| shayneburgess | the gate is just screwed right now | 00:04 |
|---|---|---|
| vkmc | :/ | 00:04 |
| *** saurabhs has left #openstack-trove | 00:11 | |
| cp16net | vkmc: nice | 00:12 |
| *** saurabhs has joined #openstack-trove | 00:15 | |
| *** saurabhs has left #openstack-trove | 00:17 | |
| cp16net | since we see the instance goto error there must be some other issue now | 00:19 |
| *** sharikapongubala has joined #openstack-trove | 00:25 | |
| openstackgerrit | Masaki Matsushita proposed openstack/trove-integration: Simplify sed statement https://review.openstack.org/216982 | 00:29 |
| amrith | saurabhs, sorry. wasn't looking at this screen. | 00:43 |
| *** IanGovett has quit IRC | 00:45 | |
| sharikapongubala | amrith , you have applied the cli changes too when trying to run the api right? | 00:48 |
| amrith | sharika, yes | 00:48 |
| amrith | sharikapongubala, yes | 00:48 |
| amrith | am trying again now | 00:48 |
| amrith | cp16net, I just updated etherpad with progress. | 00:48 |
| amrith | looks like we're moving forward. | 00:48 |
| sharikapongubala | ok if theres an error let me know | 00:48 |
| amrith | vkmc, waiting for your change (mariadb). | 00:49 |
| cp16net | amrith: thanks | 00:49 |
| amrith | sharikapongubala, no error, I just can't see any code path that would actually invoke some of the cluster code. | 00:49 |
| amrith | and I don't see what cluster support in mysql means at this stage. | 00:49 |
| sharikapongubala | root-show <cluster-id> with a vertica cluster will work | 00:49 |
| sharikapongubala | mysql doesnt support cluster root-enable as far as i know | 00:50 |
| amrith | sharikapongubala, that's my point exactly. hence my bewilderment at the code. | 00:50 |
| amrith | anyway, I'll try to understand it some more. | 00:50 |
| amrith | have to switch monitors. | 00:50 |
| vkmc | amrith, already up for review, PS10 | 00:50 |
| amrith | bbiab | 00:50 |
| cp16net | mvandijk: can you update this one. https://review.openstack.org/#/c/205355/ | 00:50 |
| amrith | vkmc, will cycle back to it shortly. half way through this root-show/root-enable thing. | 00:50 |
| vkmc | amrith, cool, thx | 00:51 |
| amrith | vkmc, if you have a second would you also take a look at https://review.openstack.org/#/c/196913/29/trove/extensions/mysql/models.py,cm. Maybe you have already looked at it and understand it better. will be back on irc in about 15m'ish. | 00:51 |
| vkmc | sure thing | 00:53 |
| cp16net | vgnbkr: how goes the updates to redis cluster? https://review.openstack.org/#/c/203702/ | 00:54 |
| *** mingdang1 has joined #openstack-trove | 01:14 | |
| SlickNik | wow | 01:14 |
| SlickNik | hi folks | 01:15 |
| *** erkules has joined #openstack-trove | 01:19 | |
| *** erkules_ has quit IRC | 01:22 | |
| *** saurabhs has joined #openstack-trove | 01:27 | |
| cp16net | i just found the fun probalem... | 01:36 |
| cp16net | cant have spaces apparently in the name of an instance | 01:36 |
| cp16net | BadRequest: Invalid input for field/attribute name. Value: test here. u'test here' does not match '^[a-zA-Z0-9-._]*$' (HTTP 400) (Request-ID: req-fb2e8a3e-20c7-456c-a2a0-33a6852467b1) | 01:36 |
| amrith | SlickNik, hello! | 01:46 |
| amrith | are you in zipcode country or pincode country? | 01:46 |
| openstackgerrit | Craig Vyvial proposed openstack/trove: Instance name can not have a space in it now. https://review.openstack.org/219950 | 01:48 |
| cp16net | so it looks like that is the solution | 01:49 |
| vkmc | cp16net, that's in the gate? | 01:50 |
| SlickNik | cp16net: I saw that too | 01:50 |
| SlickNik | it can't have a : in it either | 01:50 |
| vkmc | it can't have special chars at all | 01:50 |
| vkmc | right? | 01:50 |
| SlickNik | so you might have to rework that patch. | 01:50 |
| vkmc | ïnst|nc* is not a valid name | 01:51 |
| *** imandhan has quit IRC | 01:51 | |
| cp16net | ah your right... | 01:51 |
| SlickNik | amrith: I'm in no country right now | 01:52 |
| SlickNik | up in the air | 01:52 |
| SlickNik | cp16net: possibly replace that with epoch time? | 01:52 |
| SlickNik | amrith: On my way from pincode land to zipcode land. | 01:53 |
| vkmc | oh wow | 01:54 |
| vkmc | reviewing from a plane | 01:54 |
| vkmc | if this is not FF... | 01:54 |
| cp16net | does the time in the name matter? | 01:54 |
| SlickNik | No, i suspect it's just to unique'ify | 01:55 |
| SlickNik | and give you qn idea of which instance was created when. | 01:55 |
| cp16net | yeah.. so i was just thinking uuid | 01:55 |
| cp16net | TEST_+uuid.uuid4() | 01:55 |
| SlickNik | gopher it | 01:56 |
| SlickNik | btw, I can't pinpoint which nova change recently merged to cause this. | 01:57 |
| openstackgerrit | Craig Vyvial proposed openstack/trove: Test instance name can not have special charactrers in it now. https://review.openstack.org/219950 | 01:57 |
| cp16net | there thats better | 01:58 |
| cp16net | here it is... | 02:02 |
| cp16net | https://github.com/openstack/python-novaclient/commit/72190ba9a28353f1176cdc20f2ea2c688ee697e7 | 02:02 |
| cp16net | its in the client | 02:02 |
| cp16net | novaclient* | 02:02 |
| cp16net | doh.. nevermind thats not it | 02:03 |
| SlickNik | no — it's coming from the server | 02:03 |
| SlickNik | funny part is the server schema validation (on n-api) looks like it should allow for spaces. | 02:04 |
| SlickNik | https://github.com/openstack/nova/blob/master/nova/api/validation/parameter_types.py#L43-L55 | 02:04 |
| SlickNik | wow, almost looks like it's trying to validate against the hostname regexinstead for some reason. | 02:06 |
| SlickNik | https://github.com/openstack/nova/blob/master/nova/api/validation/parameter_types.py#L43-L55 | 02:06 |
| SlickNik | whoops https://github.com/openstack/nova/blob/master/nova/api/validation/parameter_types.py#L84-L92 | 02:07 |
| cp16net | oh doesnt the name get translated to hostname | 02:08 |
| saurabhs | amrith do you want to talk here about that root-enable patch? | 02:15 |
| cp16net | i bet thats what it is | 02:15 |
| SlickNik | yeah, but not sure how it worked before in that case. | 02:15 |
| saurabhs | cp16net do you have a min to guide me on being able to setup a mysql cluster, we need to test the root-enable on cluster patch with mysql clusters | 02:18 |
| SlickNik | Just spoke to the nova guys — apparently it's a side effect of https://github.com/openstack/nova/commit/4a18f7d3bafcdbede48500aac389e0a770b8e6a8 | 02:18 |
