*** shuyingya has joined #openstack-sahara | 00:56 | |
*** shuyingya has quit IRC | 01:41 | |
*** shuyingya has joined #openstack-sahara | 01:43 | |
*** kbaegis has joined #openstack-sahara | 02:23 | |
kbaegis | Hey all. Trying to get the sahara-dashboard enabled | 02:24 |
---|---|---|
kbaegis | I would really appreciate some completion of the README | 02:24 |
*** shuyingya has quit IRC | 02:30 | |
*** shuyingya has joined #openstack-sahara | 02:43 | |
*** shuyingya has quit IRC | 02:59 | |
*** shuyingya has joined #openstack-sahara | 03:00 | |
*** david-lyle has quit IRC | 03:04 | |
*** brad[]` has quit IRC | 03:38 | |
*** shuyingya has quit IRC | 03:39 | |
*** kbaegis has quit IRC | 03:51 | |
*** kbaegis has joined #openstack-sahara | 03:52 | |
*** kbaegis has joined #openstack-sahara | 03:52 | |
*** shuyingya has joined #openstack-sahara | 03:57 | |
*** kbaegis has quit IRC | 04:09 | |
*** kbaegis has joined #openstack-sahara | 04:18 | |
openstackgerrit | Shu Yingya proposed openstack/python-saharaclient: Add newline to strings in stdout/stderr.write() https://review.openstack.org/376114 | 04:21 |
*** kbaegis has quit IRC | 04:26 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/python-saharaclient: Updated from global requirements https://review.openstack.org/376143 | 04:27 |
*** shuyingya has quit IRC | 04:52 | |
*** kbaegis has joined #openstack-sahara | 04:55 | |
*** kbaegis has quit IRC | 05:02 | |
*** shuyingya has joined #openstack-sahara | 05:11 | |
*** DuncanT_ has joined #openstack-sahara | 05:15 | |
*** rickflare2 has joined #openstack-sahara | 05:17 | |
openstackgerrit | Shu Yingya proposed openstack/python-saharaclient: Add newline to strings in stdout/stderr.write() https://review.openstack.org/376114 | 05:18 |
*** dgonzalez_ has joined #openstack-sahara | 05:19 | |
*** DuncanT has quit IRC | 05:21 | |
*** mionkin has quit IRC | 05:21 | |
*** hogepodge has quit IRC | 05:21 | |
*** dgonzalez has quit IRC | 05:21 | |
*** rickflare has quit IRC | 05:21 | |
*** logan- has quit IRC | 05:21 | |
*** rickflare2 is now known as rickflare | 05:21 | |
*** dgonzalez_ is now known as dgonzalez | 05:21 | |
*** logan- has joined #openstack-sahara | 05:21 | |
*** shuyingya has left #openstack-sahara | 05:24 | |
*** shuyingya has joined #openstack-sahara | 05:24 | |
*** mionkin has joined #openstack-sahara | 05:28 | |
openstackgerrit | Tuan Luong-Anh proposed openstack/sahara: TrivialFix Using sys.exit(main()) instead of main() https://review.openstack.org/376174 | 05:29 |
*** rcernin has joined #openstack-sahara | 05:30 | |
*** pgadiya has joined #openstack-sahara | 05:32 | |
*** DuncanT_ is now known as DuncanT | 05:36 | |
*** Poornima has joined #openstack-sahara | 05:36 | |
*** sudipto_ has joined #openstack-sahara | 05:47 | |
*** sudswas__ has joined #openstack-sahara | 05:47 | |
*** nkrinner_afk is now known as nkrinner | 06:18 | |
*** vgridnev has joined #openstack-sahara | 06:29 | |
*** shuyingya has quit IRC | 06:29 | |
*** shuyingya has joined #openstack-sahara | 06:31 | |
*** shuyingya has joined #openstack-sahara | 06:32 | |
*** pcaruana has joined #openstack-sahara | 06:45 | |
*** tnovacik has joined #openstack-sahara | 06:53 | |
*** shuyingya has quit IRC | 07:15 | |
*** shuyingya has joined #openstack-sahara | 07:15 | |
*** shuyingya has quit IRC | 07:19 | |
*** akuznetsov has joined #openstack-sahara | 07:29 | |
*** shuyingya has joined #openstack-sahara | 07:33 | |
*** vgridnev has quit IRC | 07:41 | |
*** abehl has joined #openstack-sahara | 08:11 | |
*** zemuvier has joined #openstack-sahara | 08:23 | |
*** vgridnev has joined #openstack-sahara | 08:26 | |
openstackgerrit | Merged openstack/sahara-ci-config: Changed filters for branches with skip-if https://review.openstack.org/374097 | 08:38 |
