done anyway | 12:45 |
| iwonka | so better in one place than in two? | 12:45 |
| iwonka | but the problem is (in sahara) u.render expects a dict and it turns it into a this list with a string with a json | 12:47 |
| iwonka | so should i use u.render at all? | 12:47 |
| iwonka | no, wait | 12:48 |
| iwonka | one exported with curl is ok | 12:48 |
| iwonka | so sahara does it' job | 12:48 |
| iwonka | but one exported with cli is not correct | 12:49 |
| iwonka | why do they even differ? | 12:49 |
| *** jeremyfreudberg has joined #openstack-sahara | 12:49 | |
| tellesnobrega | iwonka, lets get this straight. The template exported from the UI works with the import? | 12:50 |
| tellesnobrega | but the one with CLI doesn't? | 12:50 |
| tellesnobrega | UI/curl? | 12:50 |
| iwonka | the one exported with curl command works, the one exported with cli doesn't, i haven't tested UI one | 12:51 |
| tellesnobrega | so we need to check what the CLI is doing that affects this change | 12:51 |
| tellesnobrega | ok | 12:51 |
| * tellesnobrega checks the CLI change to possible causes | 12:51 | |
| jeremyfreudberg | str(response._info) is not the same thing as json, I guess | 12:52 |
| tellesnobrega | that makes sense | 12:53 |
| iwonka | but it should actually get a string somewhere | 12:54 |
| iwonka | the one sahara sends | 12:54 |
| iwonka | can i just change str(response._info) to json.dumps(response._info) ? | 12:55 |
| jeremyfreudberg | that would be the simplest thing | 12:55 |
| jeremyfreudberg | make sure you use that oslo_serialization import at the top of the file for json | 12:55 |
| tellesnobrega | iwonka, make sure that response._info is exactly the json format we need here. Maybe you need to json.dumps | 12:55 |
| tellesnobrega | oslo_serialization is already there | 12:56 |
| iwonka | json.loads is used | 12:56 |
| iwonka | so can i use json.dumps or should i use something from oslo_serialization? | 12:57 |
| jeremyfreudberg | use the json function that's already imported by oslo | 12:57 |
| jeremyfreudberg | to call json.dumps | 12:57 |
| tellesnobrega | json.dumps will work | 12:57 |
| tellesnobrega | oslo_serialization is imported as json | 12:58 |
| iwonka | tellesnobrega: in which file? i wanted to change python-saharaclient/saharaclient/osc/v1/node_group_templates.py | 12:59 |
| iwonka | there's import json | 13:00 |
| tellesnobrega | that is the one I'm looking at | 13:00 |
| tellesnobrega | at your change | 13:00 |
| tellesnobrega | L22 | 13:01 |
| tellesnobrega | from oslo_serialization import jsonutils as json | 13:01 |
| *** esikachev has joined #openstack-sahara | 13:01 | |
| iwonka | oh | 13:04 |
| iwonka | i changed it at some point | 13:04 |
| iwonka | nevermind, i'll fix it there | 13:04 |
| tellesnobrega | cool | 13:05 |
| *** shuyingya has joined #openstack-sahara | 13:08 | |
| *** dave-mccowan has joined #openstack-sahara | 13:09 | |
| openstackgerrit | Iwona Kotlarska proposed openstack/python-saharaclient master: Add export of node group templates to CLI https://review.openstack.org/486721 | 13:10 |
| *** shuyingya has quit IRC | 13:12 | |
| *** dave-mcc_ has joined #openstack-sahara | 13:15 | |
| *** dave-mccowan has quit IRC | 13:19 | |
| tellesnobrega | iwonka, did it work? | 13:24 |
| iwonka | yeah, but same problem with ui, i also used str() there | 13:24 |
| tellesnobrega | CLI is done :) | 13:25 |
| tellesnobrega | which is good | 13:25 |
| tellesnobrega | looking for reasons why it is not working on UI | 13:25 |
| tellesnobrega | content = str(saharaclient.nodegroup_template_export( | 13:26 |