| cp16net | saurabhs: sure | 02:19 |
| cp16net | SlickNik: thats interesting | 02:20 |
| saurabhs | cp16net so in order to build the mysql image for clustering all I do is run ./redstack kick-start pxc. do I need to do anything else? | 02:20 |
| cp16net | that should be sufficient if you have the code | 02:20 |
| *** ashleighfarnham has quit IRC | 02:21 | |
| cp16net | SlickNik: so now we need both of those changes in | 02:21 |
| cp16net | or the gate will never pass | 02:21 |
| cp16net | but you cant make a circular dependency between them | 02:22 |
| saurabhs | and this is all the code I need correct? https://review.openstack.org/#/c/208073/ | 02:22 |
| cp16net | that and have the latest trove-integration | 02:22 |
| cp16net | yup | 02:23 |
| SlickNik | cp16net: not circular - 3 layers deep | 02:23 |
| amrith | saurabhs, yes | 02:23 |
| amrith | the issue is that when pxc lands | 02:23 |
| amrith | the root controller for that should be something other than MySQL Root Controller | 02:24 |
| amrith | you need a PXC Root Controller. | 02:24 |
| amrith | That can have the code you are showing in https://review.openstack.org/#/c/196913/29/trove/extensions/mysql/models.py,cm | 02:24 |
| amrith | however, the code in mysql/models.py should (I believe) not have anything to do with clusters. | 02:24 |
| *** csoukup has joined #openstack-trove | 02:24 | |
| amrith | after all, pxc has its own config section in cfg.py | 02:24 |
| amrith | saurabhs ^^ | 02:25 |
| amrith | the code for MySQL should (I believe) error if the cluster parameter is passed in. | 02:25 |
| *** saurabhs1 has joined #openstack-trove | 02:25 | |
| amrith | same for percona though I think it should eventually get its own (or we need the refactor to give it a way). These OO languages have something called inheritence ;) | 02:26 |
| amrith | there I said the dirty "I" word. | 02:26 |
| amrith | saurabhs ^^ | 02:26 |
| *** saurabhs has quit IRC | 02:26 | |
| amrith | bbiab. switching screens; not a good day to have a monitor fail. | 02:26 |
| cp16net | booo that sucks amrith | 02:27 |
| amrith | cp16net what sucks? the fact that my monitor broke. I'm getting it replaced (warranty). It is a HP w2408h. | 02:29 |
| amrith | nice monitor | 02:29 |
| amrith | great colors | 02:29 |
| amrith | I have two of them | 02:30 |
| amrith | one of them went on the fritz | 02:30 |
| amrith | or are you saying that inheritence sucks? | 02:30 |
| cp16net | losing a monitor sucks | 02:30 |
| amrith | not if you get a fat check in the mail for an aunt you never liked anyway. | 02:30 |
| cp16net | and that sucks | 02:30 |
| cp16net | lol | 02:30 |
| amrith | not that I have any like that | 02:30 |
| amrith | they all love me and have no money. | 02:30 |
| amrith | that sucks too ;) | 02:30 |
| amrith | back to linux screen for a bit. bbiab | 02:30 |
| saurabhs1 | amrith I agree with your concerns and confusion. but as I said before in the patch comment. the Root class is currently used by all the datastore that wants to implment root functionality. And its true for other classes in that file, like Users or RootHistory. | 02:34 |
| saurabhs1 | and we don't wanted to refactor the whole thing just for root-enable functionality on clusters | 02:34 |
| saurabhs1 | today we are using inheritance by implementing a MySQLRootController and VerticaRootController and that is datastore specific thing. | 02:35 |
| *** haomaiwang has quit IRC | 02:36 | |
| saurabhs1 | when pxc gets merged, it can have its own pxcRootController when root-enable functionality is supported by it | 02:37 |
| amrith | saurabhs1, I don't get it. MySQL has NO cluster capability. not now, not after PXC merges. | 02:38 |
| saurabhs1 | amrith if you are asking to do the refactor to move the Root classes out of mysql models then its a big change and needs lot of testing | 02:38 |
| saurabhs1 | amrith that is correct | 02:38 |
| saurabhs1 | but it does have root-enable fuctionality | 02:38 |
| amrith | and your cfg.py only sets this class for mysql and percona | 02:38 |
| *** haomaiwang has joined #openstack-trove | 02:38 | |
| saurabhs1 | and the Root class works for instances as well as clusters | 02:38 |
| amrith | and for that, sure, accept a parameter. the implementation is bogus. | 02:39 |
| saurabhs1 | amrith cft.py also sets it for vertica | 02:39 |
| amrith | saurabhs1, not to this class | 02:39 |
| amrith | for vertica it sets it to VerticaRootController. | 02:39 |
| amrith | the ONLY thing I'm saying is that you can receive cluster_instances_list. But, in MySQL root controller, if it is set, throw an exception | 02:40 |
| openstackgerrit | Sharika Pongubala proposed openstack/trove: Root enablement for Vertica clusters/instances https://review.openstack.org/196913 | 02:40 |
| amrith | the implementation makes no sense there. | 02:40 |
| saurabhs1 | trove.extensions.mysql.service.MySQLRootController is in mysql/service.py | 02:40 |
| amrith | yes, so? | 02:40 |
| *** sabeen has joined #openstack-trove | 02:41 | |
| amrith | saurabhs1, here is my proposed code in MySQL root controller | 02:42 |
| amrith | https://gist.github.com/amrith/26254f6883011ef3d06c | 02:42 |
| saurabhs1 | amrith we both are on same page, but on different side I guess :). can you look at this patch : https://review.openstack.org/#/c/196913/, when root-enabled is called for mysql cluster it raises an exception in service.py | 02:42 |
| amrith | sorry, the indentation is messed up in gist | 02:43 |
| amrith | all I'm saying is that the code in MySQL root controller to do anything with the cluster_instance_list parameter is meaningless. | 02:43 |
| amrith | When cp16net's code merges, you get to create a PXC Root Controller. | 02:43 |
| amrith | that's it! | 02:43 |
| amrith | bbiab | 02:45 |
| amrith | switching screens | 02:45 |
| saurabhs1 | so from your gist, again this method gets called for vertica as well as mysql. so thats why we don't want to raise exception here. | 02:46 |
| saurabhs1 | instead if you look at new patch https://review.openstack.org/#/c/196913/ , its raising appropriate exception in service.py in MySQLRootController. https://review.openstack.org/#/c/196913/30/trove/extensions/mysql/service.py | 02:46 |
| saurabhs1 | that will take care of cp16net's pxc datastore and will say that operation is not supported | 02:47 |
| cp16net | alright so i have a plan of attack on this gate issue now... | 02:52 |