openstackgerrit | Merged openstack/sahara-ci-config: Added missed filter for cdh 5.7.0 job https://review.openstack.org/375381 | 08:38 |
*** esikachev has joined #openstack-sahara | 08:45 | |
openstackgerrit | Alina Nesterova proposed openstack/sahara-ci-config: Add participant for nightly and weekly jobs https://review.openstack.org/376301 | 08:46 |
*** zemuvier has quit IRC | 09:26 | |
*** tosky has joined #openstack-sahara | 09:28 | |
tosky | vgridnev: heeeello, did you see https://bugs.launchpad.net/sahara/+bug/1627639 ? | 09:30 |
openstack | Launchpad bug 1627639 in Sahara "API break: image metadata format changed in Newton" [Undecided,New] | 09:30 |
vgridnev | tosky, will take a look | 09:32 |
*** esikachev has quit IRC | 09:34 | |
*** pgadiya has quit IRC | 09:35 | |
*** shuyingya has quit IRC | 09:41 | |
*** pgadiya has joined #openstack-sahara | 09:48 | |
openstackgerrit | Merged openstack/python-saharaclient: Updated from global requirements https://review.openstack.org/376143 | 09:50 |
vgridnev | tosky, argh | 09:57 |
vgridnev | :( | 09:57 |
tosky | vgridnev: confirmed? | 09:57 |
*** esikachev has joined #openstack-sahara | 09:58 | |
vgridnev | tosky, the general issue is that sahara is proxy, so we have a dependency on data returned by glance | 09:59 |
tosky | vgridnev: right, but as proxy it should be able to transform the data | 10:00 |
tosky | and a TODO for v2: don't depend directly on glance API | 10:00 |
tosky | (for the format of the results) | 10:00 |
*** shuyingya has joined #openstack-sahara | 10:03 | |
vgridnev | tosky, is it used somewhere? | 10:04 |
vgridnev | I mean, metadata fields | 10:04 |
tosky | vgridnev: that method? It was used in one of our disabled test, which I was trying to reenable | 10:05 |
tosky | the question is: was that information part of the public API? | 10:05 |
tosky | probably yes, as it was advertised here: http://developer.openstack.org/api-ref-data-processing-v1.1.html#job-binary-internals-data%G | 10:06 |
tosky | so we need to restore it | 10:07 |
*** zemuvier has joined #openstack-sahara | 10:07 | |
vgridnev | I would say no | 10:08 |
openstackgerrit | Merged openstack/sahara-ci-config: Add participant for nightly and weekly jobs https://review.openstack.org/376301 | 10:08 |
tosky | vgridnev: not part of the public API? It's in the documentation of the public API :) | 10:11 |
vgridnev | Nope. I'm not going to restore that | 10:13 |
vgridnev | These fields are exposed by other API's and there is no way to restore all of them | 10:14 |
*** akuznetsov has quit IRC | 10:15 | |
tosky | yes, it is possible | 10:16 |
tosky | they are marked as _sahara_something | 10:16 |
tosky | well, at least in the old API | 10:16 |
tosky | I'm not saying to not expose the fields in the new way, but just find and expose the expected ones | 10:17 |
tosky | API breakages are serious things | 10:17 |
vgridnev | metadata is not sahara field at all | 10:17 |
tosky | well, it is, it's in the public API | 10:18 |
tosky | would you oppose a patch which restores them, without touching the other fields? | 10:18 |
tosky | moreover, the change does not depend on glance, it is all in https://review.openstack.org/#/c/334569/, sahara/utils/openstack/images.py does the translation | 10:23 |
tosky | so it's easy to get them and expose them under metadata | 10:24 |
vgridnev | Anyway, even if I will restore everything, I'm NOT recommending fields not 'sahara' posted fields | 10:28 |
vgridnev | ok, sorry tosky, I was 100% wrong. Now i'm working hard on patch to restore missing fields, could you please remove replace usages of 'metadata' field in tests? | 10:35 |
tosky | vgridnev: np - we can keep the failing the tests disabled, but if we want to restore it we need to test for both, as it should work on older versions | 10:36 |