| tellesnobrega | request, object_id)._info) | 13:26 |
| tellesnobrega | probably should use json.dumps here too | 13:26 |
| tellesnobrega | instead of str | 13:26 |
| iwonka | yes, i see | 13:26 |
| tellesnobrega | cool | 13:26 |
| tellesnobrega | let me know if it works | 13:26 |
| iwonka | should i use oslo_seriasization there? | 13:26 |
| tellesnobrega | yes | 13:27 |
| tellesnobrega | wait | 13:27 |
| tellesnobrega | oslo_serialization is not a dependency on the sahara-dashboard | 13:29 |
| jeremyfreudberg | argh | 13:29 |
| jeremyfreudberg | we should be able to add it though | 13:30 |
| tellesnobrega | yes | 13:31 |
| jeremyfreudberg | everyone downstream is packaging it, i hope | 13:31 |
| tellesnobrega | iwonka, add oslo.serialization and add this to requirements.txt | 13:31 |
| tellesnobrega | oslo.serialization!=2.19.1,>=1.10.0 # Apache-2.0 | 13:31 |
| iwonka | ok | 13:33 |
| jeremyfreudberg | alphabetically | 13:33 |
| *** rcernin has quit IRC | 13:54 | |
| openstackgerrit | Iwona Kotlarska proposed openstack/sahara-dashboard master: Add export of node group templates https://review.openstack.org/485215 | 13:58 |
| iwonka | tellesnobrega: done. it works now | 13:59 |
| iwonka | i'll return to working on import on cli now | 13:59 |
| jeremyfreudberg | iwonka: cool | 13:59 |
| iwonka | thanks for help :) | 14:00 |
| *** tomtomtom has joined #openstack-sahara | 14:03 | |
| *** dave-mcc_ has quit IRC | 14:05 | |
| *** ukaynar has joined #openstack-sahara | 14:06 | |
| tellesnobrega | iwonka, awesome | 14:13 |
| *** ssmith has joined #openstack-sahara | 14:17 | |
| *** esikachev has quit IRC | 14:20 | |
| ssmith | Good morning. This is the first time that we're firing up Data Processing on our system. We think we have it configured correctly and am using a pre-configured Sparky image but are getting a "Boot failed: not a bootable disk" and "No bootable device" | 14:28 |
| tellesnobrega | hi ssmith, did you build the image yourself? | 14:29 |
| ssmith | No, as per the documentation we got it from http://sahara-files.mirantis.com/images/upstream/ | 14:30 |
| tellesnobrega | ssmith, ocata version? | 14:30 |
| ssmith | Newton | 14:31 |
| tellesnobrega | ssmith, great. I will take a look at this asap to check if the image is fine and I will let you know | 14:32 |
| ssmith | Newton 14.07 deployed with Ansible from repo | 14:32 |
| tellesnobrega | this is the image right? sahara-newton-spark-1.6.0-ubuntu.qcow2 | 14:33 |
| tellesnobrega | ssmith, ^ | 14:34 |
| ssmith | Yes, correct | 14:34 |
| tellesnobrega | i will take a look into that | 14:35 |
| jeremyfreudberg | ssmith, have you tried booting that image on a regular instance (not as sahara cluster) | 14:35 |
| ssmith | hmmm, no. I will | 14:35 |
| tellesnobrega | also a good test | 14:35 |
| tellesnobrega | could be an issue when downloading the image | 14:36 |
| jeremyfreudberg | for sanity, check the checksum as well | 14:36 |
| ssmith | Here's what's on the console. Still doing a seperate launch http://imgur.com/PocWcZs | 14:43 |
| jeremyfreudberg | not familiar with that environment, but I assume it's capable of booting from other qcow images? | 14:44 |
| ssmith | That's the console page under the instance. Yes, other qcow images boot | 14:48 |
| jeremyfreudberg | ssmith, and you verified with the checksum? | 14:49 |
| iwonka | tellesnobrega: tests fail when i run them from my deployment, but it seems to be unrelated to my change (FAIL: test_launch_cluster_get_nodata (sahara_dashboard.content.data_processing.clusters.clusters.tests.DataProcessingClusterTests)) so should i try to fix that? | 14:57 |