| amrith | saurabhs1, I don't believe that this is a good solution. | 02:53 |
| amrith | in service.py (vertica) you are importing all the MySQL extensions. | 02:53 |
| amrith | that defeats the whole purpose of having datastore specific extensions. | 02:54 |
| amrith | in doing this, you are tying the hands of anyone who in the future wants to do MySQL specific stuff in the MySQL extension! | 02:54 |
| amrith | because that would (potentially) break Vertica. | 02:55 |
| saurabhs1 | amrith other datastores already use users extensions from mysql, don't they | 02:56 |
| saurabhs1 | ? | 02:57 |
| *** haomaiwang has quit IRC | 02:57 | |
| amrith | we're trying like hell not to add new dependencies like this. | 02:57 |
| amrith | I don't know whether they do or not | 02:57 |
| amrith | specifically to your question though. | 02:57 |
| SlickNik | cp16net: Are you gonna create a chain of patches turning the int-tests off in the first one, putting in the needed fixes, and then re-enabling them? | 03:00 |
| *** haomaiwa_ has joined #openstack-trove | 03:00 | |
| cp16net | thats the plan | 03:00 |
| amrith | saurabhs1, I just checked. I don't see any other instances of a datastore using mysql extensions. | 03:00 |
| SlickNik | Yup, I don't see any other way out. | 03:00 |
| amrith | I did a quick check | 03:00 |
| amrith | so I may be wrong | 03:00 |
| amrith | I'll look again | 03:00 |
| saurabhs1 | amrith: any datastore that implements users functionality will use mysql models | 03:01 |
| saurabhs1 | or users api to be very specific | 03:01 |
| cp16net | i think the plan is to 1. exit 0 out of redstack at the top. 2. fix the test info name we use. 3. reenable redstack. 4. get image to build 5. other patches merged | 03:01 |
| cp16net | SlickNik: sounds good? | 03:02 |
| SlickNik | I would re-order 3 and 4. | 03:06 |
| SlickNik | cp16net: ^ | 03:06 |
| SlickNik | to save us some time. | 03:06 |
| openstackgerrit | Craig Vyvial proposed openstack/trove-integration: skip everything with the gate to fix issues https://review.openstack.org/219967 | 03:06 |
| cp16net | SlickNik: ok | 03:07 |
| cp16net | i was thinking about that as well | 03:07 |
| cp16net | the only thing is that if the image merges then the other depends-on patches could merge without tests run | 03:07 |
| cp16net | dunno if thats a terrible thing but it could bite us | 03:08 |
| cp16net | i guess they havnt passed as is tho | 03:08 |
| cp16net | so we'd have to recheck them anyway | 03:08 |
| SlickNik | No the other depends on patches won't merge. | 03:08 |
| cp16net | yea ok | 03:09 |
| openstackgerrit | Craig Vyvial proposed openstack/trove: Test instance name can not have special charactrers in it now. https://review.openstack.org/219950 | 03:09 |
| openstackgerrit | Craig Vyvial proposed openstack/trove-integration: Fix ubuntu image builds due to revert in tarball names https://review.openstack.org/219548 | 03:10 |
| *** bhunter71 has quit IRC | 03:11 | |
| openstackgerrit | Craig Vyvial proposed openstack/trove-integration: Revert "skip everything with the gate to fix issues" https://review.openstack.org/219972 | 03:12 |
| cp16net | umm double check those 4 changes now... | 03:12 |
| cp16net | the order i submitted them should be the order they are in. | 03:13 |
| amrith | saurabhs1, I'm not comfortable with this change. I've looked at it a bunch and it may be worthwhile someone else review it. This module has a number of things including User, UserAccess, Root, with MySQL based implementaions. making this dependency on vertica strikes me as highly unadvisable. | 03:13 |
| cp16net | i think it would be easier to start with the last one since they link in the depends-on change id | 03:13 |
| cp16net | looks good to me | 03:14 |
| cp16net | SlickNik: mind looking those over? | 03:16 |
| SlickNik | looking now. | 03:16 |
| cp16net | i work flowed everything but that first one in the chain | 03:16 |
| *** csoukup has quit IRC | 03:16 | |
| SlickNik | cp16net: FYI https://bugs.launchpad.net/nova/+bug/1491511 | 03:17 |
| openstack | Launchpad bug 1491511 in OpenStack Compute (nova) "Behavior change with latest nova paste config" [Undecided,New] | 03:17 |
| cp16net | ok | 03:18 |
| cp16net | cool | 03:18 |
| SlickNik | LGTM | 03:19 |
| cp16net | awesome | 03:19 |
| cp16net | so once all those are merged then we recheck the patches | 03:19 |
| cp16net | eta... forever... | 03:19 |
| cp16net | lol | 03:19 |
| *** esmute_ has quit IRC | 03:26 | |
| amrith | SlickNik, https://bugs.launchpad.net/nova/+bug/1491511 ... that's crazy! | 03:26 |
| openstack | Launchpad bug 1491511 in OpenStack Compute (nova) "Behavior change with latest nova paste config" [Undecided,Confirmed] | 03:26 |
| amrith | I'm going to wander away and get some dinner. be back shortly. | 03:27 |
| amrith | crap, is that really the time? | 03:27 |
| cp16net | same here | 03:27 |
| *** eglute has quit IRC | 03:27 | |
| amrith | it is way past my bedtime | 03:27 |
| cp16net | i'm going to look at something other than my computer for a bit... | 03:27 |
| amrith | I'm going to go eat and be back. | 03:28 |
| amrith | looks like a long night ahead. | 03:28 |
| amrith | how's your merge situation going; I get to start up on vkmc's changes next. | 03:28 |
| cp16net | amrith: i think we have a resolution to the madness | 03:28 |
| SlickNik | amrith: cp16net and I have a plan. | 03:29 |
| amrith | vkmc ... your changes 211791 failed check again | 03:29 |
| cp16net | not an ideal one but its already in motion | 03:29 |
| amrith | good deal, would your plan need to be applied to 211791? | 03:29 |
| amrith | https://review.openstack.org/#/c/211791/ | 03:29 |
| cp16net | yeah i'm sure they will | 03:29 |
| * SlickNik crosses his fingers | 03:29 | |
| cp16net | i'm surprised much of anything passes right now | 03:29 |
| * amrith uncrosses his eyes | 03:29 | |
| cp16net | acutally shouldnt juna pass tho? | 03:30 |
| cp16net | i'd expect master to fail but kilo and juno should be pinned | 03:30 |
| amrith | juno failed on a timeout | 03:30 |
| amrith | in 1h54m | 03:30 |
| cp16net | ahh damn it timed out | 03:31 |
| amrith | at 2h | 03:31 |
| *** sharikap_ has joined #openstack-trove | 03:31 | |
| cp16net | omg this is fun | 03:31 |
| amrith | I thought only instance delete was allowed 2h | 03:31 |
| *** eglute has joined #openstack-trove | 03:32 | |
| *** sharikapongubala has quit IRC | 03:32 | |
