| *** yingjun has joined #openstack-searchlight | 00:52 | |
| *** openstackgerrit has quit IRC | 01:22 | |
| *** openstackgerrit has joined #openstack-searchlight | 01:22 | |
| *** lakshmiS has joined #openstack-searchlight | 03:19 | |
| *** lei-zh has quit IRC | 03:33 | |
| *** lei-zh has joined #openstack-searchlight | 03:36 | |
| *** lakshmiS has quit IRC | 03:45 | |
| *** lei-zh has quit IRC | 03:50 | |
| *** lei-zh has joined #openstack-searchlight | 03:50 | |
| *** lei-zh has quit IRC | 04:01 | |
| *** lei-zh has joined #openstack-searchlight | 04:05 | |
| *** yingjun has quit IRC | 05:30 | |
| *** yingjun has joined #openstack-searchlight | 05:30 | |
| *** yingjun has quit IRC | 05:35 | |
| *** lakshmiS has joined #openstack-searchlight | 05:39 | |
| *** gb21_ has joined #openstack-searchlight | 05:42 | |
| *** yingjun has joined #openstack-searchlight | 05:58 | |
| *** yingjun has quit IRC | 09:37 | |
| *** gb21_ has quit IRC | 12:11 | |
| *** lakshmiS has quit IRC | 13:26 | |
| *** bpokorny has joined #openstack-searchlight | 15:57 | |
| *** bpokorny has quit IRC | 16:02 | |
| *** gb21_ has joined #openstack-searchlight | 16:06 | |
| *** tsufiev has quit IRC | 16:11 | |
| *** pkarikh has quit IRC | 16:12 | |
| *** TravT has quit IRC | 16:22 | |
| *** TravT has joined #openstack-searchlight | 16:25 | |
| *** tsufiev has joined #openstack-searchlight | 16:27 | |
| *** pkarikh has joined #openstack-searchlight | 16:28 | |
| *** gb21_ has quit IRC | 17:51 | |
| *** sjmc7 has quit IRC | 20:32 | |
| *** sjmc7 has joined #openstack-searchlight | 20:35 | |
| TravT | sjmc7: are you around? | 20:55 |
|---|---|---|
| sjmc7 | TravT yep | 20:58 |
| sjmc7 | wassup? | 20:58 |
| TravT | do you have latest devstack with searchlight running by chance? | 20:58 |
| TravT | my devstack is possibly corrupted | 20:58 |
| TravT | i think | 20:59 |
| TravT | but when I run searchlight-manage index sync, i'm getting an error | 20:59 |
| TravT | wanted to see if you saw the same thing | 20:59 |
| sjmc7 | sure, one sec | 21:07 |
| sjmc7 | ah… which elasticsearch client did you get? | 21:09 |
| sjmc7 | there’s an open bug about that | 21:09 |
| sjmc7 | ah, and it looks like designateclient has changed | 21:10 |
| sjmc7 | so yes, i am getting it | 21:10 |
| sjmc7 | i’ll look at them now - you want to file a bug for the designate one or should i? we may just need to cap versions in requirements | 21:10 |
| sjmc7 | https://bugs.launchpad.net/searchlight/+bug/1515412 for elasticsearch | 21:12 |
| openstack | Launchpad bug 1515412 in OpenStack Search (Searchlight) "python elasticsearch must match server version" [High,Triaged] - Assigned to Liyingjun (liyingjun) | 21:12 |
| sjmc7 | although since global-requirements doesn’t cap them, i don’t think we can | 21:12 |
| sjmc7 | curses | 21:13 |
| TravT | oh no, that's a bit of an issue | 21:16 |
| TravT | you can go ahead and file the designate bug | 21:16 |
| sjmc7 | yeah.. i’m not sure about the elasticsearch client. if global-reqs doesn’t cap it, they force everyone to upgrade | 21:16 |
| TravT | which error are you seeing? | 21:17 |
| sjmc7 | for elasticsearch - Failed to setup index extension os_glance_image: 'IndicesClient' object has no attribute 'delete_mapping' | 21:18 |
| sjmc7 | for designate: Failed to setup index extension os_designate_zone: An auth plugin is required to determine endpoint URL | 21:18 |
| sjmc7 | elasticsearch==2.1.0 | 21:18 |
| TravT | ok, yeah, that is the error i'm seeing | 21:18 |
| sjmc7 | global-reqs has >=1.3.0 | 21:18 |
| TravT | i have designate disabled at the moment, so not getting that second error | 21:19 |
| sjmc7 | at least it fails rather than going on with the rest :) | 21:19 |
| TravT | found a number of other projects having this issue... | 21:20 |
| TravT | even mongo | 21:20 |
| TravT | https://github.com/mongodb-labs/mongo-connector/issues/348 | 21:20 |
| TravT | they fixed it by capping version... | 21:20 |
| sjmc7 | yep | 21:20 |
| sjmc7 | we can propose a cap to global-reqs | 21:20 |
| TravT | probably should google around to see if there is some workaround | 21:21 |
| sjmc7 | there isn’t - the clients aren’t compatible | 21:21 |
| TravT | well, that's fun | 21:21 |
| sjmc7 | https://elasticsearch-py.readthedocs.org/en/master/ | 21:21 |
| sjmc7 | it’s a really thin wrapper on e-s | 21:21 |
| sjmc7 | aside from the pooling, it doesn’t add a great deal | 21:21 |