| jeremyfreudberg | tellesnobrega/iwonka seems like the dashboard failures are a real problem, but not related to iwonka's/my recent patches | 14:58 |
| openstackgerrit | Merged openstack/sahara master: Add export of node group templates https://review.openstack.org/478245 | 14:58 |
| iwonka | jeremyfreudberg: ok, i'm not touching that in this case | 14:59 |
| ssmith | md5sum returns 698418b0b6857022a5ce49d80ebe9846 sahara-newton-spark-1.6.0-ubuntu.qcow2 | 15:02 |
| jeremyfreudberg | yep, that's the correct checksum | 15:03 |
| *** esikachev has joined #openstack-sahara | 15:05 | |
| ssmith | Our normal centos image launches fine. The sahara image give this: http://imgur.com/5qgqrT9 | 15:06 |
| jeremyfreudberg | block device mapping is usually boot from volume, no? | 15:07 |
| ssmith | boot source = image = sahara-newton-spark-1.6.0-ubuntu.qcow2 | 15:09 |
| jeremyfreudberg | alright | 15:09 |
| jeremyfreudberg | i'm pretty sure i've used that image before | 15:11 |
| jeremyfreudberg | without issue | 15:11 |
| ssmith | OK wait. I just tried myself and it booted fine and was able to ssh into it as well | 15:15 |
| jeremyfreudberg | ssmith, so the issue was just temporary? | 15:15 |
| ssmith | http://imgur.com/XNPnwZW | 15:15 |
| ssmith | No, I thing the other guy left the root volume at 1 GB while I changed it to 10 | 15:18 |
| jeremyfreudberg | i see | 15:18 |
| ssmith | Let me try a cluster again | 15:19 |
| ssmith | Been on validating for 1:30 minutes | 15:20 |
| jeremyfreudberg | ok, that's more of a problem then to be stuck on validating | 15:21 |
| jeremyfreudberg | that's usually the 2nd-fastest part | 15:21 |
| ssmith | The error | 15:21 |
| ssmith | Then error | 15:21 |
| jeremyfreudberg | anything in sahara logs? | 15:21 |
| *** rcernin has joined #openstack-sahara | 15:22 | |
| ssmith | http://imgur.com/Z9Onx4P | 15:23 |
| ssmith | Checking on the logs | 15:23 |
| jeremyfreudberg | ok, well that's just saying glance api is down | 15:24 |
| jeremyfreudberg | ssmith, I have to be afk for a while, but I'm usually around | 15:27 |
| *** shuyingya has joined #openstack-sahara | 15:30 | |
| *** jeremyfreudberg has quit IRC | 15:30 | |
| *** jeremyfreudberg has joined #openstack-sahara | 15:30 | |
| *** jeremyfreudberg has quit IRC | 15:31 | |
| *** rcernin has quit IRC | 15:34 | |
| iwonka | ping tellesnobrega | 15:35 |
| iwonka | when importing, things like tenant_id are added automatically | 15:36 |
| iwonka | maybe it should be removed? instead of changed to placeholder? | 15:36 |
| iwonka | same with is_default | 15:37 |
| iwonka | create doesn't want it as an argument | 15:38 |
| iwonka | and security groups when they are exported the default way do not work | 15:41 |
| *** rcernin has joined #openstack-sahara | 16:03 | |
| *** chlong has quit IRC | 16:08 | |
| *** esikachev has quit IRC | 16:15 | |
| *** shuyingya has quit IRC | 16:17 | |
| ssmith | We do notice that the root volume for the cluster instances are not marked as bootable which is why the instance is saying "No bootable device" | 16:22 |
| *** shuyingya has joined #openstack-sahara | 16:26 | |
| tellesnobrega | iwonka, regarding dashboard failures I have to trace back to when it started | 16:38 |
| tellesnobrega | seems unrelated | 16:38 |
| tellesnobrega | maybe some dependency | 16:38 |
| *** shuyingya has quit IRC | 16:38 | |
| tellesnobrega | ssmith, have you registered the image using sahara image registry? | 16:40 |
| tellesnobrega | iwonka, step by step here | 16:41 |