| amrith | see https://review.openstack.org/#/c/217881/ | 03:32 |
| amrith | oh, never mind | 03:32 |
| amrith | wonder who gives a 2h timeout | 03:32 |
| amrith | not anything I know of | 03:32 |
| amrith | that is 120s | 03:32 |
| amrith | not 120m | 03:32 |
| amrith | I'm too cross-eyed | 03:32 |
| amrith | going to go get food. | 03:32 |
| amrith | SlickNik, you have the wheel. put the sucker on auto-pilot, stay steady at 35k feet. | 03:33 |
| amrith | 37k if you get bumps | 03:33 |
| amrith | keep the fasten seat belts sign on | 03:33 |
| cp16net | lol | 03:33 |
| amrith | have the crew pass out extra grog. | 03:33 |
| amrith | toodle doo for now | 03:33 |
| cp16net | gate-trove-functional-dsvm-mysql SUCCESS in 22m 55s | 03:34 |
| cp16net | thats amazing... | 03:35 |
| cp16net | :-P | 03:35 |
| SlickNik | cp16net: trove integration is branchless | 03:36 |
| SlickNik | so juno and kilo should pass. | 03:36 |
| cp16net | yeah thats what i was thinking | 03:37 |
| cp16net | SlickNik: so i guess we sit and wait for the those to get through now | 03:37 |
| SlickNik | Yup | 03:37 |
| SlickNik | hey the gate is down to double digits | 03:38 |
| SlickNik | :P | 03:38 |
| cp16net | yeah but it will take all "night" to get through i bet | 03:38 |
| SlickNik | btw saurabhs1 and amrith — was reading scrollback and your conversation. | 03:39 |
| SlickNik | methinks it's probably wise to put the the common root models (used by mysql and vertica) in a common module and have them both import it from there. | 03:40 |
| SlickNik | Rather than having vertica import the mysql one. | 03:40 |
| saurabhs1 | yeh started working on it will post a patch in couple of hours | 03:41 |
| SlickNik | thanks saurabhs1 | 03:44 |
| SlickNik | cp16net: are you going to update your pxc patch to be dependent on the last one in that series? :) | 03:45 |
| SlickNik | Yeah probably will | 03:45 |
| SlickNik | btw, Are there any more changes that need to get looked at? | 03:45 |
| SlickNik | (as opposed to ones that just need to merge now) | 03:45 |
| *** links has joined #openstack-trove | 04:02 | |
| *** sabeen has quit IRC | 04:02 | |
| *** esmute_ has joined #openstack-trove | 04:08 | |
| *** esmute_ has quit IRC | 04:25 | |
| cp16net | oh yeah that would be a good idea SlickNik | 04:36 |
| cp16net | i just took a break and gave my stinky dogs a bath | 04:37 |
| *** sharikap_ has quit IRC | 04:41 | |
| *** sharikapongubala has joined #openstack-trove | 04:44 | |
| openstackgerrit | Craig Vyvial proposed openstack/trove: Percona Xtradb Cluster implementation https://review.openstack.org/208073 | 04:50 |
| cp16net | oh crap looks like theres more than one place where we use names with time stamps... | 04:52 |
| openstackgerrit | Craig Vyvial proposed openstack/trove: Test instance name can not have special charactrers in it now. https://review.openstack.org/219950 | 04:56 |
| cp16net | SlickNik: yt? | 04:56 |
| cp16net | i did a search all through the test to find all the place we use time stamps for instance names | 05:02 |
| cp16net | replaced them all now | 05:02 |
| *** esmute_ has joined #openstack-trove | 05:26 | |
| *** sgotliv has joined #openstack-trove | 06:30 | |
| *** lexholden has joined #openstack-trove | 07:55 | |
| *** links has quit IRC | 08:02 | |
| *** sharikapongubala has quit IRC | 08:04 | |
| *** _bhi_ has joined #openstack-trove | 08:30 | |
| *** mingdang1 has quit IRC | 08:57 | |
| *** tosky has joined #openstack-trove | 09:34 | |
| *** sharikapongubala has joined #openstack-trove | 09:47 | |
| *** sharikapongubala has quit IRC | 09:52 | |
| *** IanGovett has joined #openstack-trove | 10:02 | |
| *** kbyrne has quit IRC | 10:12 | |
| *** kbyrne has joined #openstack-trove | 10:15 | |
| *** links has joined #openstack-trove | 10:31 | |
| amrith | cp16net, SlickNik looks like the masterful change (exit 0) may have passed gate! | 11:04 |
| amrith | saurabhs1, don't see any new code on the root enablement. | 11:05 |
| saurabhs1 | amrith: yeh still working through the changes, did lil more refactor and just testing everything now. | 11:09 |
| amrith | the gate seems to be a smouldering mountain of horse manure right now. | 11:10 |
| *** lexholden has quit IRC | 11:28 | |
| openstackgerrit | Saurabh Surana proposed openstack/trove: Root enablement for Vertica clusters/instances https://review.openstack.org/196913 | 11:28 |
| SlickNik | cp16net: Just got back | 11:36 |
| SlickNik | apparently there are no satellite connections when you're flying over the North pole. | 11:37 |
| SlickNik | cp16net: Thanks for looking into the other places with the name issue. | 11:38 |
| SlickNik | ls | 11:38 |
| SlickNik | I'm planning on making a small change to your revert fix as well — temporarily disable stable branch support in trove-int. We'll re-enable it once lib-3 is cut. | 11:39 |
| openstackgerrit | Saurabh Surana proposed openstack/trove: Root enablement for Vertica clusters/instances https://review.openstack.org/196913 | 11:42 |
| saurabhs1 | amrith: https://review.openstack.org/#/c/196913/31 is now refactored to address earlier comments. I think I only went back on one of the old comments on that patch where it was asked to set default rootController to None. I realized that before this patch everything was anyways going through the old root controller, so if we let everything go through the default implementation we won't change the bahavior of other databases. Jus | 11:49 |
| saurabhs1 | I will be online in couple of hours again. | 11:51 |
| *** lexholden has joined #openstack-trove | 11:57 | |
| openstackgerrit | Nikhil Manchanda proposed openstack/trove: Test instance name can not have special charactrers in it now. https://review.openstack.org/219950 | 12:02 |
| SlickNik | cp16net: updated https://review.openstack.org/#/c/219950/ to fix a couple of pep8 issues | 12:26 |
| *** arist has quit IRC | 12:29 | |
| *** arist has joined #openstack-trove | 12:30 | |
| *** amrith is now known as _amrith_ | 12:50 | |
| *** pmackinn has joined #openstack-trove | 13:05 | |
| *** itlinux has joined #openstack-trove | 13:16 | |
| *** mingdang1 has joined #openstack-trove | 13:27 | |
| *** robcresswell has joined #openstack-trove | 13:37 | |
| *** links has quit IRC | 13:40 | |
| *** _crobertsrh is now known as crobertsrh | 13:43 | |
| *** _amrith_ is now known as amrith | 13:48 | |
| *** sharikapongubala has joined #openstack-trove | 13:50 | |