tosky | another reason to try to push API v2, were we can freely do cleanups | 10:36 |
openstackgerrit | Evgeny Sikachev proposed openstack/sahara-tests: Add Hive job for Ambari https://review.openstack.org/299358 | 10:37 |
vgridnev | tosky, we always returning tags and usernames, so it will be working on all versions. So, we can replace 'metadata' usages to use 'tags' and username fields. | 10:38 |
vgridnev | Moreover, that is the only correct method of validating that image registered | 10:39 |
tosky | right | 10:39 |
vgridnev | saying about https://github.com/openstack/sahara-tests/blob/66dc1c0411ba760a0c7a60c67807d9e9bbb8a4de/sahara_tempest_plugin/tests/clients/test_job_executions.py#L39 | 10:40 |
*** rodrigods has quit IRC | 10:43 | |
tosky | vgridnev: yep, that's the one disabled | 10:43 |
*** rodrigods has joined #openstack-sahara | 10:43 | |
tosky | let's retry it with using tags and usernames | 10:43 |
*** kbaegis has joined #openstack-sahara | 10:56 | |
*** Poornima has quit IRC | 11:11 | |
openstackgerrit | Alina Nesterova proposed openstack/sahara-ci-config: Add -e to shebang https://review.openstack.org/376404 | 11:17 |
openstackgerrit | Alina Nesterova proposed openstack/sahara-ci-config: Add -e to shebang https://review.openstack.org/376404 | 11:19 |
openstackgerrit | Merged openstack/sahara-tests: Use default plugin instead of fake https://review.openstack.org/367385 | 11:20 |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara: Fix API compatibility issue https://review.openstack.org/376414 | 11:34 |
vgridnev | tosky, ^^ | 11:34 |
tosky | uh, thanks | 11:34 |
tosky | vgridnev: does _transform_image_props really need to return a value? Isn't it changing the dictionary that it is passed already? | 11:36 |
tosky | (this question applies also to the current code) | 11:36 |
*** sudswas__ has quit IRC | 11:38 | |
*** sudipto_ has quit IRC | 11:38 | |
vgridnev | tosky, why not ? | 11:38 |
tosky | uhm, I thought that it can be confusing, but it's just an internal method | 11:39 |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara: Fix API compatibility issue https://review.openstack.org/376414 | 11:41 |
*** Poornima has joined #openstack-sahara | 11:42 | |
vgridnev | tosky, so, I think that is final fix ^^ | 11:42 |
tosky | vgridnev: is there a way to remove them from v2 even now, or we can simply not document them in v2 and then at some point remove them? | 11:45 |
vgridnev | tosky, APIv2 is experimental, that means it's not recommended for use, moreover, this version is not enabled by default | 11:45 |
*** esikachev has quit IRC | 11:46 | |
vgridnev | so, we can drop these values at any time | 11:46 |
tosky | vgridnev: yes, just need to drop them from the v2-specific code | 11:46 |
*** sudswas__ has joined #openstack-sahara | 12:06 | |
*** sudipto_ has joined #openstack-sahara | 12:06 | |
openstackgerrit | Kirill Bespalov proposed openstack/sahara: Remove enable_notifications option https://review.openstack.org/368747 | 12:15 |
openstackgerrit | Luigi Toscano proposed openstack/sahara-tests: WIP Check if image registration now works for client test https://review.openstack.org/375312 | 12:19 |
*** zemuvier has quit IRC | 12:28 | |
openstackgerrit | Luigi Toscano proposed openstack/sahara-tests: Fix the network settings of the templates https://review.openstack.org/370997 | 12:37 |
*** kbaegis has quit IRC | 12:41 | |
*** kbaegis has joined #openstack-sahara | 12:42 | |
*** esikachev has joined #openstack-sahara | 12:44 | |
vgridnev | NikitaKonovalov, please review: https://review.openstack.org/#/c/376414/ | 12:49 |