| tellesnobrega | when you are importing the tenant_id is added, the current one? | 16:41 |
| ssmith | Yes, and it attaches as vdb which I'm pretty sure won't boot on KVM no matter if it's marked as bootable or not | 16:41 |
| ssmith | the compute driver tries to mount vda, yet the heat template is attaching on vdb. | 16:42 |
| iwonka | tellesnobrega: yes, but create doesn't want it as argument. it's added in sahara itself (id of current project) | 16:43 |
| tellesnobrega | if that is the case we can remove from the export without a problem | 16:44 |
| tellesnobrega | same for is_default | 16:44 |
| tellesnobrega | ssmith, I'm adding the image here and testing it | 16:44 |
| *** anshul has quit IRC | 16:46 | |
| iwonka | tellesnobrega: and what about security groups? | 16:47 |
| iwonka | and should it be in a new commit? | 16:47 |
| tellesnobrega | the change will be on sahara? or other parts? | 16:48 |
| tellesnobrega | if in sahara yes, your change got merged | 16:48 |
| iwonka | yes, sahara | 16:49 |
| tellesnobrega | so a new patch | 16:49 |
| iwonka | ok | 16:49 |
| tellesnobrega | what is not working on security groups? | 16:49 |
| iwonka | most probably i'm not sure what to put there | 16:51 |
| tellesnobrega | ssmith, I just successfully started a cluster using that image | 16:53 |
| ssmith | IS the volume attached to the instance as vda or vdb? | 16:54 |
| tellesnobrega | I have it with no volumes attached | 16:54 |
| tellesnobrega | where did you put a volume? worker instances? | 16:55 |
| ssmith | Are our templates configured incorrectly maybe? | 16:56 |
| tellesnobrega | it is possible, can you show them? | 16:56 |
| ssmith | http://imgur.com/0MndAnL | 16:57 |
| tellesnobrega | that looks fine to me | 16:58 |
| ssmith | http://imgur.com/fx2rDpc | 16:58 |
| tellesnobrega | I'm adding volumes to my template here | 16:58 |
| ssmith | http://imgur.com/Pvo1Kxr | 16:59 |
| ssmith | http://imgur.com/XeWKSPD | 17:01 |
| tellesnobrega | on node groups templates, do you have one for the master and one for slaves or just a single template with all services running? | 17:01 |
| ssmith | ONe for slave and one for master. Slave has datanode and slave as Node Processes | 17:02 |
| *** mgoddard_ has quit IRC | 17:02 | |
| tellesnobrega | ok, that is perfect | 17:03 |
| tellesnobrega | on cluster template did you add both NG_templates? | 17:03 |
| tellesnobrega | on the Node Groups* tab | 17:03 |
| ssmith | Yes one of each with count of 1 each | 17:04 |
| tellesnobrega | ok | 17:04 |
| *** anshul has joined #openstack-sahara | 17:04 | |
| tellesnobrega | under the HDFS Parameters you should update a dfs.replication parameter | 17:05 |
| tellesnobrega | it should be 1, since you are adding just one slave | 17:05 |
| tellesnobrega | it will fail validation otherwise | 17:05 |
| ssmith | says 3 | 17:05 |
| tellesnobrega | since you are adding just one datanode it should be changed to 1 | 17:06 |
| ssmith | OK done | 17:06 |
| tellesnobrega | cool | 17:06 |
| openstackgerrit | Iwona Kotlarska proposed openstack/sahara master: Fix export of node group templates https://review.openstack.org/487523 | 17:06 |
| tellesnobrega | started a cluster here with volumes and it is attached on /dev/vdb | 17:08 |
| tellesnobrega | iwonka, have you figured out what to do with security_groups? | 17:08 |
| openstackgerrit | Iwona Kotlarska proposed openstack/sahara master: Fix export of node group templates https://review.openstack.org/487523 | 17:10 |
| ssmith | So we don't have to use volumes? | 17:10 |