| *** sharikapongubala has quit IRC | 13:54 | |
| *** atomic77 has left #openstack-trove | 14:20 | |
| *** atomic77 has joined #openstack-trove | 14:20 | |
| *** devlaps has quit IRC | 14:33 | |
| openstackgerrit | Merged openstack/trove-integration: skip everything with the gate to fix issues https://review.openstack.org/219967 | 14:33 |
| *** mingdang1 has quit IRC | 14:34 | |
| pmackinn | cp16net, lol, not playing around anymore | 14:35 |
| *** itlinux has quit IRC | 14:43 | |
| *** esmute_ has quit IRC | 14:51 | |
| cp16net | nope | 14:54 |
| pmackinn | scorched earth | 14:54 |
| pmackinn | cp16net, so trove-integration changes should wait? or merge that to slide in for now? or this is is just to unblock py-trove, trove? | 14:55 |
| cp16net | pmackinn: you talking about the changes i made? | 14:57 |
| pmackinn | cp16net, your unblocker change | 14:57 |
| cp16net | there is a chain of patches i created | 14:58 |
| cp16net | to fix the gate | 14:58 |
| amrith | what a mess (the gate is) | 15:03 |
| amrith | saurabhs1, I'm running your new code | 15:04 |
| *** csoukup has joined #openstack-trove | 15:07 | |
| cp16net | amrith: i agree | 15:08 |
| cp16net | hmm anyone seen this error? http://logs.openstack.org/72/219972/1/check/gate-trove-functional-dsvm-mysql/78fdb96/logs/screen-tr-tmgr.txt.gz#_2015-09-03_13_17_03_550 | 15:12 |
| cp16net | i notcied it last night when testing mariadb but never saw it in the gate till now | 15:13 |
| saurabhs1 | amrith: Just woke up. let me know if you need anything | 15:14 |
| amrith | saurabhs1, I think we both need coffee ;) | 15:15 |
| cp16net | i just made some myself | 15:17 |
| *** rjaiswal has joined #openstack-trove | 15:34 | |
| *** amrith is now known as _amrith_ | 15:40 | |
| *** sabeen has joined #openstack-trove | 15:47 | |
| openstackgerrit | Morgan Jones proposed openstack/trove: Redis Cluster Initial Implementation https://review.openstack.org/203702 | 15:47 |
| *** esmute_ has joined #openstack-trove | 15:52 | |
| *** mattgriffin has joined #openstack-trove | 15:54 | |
| *** esmute_ has quit IRC | 15:56 | |
| *** mattgriffin has quit IRC | 16:00 | |
| *** mattgriffin has joined #openstack-trove | 16:00 | |
| cp16net | http://logs.openstack.org/72/219972/1/check/gate-trove-functional-dsvm-mysql/78fdb96/logs/screen-tr-tmgr.txt.gz#_2015-09-03_13_17_03_550 | 16:03 |
| *** saurabhs1 has quit IRC | 16:27 | |
| *** _bhi_ has quit IRC | 16:34 | |
| openstackgerrit | Matthew Van Dijk proposed openstack/trove: Mongodb Cluster Scaling https://review.openstack.org/205355 | 16:37 |
| *** ashleighfarnham has joined #openstack-trove | 16:39 | |
| *** sharikapongubala has joined #openstack-trove | 16:58 | |
| *** sharikapongubala has quit IRC | 16:59 | |
| *** saurabhs has joined #openstack-trove | 17:00 | |
| *** imandhan has joined #openstack-trove | 17:01 | |
| *** esmute_ has joined #openstack-trove | 17:01 | |
| *** haomaiw__ has joined #openstack-trove | 17:02 | |
| *** haomaiwa_ has quit IRC | 17:03 | |
| *** esmute_ has quit IRC | 17:11 | |
| SlickNik | hi folks | 17:11 |
| cp16net | hey SlickNik | 17:12 |
| SlickNik | cp16net: looks like the first of our changes merged. | 17:13 |
| cp16net | yup | 17:14 |
| cp16net | the 2 patches are in the gate | 17:14 |
| cp16net | but i saw some weird issues when booting a nova instance now | 17:14 |
| cp16net | look at one of the log links i posted | 17:15 |
| *** _amrith_ is now known as amrith | 17:16 | |
| mvandijk | cp16net, i updated the commit msg as you suggested and it wiped my reviews | 17:19 |
| cp16net | ok | 17:20 |
| SlickNik | cp16net: I think I know what's going on — the latest version of nova-client isn't on pypi and that's causing the issue. | 17:22 |
| amrith | cp16net, you have a +2 to add to https://review.openstack.org/#/c/205355/15 ... I put one on it. | 17:22 |
| *** esmute_ has joined #openstack-trove | 17:22 | |
| amrith | I reviewed the previous patch set; the only diff in 15 is the commit message. | 17:22 |
| amrith | so I think it is good to go. | 17:23 |
| mvandijk | you both +2 no wf+1 though | 17:24 |
| cp16net | heh yeah | 17:24 |
| cp16net | done | 17:24 |
| *** esmute_ has quit IRC | 17:27 | |
| *** sharikapongubala has joined #openstack-trove | 17:27 | |
| mvandijk | cheers dudes | 17:27 |
| cp16net | word | 17:27 |
| cp16net | lol the check passes so fast now... | 17:29 |
| SlickNik | cp16net: So I suspect that this functional job isn't going to work correctly unless nova releases a new client that supports the new 2.1 API. | 17:30 |
| cp16net | ah | 17:30 |
| cp16net | SlickNik: one way to make them all pass is the way we have the gate setup now :-p | 17:32 |
| SlickNik | cp16net: I vote that we get our Liberty-3 patches merged while the job is still a no-op. And then bring back the job once we have the right client. | 17:32 |
| SlickNik | yup. | 17:32 |
| cp16net | heh | 17:32 |
| SlickNik | We've already run these anyway. | 17:32 |
| cp16net | true | 17:32 |
| cp16net | should we workflow -1 the test revert? | 17:32 |
| *** esmute_ has joined #openstack-trove | 17:33 | |
| cp16net | so it doesnt merge? | 17:33 |
| cp16net | nothing depends on that patch that i know of | 17:33 |
| *** sabeen has quit IRC | 17:33 | |
| cp16net | this will mean that its likely that we will have an unstable l-3 | 17:33 |
| cp16net | which is ok for now i suppose | 17:34 |
| *** esmute_ has quit IRC | 17:34 | |
| SlickNik | I don't think it will anyway since the tests won't pass. | 17:34 |
| cp16net | oh true... | 17:34 |
| cp16net | :-P | 17:34 |
| cp16net | unless its updated before it gets run | 17:35 |
| edmondk | Trying to get updated so is the gate-trove-functional-dsvm-mysql gate a noop? | 17:35 |
| edmondk | which is why its only 22 minutes | 17:35 |
| cp16net | edmondk: yeah | 17:35 |
| cp16net | i "fixed" it :-P | 17:35 |
| edmondk | lol | 17:35 |
| cp16net | exit 0 as the first line of redstack | 17:35 |
| edmondk | peterstac, https://review.openstack.org/#/c/185892/ this patch has a unit test failing | 17:36 |
| *** saurabhs1 has joined #openstack-trove | 17:36 | |
| edmondk | Redis backup/restore ^ | 17:36 |
| cp16net | SlickNik: i marked this as wf-1 for time being https://review.openstack.org/#/c/219972/ | 17:37 |
| SlickNik | cp16net: sounds good. | 17:38 |
| peterstac | edmondk, yeah, it's failing on MethodTest (get_flavors) - which shouldn't be affected by my code | 17:40 |