*** tellesnobrega has joined #openstack-sahara | 12:50 | |
*** esikachev has quit IRC | 12:52 | |
*** Poornima has quit IRC | 12:52 | |
*** david-lyle has joined #openstack-sahara | 12:56 | |
*** sudswas__ has quit IRC | 12:59 | |
*** sudipto_ has quit IRC | 13:00 | |
*** I has joined #openstack-sahara | 13:06 | |
*** I is now known as Guest3090 | 13:06 | |
*** shuyingya has quit IRC | 13:08 | |
tosky | vgridnev: about https://review.openstack.org/#/c/376414 : it works fine, I just noticed that status changed from UPPERCASE to lowercase between mitaka and newton | 13:13 |
openstackgerrit | Luigi Toscano proposed openstack/sahara-tests: WIP Check if image registration now works for client test https://review.openstack.org/375312 | 13:15 |
*** esikachev has joined #openstack-sahara | 13:19 | |
*** zemuvier has joined #openstack-sahara | 13:21 | |
vgridnev | tosky, hm, really | 13:26 |
vgridnev | ok, I will change that to uppercase | 13:27 |
openstackgerrit | Alex Schultz proposed openstack/puppet-sahara: Puppet OpenStack Newton RC2 https://review.openstack.org/376500 | 13:33 |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara: Fix API compatibility issue https://review.openstack.org/376414 | 13:36 |
vgridnev | tosky, ^^ | 13:36 |
vgridnev | tellesnobrega, ^^ | 13:37 |
tellesnobrega | vgridnev, looking | 13:37 |
openstackgerrit | Merged openstack/sahara: Fix incorrect event log for ambari https://review.openstack.org/375573 | 13:38 |
*** kbaegis has quit IRC | 13:40 | |
*** brad[] has joined #openstack-sahara | 13:47 | |
openstackgerrit | Luigi Toscano proposed openstack/sahara-tests: WIP Check if image registration now works for client test https://review.openstack.org/375312 | 13:49 |
*** shuyingya has joined #openstack-sahara | 13:51 | |
shuyingya | tosky vgridnev, can you review https://review.openstack.org/#/c/376028/ | 14:02 |
shuyingya | I am not sure I totally understand your meaning? so, Please give me some suggestion | 14:05 |
*** tnovacik has quit IRC | 14:05 | |
openstackgerrit | Evgeny Sikachev proposed openstack/sahara-tests: Fix creation of ng with proxy node https://review.openstack.org/376529 | 14:11 |
tosky | shuyingya: done | 14:25 |
shuyingya | thanks tosky | 14:28 |
tosky | shuyingya: but please check; I'm almost sure that sys.exit should not be used, I'm not totally sure about the proper way to communicate a failure | 14:29 |
tosky | I was just grepping around in osc sources | 14:29 |
openstackgerrit | Alina Nesterova proposed openstack/sahara-tests: Fixed unit tests in sahara scenario https://review.openstack.org/376539 | 14:30 |
shuyingya | aha, raise exception is a better way. thanks | 14:35 |
*** shuyingya has left #openstack-sahara | 14:36 | |
*** shuyingya has joined #openstack-sahara | 14:36 | |
*** nkrinner is now known as nkrinner_afk | 14:49 | |
tosky | mionkin: hi, about https://review.openstack.org/#/c/370997/, do you mean that the brackets are odd because the code could be changed, or something else? | 14:49 |
*** shuyingya has quit IRC | 14:51 | |
mionkin | tosky: hi, i mean that you can remove these brackets, they means nothing | 14:52 |
tosky | mionkin: ah, yeah, it's one of the cases where I prefer to follow the more general rule (brackets are needed if you have two or more arguments) | 14:53 |
tosky | it's like putting { ... } in C code when there is only one statement, less error-prone | 14:54 |
mionkin | tosky, ok, I see, anyway the patch lgtm :) | 14:55 |
tosky | thanks :) | 14:55 |
*** kbaegis has joined #openstack-sahara | 14:57 | |
*** hogepodge has joined #openstack-sahara | 14:58 | |
*** shuyingya has joined #openstack-sahara | 15:01 | |
*** kbaegis has quit IRC | 15:02 | |
*** kbaegis has joined #openstack-sahara | 15:02 | |
*** akuznetsov has joined #openstack-sahara | 15:02 | |
*** zemuvier has quit IRC | 15:05 | |
*** vgridnev has quit IRC | 15:09 | |
Guest3090 | hi, in the old sahara - we could specify the block storage via volumes in the "Node Group Template" - now that flow seems to have changed. What's the best way to attach a cinder block storage to my data nodes? | 15:10 |