| *** anshul has quit IRC | 17:11 | |
| tellesnobrega | You can, that really depends on your needs | 17:11 |
| *** ukaynar has quit IRC | 17:11 | |
| ssmith | OK, this time got overquotient on security group rules but they're set to 1000 for the project | 17:11 |
| *** ukaynar has joined #openstack-sahara | 17:11 | |
| tellesnobrega | what I see here is that we do have instances with volumes | 17:12 |
| ssmith | Did it boot. What does it say in the console of the instance? | 17:12 |
| tellesnobrega | ssmith, everytime you run a cluster with auto_security_groups it will create a new group with a lot of rules (required by the service) | 17:13 |
| ssmith | NVM, SG rules needed to be increased | 17:13 |
| tellesnobrega | so probably deleting some sec groups will clear that | 17:13 |
| tellesnobrega | about the instance it booted and I can ssh to it | 17:14 |
| tellesnobrega | http://paste.openstack.org/show/616606/ | 17:15 |
| tellesnobrega | this is from inside the instance | 17:15 |
| *** ukaynar has quit IRC | 17:16 | |
| ssmith | I don't see any volume created that's vda1 or vda attached to the instance like yours | 17:19 |
| tellesnobrega | my attached volume is at vdb | 17:20 |
| tellesnobrega | it is a 2GB volume | 17:20 |
| ssmith | You have a root /dev/vda1. Is it ephemeral or cinder? I don't get a vda attached to the instance | 17:21 |
| tellesnobrega | the vda is ephemeral | 17:22 |
| tellesnobrega | the cinder one is vdb | 17:22 |
| ssmith | We don't have ephemeral on our compute nodes. Does the root need to be ephemeral? | 17:22 |
| tellesnobrega | hmm, got it. Not sure, I have to check that | 17:23 |
| ssmith | What's the difference in Base Image in the NGT and under the Launch Cluster UI? | 17:25 |
| *** chlong has joined #openstack-sahara | 17:25 | |
| tellesnobrega | ssmith, none actually, this is something I've meaning to work on, reduce the number of duplicity | 17:26 |
| *** chlong has quit IRC | 17:27 | |
| *** chlong has joined #openstack-sahara | 17:28 | |
| *** jeremyfreudberg has joined #openstack-sahara | 17:34 | |
| jeremyfreudberg | ssmith, tellesnobrega, I'm back - just reading over the logs now... anything that still needs clarity? | 17:34 |
| tellesnobrega | jeremyfreudberg, yes, basically the question now is, does we must have ephemeral disks to start the instances? | 17:35 |
| jeremyfreudberg | considering that this spec: http://specs.openstack.org/openstack/sahara-specs/specs/ocata/boot-from-volume.html remains unimplemented, essentially you do need ephemeral disk | 17:36 |
| jeremyfreudberg | in other, simpler, words - there is no support for boot from volume yet | 17:36 |
| tellesnobrega | jeremyfreudberg, true. I forgot about this spec | 17:37 |
| jeremyfreudberg | nikita had started work on it but abandoned it due to other responsibilities.... I was planning to bring this up again at PTG, about whether we should prioritize it. So I guess the answer is yes, we need it | 17:38 |
| tellesnobrega | ssmith, ^ | 17:38 |
| tellesnobrega | jeremyfreudberg, the answer is yes for sure | 17:39 |
| *** mgoddard_ has joined #openstack-sahara | 17:40 | |
| tellesnobrega | jeremyfreudberg, added the topic to the etherpad | 17:43 |
| jeremyfreudberg | tellesnobrega: thanks for adding it. (actually I already had it hiding in sub-bulletpoint but the attention is good) | 17:44 |
| tellesnobrega | I see that | 17:50 |
| tellesnobrega | lets keep working on that etherpad (even thou it is already quite big) and I will work on a schedule | 17:51 |
| jeremyfreudberg | yep | 17:51 |