| TravT | surely ceilometer has hit this as well | 21:22 |
| sjmc7 | it’s not in their reqs | 21:27 |
| TravT | i'm not seeing anything jump out about the client lib API actually being different between versions. | 21:29 |
| TravT | so is this purely a deployment configuration issue | 21:29 |
| TravT | ? | 21:29 |
| sjmc7 | ?how do you mean? | 21:30 |
| TravT | i guess i jumped ahead in my thoughts of what all this means to us... let me explain | 21:31 |
| TravT | was thinking in context of supporting multiple versions of ES | 21:31 |
| TravT | if we'd have to have different code paths depending on the deployed version of ES | 21:31 |
| sjmc7 | shudder | 21:32 |
| TravT | or if we are protected a bit by their client API | 21:32 |
| sjmc7 | we can mandate elasticsearch 2.0 | 21:32 |
| sjmc7 | or install a diferent version of the client depending | 21:32 |
| TravT | if ^, then we have to get devstack to install that instead of the 1.x it is now | 21:32 |
| sjmc7 | yeah, and if not, we need to get global-reqs to cap the client version | 21:33 |
| TravT | hmmm, this seems almost mailing list worthy | 21:34 |
| sjmc7 | yeah, i think so | 21:34 |
| sjmc7 | like i say, i knew this was going to happen; it didn’t occur we couldn’t just cap it ourselves | 21:34 |
| TravT | deployers will often already have a version of ES they support. | 21:35 |
| sjmc7 | right | 21:35 |
| TravT | so ideally, this is just a deployment configuration | 21:35 |
| sjmc7 | yeah.. but we have to pick a default | 21:35 |
| sjmc7 | which will be 1.x until devstack updates | 21:35 |
| TravT | yep. | 21:36 |
| sjmc7 | so until then we need to cap global-reqs | 21:36 |
| TravT | yep, i'd think so. | 21:36 |
| sjmc7 | but that depends on whoever else is using it | 21:36 |
| TravT | i'd say it would be worthwhile to put up the patch, but then put up ML on it. | 21:36 |
| sjmc7 | yeah - do you know who maintains that repo? | 21:37 |
| TravT | no, i'd think perhaps infra, but i don't know | 21:38 |
| sjmc7 | ok | 21:38 |
| TravT | david-lyle: do you know who "owns" devstack? | 21:38 |
| TravT | there is a cross project meeting happening. | 21:38 |
| david-lyle | TravT: owns? | 21:39 |
| david-lyle | qa | 21:39 |
| TravT | we have a requirements issue that affects devstack. | 21:40 |
| TravT | and possibly other projects | 21:40 |
| TravT | sjmc7: this HAS to affect others | 21:40 |
| sjmc7 | yeah, i’ll see who added it in the first place | 21:40 |
| sjmc7 | ‘gordon chung’ familiar? | 21:42 |
| sjmc7 | ceilometer, looks like | 21:42 |
| sjmc7 | i’ll go ask in there | 21:42 |
| david-lyle | ceilo PTL | 21:42 |
| sjmc7 | ta | 21:44 |
| sjmc7 | looks like there are [qa] and [infra] tags. i’ll put the patch up and send something to those plus us and ceilo | 21:48 |
| sjmc7 | david-lyle, i know there’s an automatic job to update requirements.txt - do you know how that works? | 21:49 |
| david-lyle | sjmc7: on commit to opentack/requirements then a script fires it off | 21:51 |
| sjmc7 | ok, it’s that simple, cool | 21:51 |
| TravT | does that script look at all projects? | 21:53 |
| TravT | or are some hard coded in it? | 21:53 |
| TravT | maybe looks at the projects yaml? | 21:53 |
| sjmc7 | pretty sure it looks at projects.yaml | 21:56 |
| TravT | thanks for posting that message steve | 22:01 |
| sjmc7 | yep. i’ll have a look at the designate thing in the meantime | 22:03 |
| sjmc7 | hmm.. designate error seems to have been transient | 22:21 |
| sjmc7 | so that’s nice | 22:21 |
| TravT | guess you can't complain about that. | 22:22 |
| sjmc7 | ah, except i don’t have designate running on this VM | 22:22 |
| sjmc7 | pebkac! | 22:22 |
| TravT | that's usually the case with you. | 22:24 |
| TravT | ;P | 22:24 |
| sjmc7 | usually? | 22:26 |
| sjmc7 | i think always | 22:26 |
| sjmc7 | nope, still transient. that’s good at least. with elasticsearch<2.0 things work as normal | 22:26 |
| *** TravT has quit IRC | 23:10 | |
| *** lei-zh has quit IRC | 23:11 | |
| *** TravT has joined #openstack-searchlight | 23:12 | |
| *** lei-zh has joined #openstack-searchlight | 23:12 | |
| *** TravT_ has joined #openstack-searchlight | 23:12 | |
| ekarlso | how are things going TravT ? | 23:15 |
| ekarlso | or sjmc7 :P | 23:15 |
| *** TravT has quit IRC | 23:16 | |
| *** david-lyle has quit IRC | 23:25 | |
| *** TravT_ is now known as TravT | 23:27 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!