Guest3090 | tosky, ^ can you please help with the above query? | 15:10 |
tosky | uhm, our tests still works in the old way; I admit I could not recently test the horizon plugin | 15:11 |
*** esikachev has quit IRC | 15:14 | |
*** zemuvier has joined #openstack-sahara | 15:15 | |
tosky | Guest3090: don't you see " Storage location " in the Sahara UI, under "Node Group Templates"? Or do you mean something else? | 15:15 |
*** esikachev has joined #openstack-sahara | 15:15 | |
Guest3090 | tosky, i do - now thanks!! | 15:17 |
*** Guest3090 has quit IRC | 15:19 | |
*** openstackstatus has joined #openstack-sahara | 15:27 | |
*** ChanServ sets mode: +v openstackstatus | 15:27 | |
*** tnovacik has joined #openstack-sahara | 15:29 | |
*** shuyingya has quit IRC | 15:32 | |
*** zemuvier has quit IRC | 15:33 | |
*** esikachev has quit IRC | 15:34 | |
*** kbaegis has quit IRC | 15:36 | |
*** nkrinner_afk is now known as nkrinner | 15:36 | |
*** kbaegis has joined #openstack-sahara | 15:38 | |
*** vgridnev has joined #openstack-sahara | 15:41 | |
-openstackstatus- NOTICE: Earlier job failures for "zuul-cloner: error: too few arguments" should now be solved, and can safely be rechecked | 15:43 | |
-openstackstatus- NOTICE: Launchpad SSO is offline, preventing login to https://review.openstack.org/, https://wiki.openstack.org/ and many other sites; no ETA has been provided by the LP admin team | 15:48 | |
*** ChanServ changes topic to "Launchpad SSO is offline, preventing login to https://review.openstack.org/, https://wiki.openstack.org/ and many other sites; no ETA has been provided by the LP admin team" | 15:48 | |
*** rcernin has quit IRC | 15:52 | |
*** pgadiya has quit IRC | 15:54 | |
*** kbaegis1 has joined #openstack-sahara | 16:02 | |
*** kbaegis has quit IRC | 16:06 | |
*** I has joined #openstack-sahara | 16:08 | |
*** I has quit IRC | 16:08 | |
*** sudipto has joined #openstack-sahara | 16:09 | |
*** sudipto_ has joined #openstack-sahara | 16:09 | |
*** ChanServ changes topic to "OpenStack Sahara // IRC Meetings - http://eavesdrop.openstack.org/#OpenStack_Data_Processing_(Sahara)_Team_Meeting" | 16:10 | |
-openstackstatus- NOTICE: Launchpad SSO logins are confirmed working correctly again | 16:10 | |
openstackgerrit | Shu Yingya proposed openstack/python-saharaclient: Raise exception in command "plugin configs get" https://review.openstack.org/376028 | 16:12 |
*** david-lyle has quit IRC | 16:13 | |
*** david-lyle has joined #openstack-sahara | 16:14 | |
tosky | vgridnev: a quick note about the tests: it seems that the full tests for both CLI and clients struggles to run on the gates | 16:17 |
tosky | the full test for clients is already disabled (it's the one I was trying to reenable and I hit the API change) | 16:18 |
tosky | the full tests for CLI is currently disabled, but my patch https://review.openstack.org/#/c/370997/ reenabled it, and boom | 16:18 |
tosky | my idea is to merge the above patch (370997), then merge the two remaining patches from zemuvier if they pass (I think their effect should be visible even if the complete tests is failing) | 16:19 |
tosky | then temporarily disable that specific test with a skip+bug, and release sahara-tests | 16:19 |
tosky | it wouldn't be worse than the current package, only better (more tests) | 16:20 |
tosky | then we can investigate how to have the full tests (which creates instances, even if for fake) working | 16:20 |
tosky | what do you think? | 16:20 |
tosky | I will ask esikachev too (probably tomorrow) | 16:20 |