| jeremyfreudberg | also, you saw vitaly's email? | 17:51 |
| tellesnobrega | not yet | 17:52 |
| jeremyfreudberg | the name was russian, you might have missed it | 17:53 |
| ssmith | So one of the infrastructure dependencies for Sahara is to have ephemeral volume capabilities? | 17:53 |
| jeremyfreudberg | ssmith, unfortunately yes, at this time you must be able to boot instances from ephemeral disk | 17:54 |
| ssmith | I must have missed that somewhere. | 17:54 |
| jeremyfreudberg | ssmith, we can probably clarify in the docs | 17:54 |
| jeremyfreudberg | but anyway, lack of boot from volume support is a known issue, there will be a discussion if it can be resolved during Q release | 17:55 |
| jeremyfreudberg | sorry we don't have a better answer | 17:55 |
| *** esikachev has joined #openstack-sahara | 17:56 | |
| tellesnobrega | ssmith, we can't promise it will be solved in Queen as well, our team is a lot smaller than it used to be and we need to keep a priority list with our capabilities, but it is something important and will try to prioritize it | 18:02 |
| tellesnobrega | for ASAP | 18:02 |
| tellesnobrega | jeremyfreudberg, about vitaly's email, I totally agree with his point. I've had that in my mind for some time, I wanted to ask you personally at the PTG | 18:05 |
| tellesnobrega | but it makes sense to bring you in now with vitaly leaving | 18:05 |
| jeremyfreudberg | tellesnobrega, cool, thanks for you endorsement. it means a lot to me. :) | 18:05 |
| tellesnobrega | jeremyfreudberg, check openstack-dev | 18:12 |
| jeremyfreudberg | tellesnobrega, awesome | 18:22 |
| *** cpusmith has joined #openstack-sahara | 18:39 | |
| *** ssmith has quit IRC | 18:39 | |
| *** cpusmith_ has joined #openstack-sahara | 18:40 | |
| cpusmith_ | We have Cinder on Ceph so have just created an Ephemeral Pool. So how do we tell Sahara to use it? A volume type? On the Flavor? | 18:40 |
| jeremyfreudberg | cpusmith_ - the flavor | 18:41 |
| *** mgoddard_ has quit IRC | 18:43 | |
| *** cpusmith has quit IRC | 18:43 | |
| iwonka | tellesnobrega: the problem with security groups is that i'm not sure how to provide None as an argument for it to be parsed correctly | 18:50 |
| iwonka | i mean, it could be default | 18:50 |
| iwonka | and set to None | 18:50 |
| iwonka | and then when user doesn't provide it, the it would be None | 18:50 |
| iwonka | that works | 18:50 |
| iwonka | can i do it that way? | 18:51 |
| iwonka | (the same problem with floating ip pool and sth else probably) | 18:51 |
| tellesnobrega | hm, security_group can be set as auto_security_group | 18:53 |
| tellesnobrega | this way the security group is create after | 18:53 |
| tellesnobrega | maybe you can take a look into that | 18:54 |
| jeremyfreudberg | so auto_security_group is bool | 18:54 |
| jeremyfreudberg | but security_groups extra is a list | 18:54 |
| jeremyfreudberg | do you know if security_groups extra can take the empty list as input? | 18:54 |
| tellesnobrega | floating_ip pool has to be requested as well, it can be similar to what is done on the create, where it has a drop down (you can check how templates is formed after that) | 18:57 |
| iwonka | tellesnobrega: | 18:57 |
| iwonka | i'm working with cli now | 18:57 |
| iwonka | so no drop down | 18:58 |
| cpusmith_ | Anything that we can do about Clusters that are just sitting there deleting? | 18:59 |
| jeremyfreudberg | cpusmith_ if they are stuck deleting try deleting the heat stack first | 19:00 |
| jeremyfreudberg | you have to do `openstack stack list --hidden --nested` to see them | 19:00 |