| peterstac | I've seen that before, usually I just recheck until it goes away :) | 17:41 |
| SlickNik | FYI -2'ed https://review.openstack.org/#/c/214782/ and https://review.openstack.org/#/c/210094/ for now — let's wait until lib-3 BPs are in before merging other fixes. | 17:42 |
| edmondk | peterstac, ok I cherry picked your change down and running tox to just confirm | 17:44 |
| peterstac | it passed yesterday on the same patchset, so I'm pretty sure the code's fine | 17:45 |
| edmondk | peterstac, We should probably do something with that test if it is flaky | 17:45 |
| edmondk | peterstac, bad to have unit tests we cannot rely on | 17:45 |
| peterstac | edmondk, Yeah, I agree | 17:46 |
| saurabhs1 | amrith did you had luck testing root-enable patch | 17:46 |
| saurabhs1 | I tested it with mysql, pxc and vertica | 17:46 |
| saurabhs1 | trying to build mariadb image and will test it with that next | 17:48 |
| saurabhs1 | I also ran tox and int tests with it | 17:48 |
| peterstac | edmondk, py27 just passed in the check pipeline, so hopefully we're good | 17:50 |
| edmondk | k | 17:50 |
| SlickNik | cp16net / peterstac: Does someone need to look into https://review.openstack.org/#/c/203702/? | 17:51 |
| edmondk | yes we need to review it | 17:51 |
| amrith | saurabhs1, got waylaid by another crisis. will get back to it now. | 17:51 |
| peterstac | SlickNik, yeah, I'm looking at that (plus the pxc cluster - try to do both at the same time) | 17:52 |
| amrith | edmondk, we have highly non-deterministic tests ... https://review.openstack.org/#/c/219544/ passes 100% of the time on my desktop machines, and fails 100% of the time on gate. | 17:52 |
| peterstac | amrith, yeah, had a quick convo with edmondk ^^^ we should disable/fix tests that are flakey | 17:53 |
| edmondk | we should be using patch.start() with LOG for all of these | 17:53 |
| edmondk | in the setup | 17:53 |
| edmondk | I recently fixed some unit tests that had this same problem | 17:53 |
| amrith | edmondk, let's talk more about this | 17:53 |
| edmondk | and noticed all of the missed mocks of LOG | 17:53 |
| amrith | I DON'T like my solution. | 17:53 |
| amrith | it imposes a huge burden on the future | 17:53 |
| amrith | to make sure that we don't reintroduce code paths that throw errors. | 17:54 |
| peterstac | The dangling-mock detector project got sidelined - unfortunately now we could have used it ... | 17:54 |
| edmondk | check out what I did on this one, specifically the patcher_log https://review.openstack.org/#/c/215402/3/trove/tests/unittests/guestagent/test_dbaas.py,cm | 17:54 |
| amrith | for Mikata or Miata or Myopia or whatever we're calling it today. | 17:54 |
| SlickNik | Mitaka | 17:55 |
| edmondk | with the patcher start and stop you don't have to put @patch on every test | 17:55 |
| amrith | edmondk, that's not a safe patch | 17:55 |
| amrith | safe fix | 17:55 |
| edmondk | ? | 17:55 |
| amrith | and your fix may be the thing that is throwing me off. | 17:55 |
| amrith | here's why | 17:55 |
| amrith | we run the tests in (I believe) a parallel framework | 17:55 |
| amrith | so in a single python process, you could have >1 test running at the same time. | 17:55 |
| amrith | this is why things like dangling mocks produced non-deterministic results. | 17:56 |
| amrith | If you do a patch like this, and some other code path is checking (as I do) for the number of times a call like mock_logging.exception.call_count is 3, I could be off | 17:56 |
| amrith | because some other code path comes along and drops a mock on top of mine. | 17:56 |
| amrith | am I making sense or do we need to upgrade from 1200 baud? | 17:56 |
| edmondk | I get what you are saying | 17:57 |
| amrith | I went the route I did because I wanted to narrow the time for which I held a mock | 17:57 |
| amrith | to be exactly one test long | 17:57 |
| edmondk | patcher.start is exact one test long | 17:57 |
| edmondk | setup runs for each test | 17:57 |
| edmondk | and tears down | 17:57 |
| edmondk | on each test | 17:57 |
| peterstac | edmondk +1 | 17:57 |
| edmondk | exactly the same way yours is doing | 17:57 |
| peterstac | well, except that it does it for *each* test | 17:58 |
| edmondk | yes because each test should be using it because it uses LOG on each test | 17:58 |
| edmondk | and we don't ever want to use a real logger in a unit test | 17:58 |
| edmondk | or else you get the no handler error | 17:58 |
| *** sabeen1 has joined #openstack-trove | 18:00 | |
| edmondk | I also think adding call counts for log exceptions is making our unit tests highly tied to an implementation | 18:01 |
| edmondk | because logs aren't logic and can be removed or added | 18:02 |
| amrith | edmondk, it does it for each test, irrespective of whether there is a potential for logging or not | 18:02 |
| amrith | therefore your patch ends up leaving mock enabled, or enables and disables mocks more often | 18:02 |
| amrith | my desktop and laptop are 4core | 18:02 |
| edmondk | amrith, yes that is true it does it for each test | 18:02 |
| amrith | the test is 8 core | 18:02 |
| amrith | and therefore with more concurrency, I'm concluding comes more pain. | 18:02 |
| amrith | I'm going to retest without your change | 18:03 |
| amrith | and see if it reduces the concurrency issues | 18:03 |
| amrith | my solution is not perfect | 18:03 |
| amrith | it is PAINFUL | 18:03 |
| amrith | we need a better way out ... | 18:03 |
| edmondk | your solution will have the same problem because we are just putting patch on each test versus setuop | 18:03 |
| amrith | cp16net, thanks for the WF+1 on the patch from above. | 18:03 |
| edmondk | the race condition still exists the same | 18:03 |
| *** sabeen3 has joined #openstack-trove | 18:03 | |
| cp16net | amrith: i've noticed something similar where unittests would fail on my machine and pass in the gate | 18:03 |
| edmondk | I just ran the unit tests on my machine | 18:04 |
| edmondk | and they passed | 18:04 |
| amrith | cp16net, that's a problem with your machine ;) | 18:04 |
| edmondk | they just take awhile | 18:04 |
| cp16net | the error i found was with the database locked | 18:04 |
| amrith | ouch | 18:04 |
| *** sabeen1 has quit IRC | 18:04 | |
| cp16net | never seen it in the gate tho | 18:04 |
| amrith | I think a carefully placed "exit 0" could help ;) | 18:04 |
| cp16net | hehe :-P | 18:04 |