*** zemuvier has joined #openstack-sahara | 16:22 | |
vgridnev | tosky, I will take a look on that in 30-60 mins, a little bit busy now | 16:22 |
tosky | oh, sure | 16:22 |
tosky | just putting on the queue for an opinion, a new sahara-tests release is definitely needed at this point | 16:23 |
*** kbaegis1 has quit IRC | 16:24 | |
vgridnev | tosky, what exact reason of test failure? | 16:25 |
vgridnev | now | 16:25 |
tosky | timeout on cluster creation | 16:25 |
*** kbaegis has joined #openstack-sahara | 16:25 | |
vgridnev | hm | 16:25 |
tosky | even if they are small ubuntu images with 512MiB of RAM, maybe it's too much for the non-native virtualization | 16:26 |
tosky | same issue for the other tests (from client) when https://review.openstack.org/#/c/375312/ is applied | 16:26 |
*** kbaegis has quit IRC | 16:28 | |
*** pcaruana has quit IRC | 16:29 | |
vgridnev | tosky, output of http://logs.openstack.org/97/370997/4/check/gate-sahara-cli-nv/7afc3cd/testr_results.html.gz doesn't seems like timeout issue | 16:30 |
*** kbaegis has joined #openstack-sahara | 16:30 | |
vgridnev | No matches found ?! | 16:30 |
tosky | it is: the image unregister is just the cleanup function called on the timeout, which is executed when the cluster is still "starting" | 16:31 |
tosky | so boom | 16:31 |
tosky | even if the time is really close there, uhm, I though that there was a wait function | 16:33 |
vgridnev | tosky, uh, I saw now that in logs for patchset 3 | 16:33 |
vgridnev | for p4 there is no such logs | 16:33 |
tosky | ah, yes, self.wait_for_resource_deletion is called | 16:34 |
tosky | but still, no luck; I tried to reproduce with an all-in-one installation on a virtual machine and same issue (which makes sense) | 16:34 |
vgridnev | so, why then scenario tests are working well? | 16:35 |
*** abehl has quit IRC | 16:36 | |
tosky | good point | 16:37 |
tosky | longer timeout? | 16:37 |
vgridnev | we are creating 2 vms in this tests (which are scaling up to 3 nodes), and everything works properly | 16:37 |
tosky | uhm uhm | 16:37 |
vgridnev | is there a way to override to timeout? | 16:37 |
tosky | yes, we set it | 16:38 |
tosky | and I thought I already tested it, but not on the gate | 16:38 |
openstackgerrit | Shu Yingya proposed openstack/python-saharaclient: Raise exception in command "plugin configs get" https://review.openstack.org/376028 | 16:38 |
vgridnev | uh, no | 16:38 |
vgridnev | 1 vm -> 2 vms | 16:38 |
vgridnev | http://git.openstack.org/cgit/openstack/sahara-tests/tree/etc/scenario/gate/fake.yaml.mako | 16:38 |
*** nkrinner is now known as nkrinner_afk | 16:39 | |
openstackgerrit | Luigi Toscano proposed openstack/sahara-tests: Fix the network settings of the templates https://review.openstack.org/370997 | 16:45 |
sudipto | tosky, vgridnev - i created a node group with cinder volumes attached - to data nodes, however, when i login to the VM - i only see the ephemeral storage - any clues? | 16:55 |
sudipto | even though the cinder attachment has been successful. | 16:56 |
sudipto | I am using the LVM driver | 16:56 |
tosky | you don't see in the output of blkid? | 16:57 |
sudipto | tosky, i just see the ephemeral disk there... | 16:58 |
sudipto | blkid | 16:58 |
sudipto | /dev/vda2: LABEL="cloudimg-rootfs" UUID="08e88eb0-e298-4e3d-95fc-5147828a871b" TYPE="ext4" | 16:58 |
tosky | if you check with cinder commands (UI or openstack volume...) do you see volumes allocated ? | 16:59 |
sudipto | tosky, yeah | 17:00 |
sudipto | the volumes are allocated and attached. | 17:00 |
tosky | can you please paste the complete output of dmesg from the instance? | 17:01 |
sudipto | sure. | 17:01 |
sudipto | tosky, http://paste.openstack.org/show/582977/ | 17:03 |
tosky | uhm, weird | 17:04 |
tosky | is it the same on all data nodes? | 17:05 |