| tellesnobrega | jeremyfreudberg, I'm not sure if it can be an empty list for sec groups extra | 19:00 |
| jeremyfreudberg | tellesnobrega: really not sure either, probably something in the validation code | 19:01 |
| cpusmith_ | There are none showing in the UI | 19:01 |
| jeremyfreudberg | cpusmith_ they are hidden from ui | 19:01 |
| jeremyfreudberg | you might have missed my cli hint above | 19:01 |
| cpusmith_ | OK, will do from CLI then | 19:01 |
| tellesnobrega | iwonka, what I meant was just to check how templates are created when those options are selected | 19:02 |
| tellesnobrega | so you know what to put there | 19:02 |
| jeremyfreudberg | cpusmith_ : if the stacks won't delete either, then you can enable heat stack abandon and do that. but that's a last resort | 19:02 |
| *** dave-mccowan has joined #openstack-sahara | 19:03 | |
| tellesnobrega | jeremyfreudberg, checking the validation | 19:03 |
| *** dave-mccowan has quit IRC | 19:04 | |
| jeremyfreudberg | tellesnobrega: theoretically it works, https://github.com/openstack/sahara/blob/master/sahara/service/validations/base.py#L184 | 19:05 |
| jeremyfreudberg | it will just iterate over the empty list | 19:05 |
| jeremyfreudberg | there could be something on the api side though | 19:05 |
| jeremyfreudberg | probably best to just try it out | 19:05 |
| tellesnobrega | yeah, on this looks it will work | 19:07 |
| cpusmith_ | jeremyfreudberg Are they under the tenant were the clusters are launched or under another one? | 19:11 |
| jeremyfreudberg | cpusmith_ they are under that tenant | 19:12 |
| jeremyfreudberg | heat stack-list --show-nested --show-hidden, or openstack stack list --hidden --nested | 19:12 |
| cpusmith_ | hidden showed them | 19:13 |
| jeremyfreudberg | great, try deletion, you can tail the heat-engine log too which might give some insight as to why cluster deletion hangs | 19:13 |
| jeremyfreudberg | probably it can't clean up a resource (i've seen auto-created security groups fail to delete because the instance they are associated with failed to spawn) | 19:14 |
| tellesnobrega | cpusmith_, this is also an issue that we are trying to resolve asap | 19:18 |
| *** rcernin has quit IRC | 19:27 | |
| *** mgoddard_ has joined #openstack-sahara | 19:29 | |
| iwonka | tellesnobrega: in cli in ngt create --floating-ip-pool is optional and when it's not created it is set to None | 19:29 |
| iwonka | same with security groups | 19:29 |
| jeremyfreudberg | iwonka: are you adding client library method for import first? | 19:32 |
| iwonka | yes | 19:32 |
| jeremyfreudberg | ok, just checking | 19:32 |
| jeremyfreudberg | anyway, if it works with none, then let's stick with none | 19:32 |
| jeremyfreudberg | no need to complicate it | 19:32 |
| iwonka | the point is, it's not required in create | 19:32 |
| jeremyfreudberg | i see | 19:33 |
| jeremyfreudberg | makes sense | 19:33 |
| iwonka | so my question is if it can be not required in import | 19:33 |
| jeremyfreudberg | i think it can be not required, but if the user doesn't provide a value, you have to set the keys which still have a placeholder to None | 19:33 |
| jeremyfreudberg | which is what I think you had figured out already anyway | 19:34 |
| *** esikachev has quit IRC | 19:34 | |
| iwonka | ok, thanks | 19:34 |
| jeremyfreudberg | well | 19:34 |
| jeremyfreudberg | except for image id | 19:34 |
| jeremyfreudberg | i guess | 19:34 |
| jeremyfreudberg | right | 19:34 |
| jeremyfreudberg | and flavor | 19:35 |
| jeremyfreudberg | or image id might technically be optional | 19:35 |