| amrith | remember to chalk that one up under "drammatically improved Trove development team productivity and reduced workload on CI infrastructure for each patch set" in your performance review/resume ;) | 18:05 |
| edmondk | amrith, just ran the test with peter's patch and you will see it hang with the handlers then eventually pass http://paste.openstack.org/show/444547/ | 18:05 |
| edmondk | took 750 seconds though which seems off | 18:06 |
| amrith | this is a puny machine, 2 core | 18:06 |
| edmondk | a vm | 18:06 |
| amrith | I can't find a way to print more diagnostics from py27 | 18:07 |
| amrith | like which test is running | 18:07 |
| edmondk | I will try removing my patcher_log and see if it makes a difference | 18:07 |
| edmondk | I don't see that making a difference for runtime though | 18:08 |
| edmondk | the proboscis fake tests seemed to be hanging | 18:08 |
| edmondk | not the unit tests | 18:08 |
| *** sharikapongubala has quit IRC | 18:12 | |
| edmondk | amrith's new patch is failing on py27 with the database locking problem http://logs.openstack.org/44/219544/2/check/gate-trove-python27/8cc99cc/console.html | 18:13 |
| edmondk | cp16net, looks like the locking problem can happen on the gate | 18:13 |
| cp16net | omg... | 18:14 |
| cp16net | yup thats the same issue i saw | 18:14 |
| amrith | edmondk, you've seen this problem before? | 18:15 |
| amrith | are seeing it now? | 18:15 |
| edmondk | I have never seen this issue | 18:15 |
| edmondk | just was pointing it out to craig because he said he has never seen the gate have it | 18:16 |
| *** sharikapongubala has joined #openstack-trove | 18:16 | |
| edmondk | Are we ever thinking of stopping the fake proboscis test to run in py27 ? | 18:16 |
| edmondk | they are odd | 18:16 |
| edmondk | to say the least | 18:16 |
| edmondk | and only run the actual unit tests | 18:17 |
| amrith | so, there's a typo in https://review.openstack.org/#/c/203702/6//COMMIT_MSG,cm (vgnbkr). I'm thinking of -1'ing it ;) | 18:17 |
| SlickNik | edmondk / cp16net: FWIW I've seen it in the gate before. | 18:17 |
| cp16net | i was thinking it could be from the util.init_db() | 18:18 |
| edmondk | amrith, taking out patcher_log made no difference unit tests still all pass and proboscis tests still take a long time | 18:18 |
| cp16net | if the db gets reset when something needs to read it | 18:18 |
| amrith | edmondk, I'm not seeing an issue with timing | 18:18 |
| amrith | My tests run in about 200 to 300s | 18:18 |
| edmondk | The true fix is injecting a new logger into the init of the classes we use versus defining a global LOG at the top of every file | 18:19 |
| edmondk | this would remove the parallel problems because every class we test would have a unique mocked logger | 18:19 |
| *** lexholden has quit IRC | 18:21 | |
| edmondk | the quick fix is make the unit tests run in serial | 18:22 |
| edmondk | the gate isn't failing with these issues though? | 18:22 |
| cp16net | edmondk: i dont recal ever seeing an issue making things fail | 18:23 |
| openstackgerrit | amrith proposed openstack/trove: Move ignore_dbs and ignore_users out of DEFAULT https://review.openstack.org/207833 | 18:26 |
| amrith | cp16net, what's the scoop with https://review.openstack.org/#/c/211791/? Does it need more "Depends-On" lines to get it through the process? | 18:29 |
| cp16net | i just gave it a +2 | 18:30 |
| cp16net | wf+1 | 18:31 |
| cp16net | looking at the depends... | 18:31 |
| cp16net | should be fine | 18:31 |
| cp16net | just needs to get through | 18:31 |
| cp16net | man sometimes you just have to close all the tabs to sort through things... | 18:32 |
| amrith | cp16net, I just had to do that | 18:33 |
| amrith | my browser went to hell | 18:33 |
| amrith | turns out I was running a window with zuul | 18:33 |
| amrith | which eats your cpu for lunch and spits it out whole. | 18:33 |
| amrith | what a piece of ... | 18:33 |
| cp16net | yeah no kidding LOTS of javascript going on in there | 18:34 |
| amrith | cp16net, edmondk would you take a look at https://review.openstack.org/#/c/203701/ please | 18:35 |
| amrith | I'm looking at that now,. | 18:35 |
| openstackgerrit | amrith proposed openstack/python-troveclient: Redis Clustering Initial Implementation https://review.openstack.org/203701 | 18:36 |
| cp16net | ok | 18:36 |
| amrith | I put the depends on that you'd mentioned so it wiped all earlier reviews. but it was all +1's only. | 18:37 |
| amrith | vgnbkr, is the split in troveclient/v1/shell.py (review https://review.openstack.org/#/c/203701/7/troveclient/v1/shell.py,cm) around line 271 correct? | 18:39 |
| edmondk | that looks like the same split we have for cluster create | 18:40 |
| cp16net | so i guess grow/shrink return no data? | 18:41 |
| cp16net | i guess just a 202 accepted? | 18:41 |
| edmondk | makes sense because it is just an action | 18:41 |
| cp16net | yah | 18:42 |
| amrith | cp16net, edmondk I just spoke with vgnbkr | 18:44 |
| amrith | he is going to fix the metavar and help | 18:44 |
| edmondk | ok | 18:44 |
| amrith | (or the code) | 18:46 |
| amrith | one part of that is wrong | 18:46 |
| edmondk | k I will await that patch and +2 when it comes | 18:47 |
| amrith | edmondk, you are correct; the apischema indicates name, flavorRef, volume, nics and az | 18:47 |
| amrith | so it is name, volume, flavor (at least) | 18:47 |
| amrith | flavor is the mandatory field | 18:47 |
| *** sharikapongubala has quit IRC | 18:48 | |
| amrith | #link https://review.openstack.org/#/c/203702/6/trove/common/apischema.py,cm | 18:48 |
| openstackgerrit | Morgan Jones proposed openstack/python-troveclient: Redis Clustering Initial Implementation https://review.openstack.org/203701 | 18:49 |
| *** sharikapongubala has joined #openstack-trove | 18:55 | |
| vkmc | why this change is depending on Vertica cluster impl? https://review.openstack.org/#/c/208073/ | 18:57 |
| vkmc | cp16net, ^ | 19:01 |
| cp16net | vkmc: no its the opposite | 19:01 |
| amrith | i don't think it i | 19:01 |
| amrith | it is ... | 19:02 |
| amrith | this change depends on the super-duper change ;) | 19:02 |
| cp16net | the veritic depends on it | 19:02 |
| vkmc | oh oh that one depends on https://review.openstack.org/#/c/219972/ | 19:02 |
| vkmc | ok ok | 19:02 |
| cp16net | but i screwed up on the depends-on: in that review... | 19:02 |
| cp16net | crap | 19:02 |
| amrith | that explains why it is failing | 19:02 |
| amrith | cp16net .... | 19:02 |