vgridnev | what plugin you are using? | 17:05 |
sudipto | yeah all of them | 17:05 |
sudipto | this is using the vanilla plugin 2.7.1 | 17:06 |
*** tnovacik has quit IRC | 17:06 | |
*** akuznetsov has quit IRC | 17:06 | |
*** kbaegis has quit IRC | 17:08 | |
*** kbaegis has joined #openstack-sahara | 17:08 | |
*** esikachev has joined #openstack-sahara | 17:09 | |
vgridnev | sudipto, can you post your node group templates? | 17:10 |
sudipto | vgridnev, sure. | 17:11 |
*** kbaegis has quit IRC | 17:14 | |
*** kbaegis has joined #openstack-sahara | 17:14 | |
sudipto | vgridnev, http://paste.openstack.org/show/582978/ | 17:15 |
*** kbaegis has quit IRC | 17:25 | |
vgridnev | hm, can you execute sudo su - -c "hdfs dfsadmin -report" hadoop on master node ? | 17:28 |
vgridnev | sudipto ^^ | 17:28 |
sudipto | vgridnev, sure | 17:28 |
sudipto | vgridnev, http://paste.openstack.org/show/582981/ | 17:31 |
*** kbaegis has joined #openstack-sahara | 17:34 | |
*** tosky has quit IRC | 17:35 | |
*** egafford has joined #openstack-sahara | 17:42 | |
sudipto | vgridnev, any clues? That space seems to be a lot less no? | 17:43 |
*** degorenko is now known as _degorenko|afk | 17:45 | |
*** esikachev has quit IRC | 17:49 | |
vgridnev | so, there no space at all | 17:49 |
*** zemuvier has quit IRC | 17:49 | |
sudipto | you can give me clues to debug (since i am very very new to sahara) | 17:50 |
*** esikachev has joined #openstack-sahara | 17:50 | |
sudipto | vgridnev ^ | 17:52 |
vgridnev | ops, sorry. 1 min | 17:52 |
vgridnev | so, yeah, I have some clues | 17:53 |
vgridnev | 1: you should try to investigate "are a volumes attached at all" . So, just create an instance, attach a volume, check it's attached | 17:54 |
sudipto | vgridnev, even without the volumes, that space looks too less... | 17:54 |
sudipto | my rootfs is 20 Gig | 17:55 |
vgridnev | then you can validate how find instance devices is working: you can see that right there: https://github.com/openstack/sahara/blob/master/sahara/service/volumes.py#L137 | 17:56 |
*** esikachev has quit IRC | 17:56 | |
vgridnev | what version of openstack you are using? | 17:56 |
sudipto | vgridnev, using Mitaka | 17:57 |
vgridnev | uh, ok, i'm actually sure that everything should be working on sahara side, just because I was doing some storage tests on our scale lab (LVM+ICSCI included) probably there is something wrong attaching devices itself from cinder side. | 17:59 |
vgridnev | Uh, one more note | 18:00 |
vgridnev | please check: are you using nova config drive for storing metadata? | 18:00 |
sudipto | vgridnev, and looking at this command: lsblk -r | awk '$6 ~ /disk/ || /part/ {print \"/dev/\" $1}' - it seems that we are taking the disk which is 20Gigs on my system | 18:00 |
vgridnev | if so, you will need to apply https://github.com/openstack/sahara/commit/cf57c1f8601e899b6815156d0509fd9a9a26a399 | 18:01 |
sudipto | vgridnev, i didn't make any changes on the nova side? | 18:01 |
sudipto | how do i find out if config drive is being used to store metadata? | 18:01 |
vgridnev | can you execute sudo blkid ? | 18:02 |
sudipto | vgridnev, /dev/vda2: LABEL="cloudimg-rootfs" UUID="08e88eb0-e298-4e3d-95fc-5147828a871b" TYPE="ext4" | 18:02 |
sudipto | and that vda2 device is 20 GB | 18:02 |
sudipto | that's my roots | 18:03 |
vgridnev | so, what about ? mount | awk '$1 ~ /^\/dev/ {print $1}' ? | 18:04 |
sudipto | /dev/vda2 | 18:04 |
sudipto | /dev/vda1 | 18:04 |
vgridnev | Ok, that means, that sahara can't use that because it's mounted | 18:05 |
sudipto | vgridnev, oh k - didn't quite understand that... | 18:05 |
vgridnev | https://github.com/openstack/sahara/blob/master/sahara/service/volumes.py#L171 | 18:05 |
sudipto | as in why wouldn't it take the mounted space? | 18:06 |