| jeremyfreudberg | but flavor should not be allowed to be none | 19:35 |
| iwonka | i'll leave flavor id and image id the way they are | 19:36 |
| *** esikachev has joined #openstack-sahara | 19:36 | |
| tellesnobrega | yes | 19:37 |
| tellesnobrega | sorry | 19:37 |
| tellesnobrega | iwonka, this way you discussed with jeremy looks good | 19:38 |
| iwonka | ok, cli works fine now | 19:40 |
| openstackgerrit | Merged openstack/sahara master: Fix export of node group templates https://review.openstack.org/487523 | 19:40 |
| *** chlong has quit IRC | 19:43 | |
| *** chlong has joined #openstack-sahara | 19:49 | |
| *** chlong has quit IRC | 19:52 | |
| *** chlong has joined #openstack-sahara | 19:52 | |
| openstackgerrit | Iwona Kotlarska proposed openstack/python-saharaclient master: Add import of node group templates https://review.openstack.org/487575 | 19:56 |
| openstackgerrit | Iwona Kotlarska proposed openstack/python-saharaclient master: Add import of node group templates https://review.openstack.org/487575 | 19:58 |
| *** esikachev has quit IRC | 20:04 | |
| *** mgoddard_ has quit IRC | 20:06 | |
| *** ukaynar has joined #openstack-sahara | 20:07 | |
| *** anshul has joined #openstack-sahara | 20:42 | |
| *** jeremyfreudberg has quit IRC | 20:48 | |
| *** mgoddard_ has joined #openstack-sahara | 20:56 | |
| *** ukaynar_ has joined #openstack-sahara | 20:58 | |
| cpusmith_ | jeremyfreudberg do you have 0 for root and 10 for ephemeral on your flavor? | 21:00 |
| *** esikachev has joined #openstack-sahara | 21:01 | |
| *** ukaynar has quit IRC | 21:02 | |
| *** ukaynar_ has quit IRC | 21:03 | |
| *** ukaynar has joined #openstack-sahara | 21:03 | |
| *** esikachev has quit IRC | 21:05 | |
| tellesnobrega | cpusmith_, you mean volume size? | 21:05 |
| cpusmith_ | yes | 21:06 |
| tellesnobrega | I'm using default small flavor | 21:06 |
| tellesnobrega | let me check the configs guere | 21:06 |
| tellesnobrega | ephemeral is 0 and disk is 20GB | 21:07 |
| cpusmith_ | The how is it booting to ephemeral? | 21:08 |
| *** ukaynar has quit IRC | 21:08 | |
| cpusmith_ | then how | 21:08 |
| tellesnobrega | let me make sure this is right | 21:09 |
| tellesnobrega | looking at http://paste.openstack.org/show/616606/ | 21:12 |
| tellesnobrega | ok, so I think this ephemeral disk is sized 20GB | 21:12 |
| tellesnobrega | that is probably the correct information | 21:13 |
| tellesnobrega | I have to leave now, please ping me and I will respond when I'm back | 21:13 |
| tellesnobrega | or reach me at tenobreg@redhat.com | 21:13 |
| tellesnobrega | cpusmith_, ^ | 21:13 |
| tellesnobrega | hopefully we will get that working soon | 21:14 |
| tellesnobrega | ephemeral is 20GB :) | 21:14 |
| *** chlong has quit IRC | 21:14 | |
| *** anshul has quit IRC | 21:21 | |
| *** ukaynar has joined #openstack-sahara | 21:33 | |
| *** mgoddard_ has quit IRC | 21:34 | |
| *** ukaynar has quit IRC | 21:42 | |
| *** ukaynar has joined #openstack-sahara | 21:43 | |
| *** cpusmith_ has quit IRC | 21:51 | |
| *** ukaynar has quit IRC | 21:53 | |
| *** esikachev has joined #openstack-sahara | 22:02 | |
| *** esikachev has quit IRC | 22:06 | |
| *** jeremyfreudberg has joined #openstack-sahara | 22:15 | |
| *** jeremyfreudberg has quit IRC | 22:17 | |
| *** esikachev has joined #openstack-sahara | 23:03 | |
| *** esikachev has quit IRC | 23:07 | |
| *** iwonka has quit IRC | 23:19 | |
| *** iwonka has joined #openstack-sahara | 23:27 | |
| *** ukaynar has joined #openstack-sahara | 23:30 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!