vgridnev | so, sahara is taking only attached devices, then we are formatting devices, and only then we are mounting that. So, we are not using already mounted devices because it's guaranteed that it's really can be used for storage | 18:08 |
sudipto | vgridnev, ah ok - got it. So basically i have to ensure that the cinder volumes are attached | 18:09 |
sudipto | vgridnev, btw - is block device driver a recommended cinder volume driver for SAHARA or is it LVM? | 18:10 |
vgridnev | so, I would say yes, block device driver is much better from performance point, but that is probably harder to deploy. additionally, you should be sure that all volumes are located locally to VM | 18:12 |
sudipto | ok so even if i use iSCSI - that should be via a local cinder volume service right? | 18:13 |
sudipto | I will try giving BDD a shot tomorrow, hopefully the complexity you refer here is on the part of cinder config - and not in SAHARA? | 18:13 |
*** esikachev has joined #openstack-sahara | 18:14 | |
vgridnev | from sahara side you should use 'volume_local_to_instance=true' for each node group then | 18:15 |
sudipto | vgridnev, this config parameter is when i use iscsi right? | 18:19 |
sudipto | in such a case, the cinder volume has to run on each compute host for lvm right? | 18:19 |
vgridnev | no, this config is in node groups templates | 18:21 |
sudipto | yeah i mean this will be in effect if i use iscsi right? | 18:24 |
*** atana has quit IRC | 18:29 | |
*** zemuvier has joined #openstack-sahara | 18:30 | |
vgridnev | you should not use iscsi for block device driver. In this case you have to just specify available drives in cinder configs (if I remember correctly). You have to enable additional filters to make volume locality working correctly | 18:30 |
*** thumpba has joined #openstack-sahara | 18:30 | |
vgridnev | http://docs.openstack.org/developer/sahara/userdoc/advanced.configuration.guide.html#volume-instance-locality-configuration | 18:31 |
vgridnev | how many disks do you have available? | 18:31 |
*** atana has joined #openstack-sahara | 18:33 | |
openstackgerrit | Merged openstack/sahara: Fix API compatibility issue https://review.openstack.org/376414 | 18:48 |
*** sudipto has quit IRC | 18:57 | |
*** sudipto_ has quit IRC | 18:57 | |
*** zemuvier has quit IRC | 18:58 | |
*** esikachev has quit IRC | 19:24 | |
*** jamielennox|away has quit IRC | 19:26 | |
*** jamielennox|away has joined #openstack-sahara | 19:29 | |
*** jamielennox|away is now known as jamielennox | 19:30 | |
*** esikachev has joined #openstack-sahara | 19:32 | |
*** kbaegis has quit IRC | 19:41 | |
*** kbaegis has joined #openstack-sahara | 19:43 | |
openstackgerrit | Vitaly Gridnev proposed openstack/sahara-dashboard: just testcommit https://review.openstack.org/377063 | 19:52 |
*** vgridnev has quit IRC | 20:28 | |
*** egafford has quit IRC | 20:34 | |
*** egafford has joined #openstack-sahara | 20:34 | |
*** zemuvier has joined #openstack-sahara | 20:42 | |
*** zemuvier has quit IRC | 20:44 | |
*** egafford has quit IRC | 20:47 | |
*** esikachev has quit IRC | 20:52 | |
*** catintheroof has joined #openstack-sahara | 21:15 | |
*** nimisha has joined #openstack-sahara | 21:27 | |
*** nimisha has quit IRC | 21:35 | |
*** esikachev has joined #openstack-sahara | 21:48 | |
*** esikachev has quit IRC | 21:52 | |
*** kbaegis has quit IRC | 21:57 | |
*** thumpba has quit IRC | 22:40 | |
*** thumpba has joined #openstack-sahara | 22:40 | |
*** thumpba has quit IRC | 22:42 | |
*** esikachev has joined #openstack-sahara | 22:48 | |
*** esikachev has quit IRC | 22:53 | |
*** catintheroof has quit IRC | 23:04 | |
*** esikachev has joined #openstack-sahara | 23:49 | |
*** esikachev has quit IRC | 23:54 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!