| openstackgerrit | Craig Vyvial proposed openstack/trove: Percona Xtradb Cluster implementation https://review.openstack.org/208073 | 19:02 |
| amrith | thanks vkmc, you spotted a problem! | 19:02 |
| cp16net | yup | 19:03 |
| cp16net | yeah i think i was going cross eyed last night | 19:03 |
| * amrith mutters, the OTHER "exit 0" change | 19:03 | |
| vkmc | np, ideas come when you stare at zuul | 19:03 |
| cp16net | i just revmoved the depends-on | 19:03 |
| openstackgerrit | Saurabh Surana proposed openstack/trove: Add support for MariaDB datastore in Trove https://review.openstack.org/217451 | 19:03 |
| openstackgerrit | Saurabh Surana proposed openstack/trove: Added root_controller for mariadb datastore https://review.openstack.org/220280 | 19:03 |
| cp16net | that should get through since the gate is "working" right now | 19:04 |
| amrith | this is what I look like right now ... http://learningspanishlikecrazy.com/wp-content/uploads/2015/01/Cross_eyed.jpg | 19:04 |
| cp16net | i'm pretty sure that was me last night | 19:04 |
| cp16net | lol | 19:04 |
| * amrith looks up working in m-w.com | 19:04 | |
| amrith | working: "the manner of functioning or operating" | 19:05 |
| amrith | that sounds about right | 19:05 |
| amrith | http://www.merriam-webster.com/dictionary/working | 19:05 |
| vkmc | loool | 19:05 |
| amrith | somewhere down in that page. | 19:05 |
| amrith | 3: good enough to allow work or further work to be done | 19:05 |
| amrith | YES, the gate is "working" | 19:06 |
| openstackgerrit | Saurabh Surana proposed openstack/trove: Add support for MariaDB datastore in Trove https://review.openstack.org/217451 | 19:06 |
| openstackgerrit | Saurabh Surana proposed openstack/trove: Added root_controller for mariadb datastore https://review.openstack.org/220280 | 19:06 |
| amrith | peterstac, vkmc, cp16net, https://review.openstack.org/#/c/203701/8 is ready for you to look at | 19:07 |
| amrith | I'll let one of you +1 it | 19:07 |
| *** sharikapongubala has quit IRC | 19:08 | |
| *** sharikapongubala has joined #openstack-trove | 19:08 | |
| openstackgerrit | Saurabh Surana proposed openstack/trove: Added root_controller for mariadb datastore https://review.openstack.org/220280 | 19:08 |
| amrith | peterstac, take a look at https://review.openstack.org/#/c/217451/ | 19:08 |
| amrith | not sure waht saurabhs1 just did | 19:08 |
| amrith | but he pushed ps 6 without losing reviews. | 19:09 |
| vkmc | peterstac, hi! still hitting issues with mariadb? | 19:09 |
| amrith | vkmc, what's up with https://review.openstack.org/#/c/217451/? | 19:09 |
| vkmc | amrith, not sure what error is hitting peterstac, I managed to spin up some instances | 19:10 |
| saurabhs1 | vkmc amrith peterstac don't get hartburn. I was trying to push a mariadb patch so that root-enable works on mariadb after both mariadb and https://review.openstack.org/#/c/196913/ patch lands | 19:10 |
| vkmc | I'll try again just in case | 19:10 |
| amrith | saurabhs1, I don't grok | 19:11 |
| saurabhs1 | accedently added root-enable depedancy for mariadb patch | 19:11 |
| peterstac | vkmc, still running the tests, but having some challenges on my larger instance (not specifically with mariadb though) | 19:11 |
| saurabhs1 | but its fixed now | 19:11 |
| vkmc | peterstac, it should with an small flavor, like mysql | 19:11 |
| vkmc | peterstac, I think you can anonymously ssh to it and see what is the guest doing | 19:11 |
| peterstac | vkmc, amrith: I was getting timeouts in the guestagent - couldn't stop mariadb in 30s | 19:11 |
| amrith | cp16net, do you want to -2 https://review.openstack.org/#/c/219972/ ;) | 19:13 |
| amrith | there one person (who has received a loose fingers award in the past) who could accidentally WF+1 that ;) | 19:13 |
| amrith | SlickNik, peterstac, vkmc, edmondk ... I think we're about done for now. Have to let the machine do its thing now. yes? | 19:15 |
| peterstac | I'm still trying to get a bit of testing done, but otherwise yes | 19:16 |
| cp16net | amrith: yeah i think so... ill scope over the reviews once more to make sure | 19:16 |
| vkmc | I think so yes | 19:17 |
| cp16net | i think pxc needs +2 again | 19:17 |
| vkmc | I'm cheering the patches so they make it through the gate... but I don't think that helps much | 19:18 |
| *** sharikapongubala has quit IRC | 19:18 | |
| amrith | cp16net, done | 19:18 |
| cp16net | thx | 19:18 |
| saurabhs | amrith: any luck testing root-enable patch | 19:20 |
| amrith | saurabhs, I knew I'd forgotten something. | 19:20 |
| amrith | answer: no. | 19:20 |
| saurabhs | ? | 19:20 |
| cp16net | peterstac: how goes? https://review.openstack.org/#/c/217451/ | 19:20 |
| amrith | but I'll do it now. | 19:20 |
| cp16net | said you were testing it | 19:20 |
| cp16net | i coudlnt get it to work because of the novaclient | 19:20 |
| vkmc | what's wrong with novaclient? | 19:21 |
| dhellmann | SlickNik: hey, we need to get trove tagged today, how're things looking? | 19:21 |
| saurabhs | :) I am gonna go grab lunch while you do that. brb | 19:21 |
| amrith | saurabhs, before you do | 19:21 |
| amrith | what about 220280 ? | 19:21 |
| cp16net | dhellmann: we have everything in the gate | 19:21 |
| amrith | https://review.openstack.org/#/c/220280/ | 19:21 |
| peterstac | cp16net, I couldn't get an instance to go active, but I'm bowing out to others who were able to (might be environmental in my case) | 19:21 |
| dhellmann | cp16net: at this point, how likely is it for anything to land in the next few hours? | 19:21 |
| cp16net | i think SlickNik might be afk | 19:21 |
| amrith | cp16net, one more maybe ...https://review.openstack.org/#/c/196913/ | 19:22 |
| cp16net | dhellmann: next few? unlikely with the gate time it looks like | 19:22 |
| dhellmann | cp16net: that's what I thought. Are those related to bugs, or would they need FFEs? | 19:23 |
| peterstac | dhellmann: Things current take > 11 hours to make it through in the best case | 19:23 |
| cp16net | dhellmann: we are pushing for all the features and holding off on bugs | 19:23 |
| cp16net | if we got a build cut now it would mean over half of the features we were pushing for would be missing | 19:24 |
| dhellmann | cp16net: ok. we're going to need to make a call soon based on what is actually merged in order to stick with the deadlines we've set. The team can decide on FFEs for anything you want in RC1, which is still several weeks away. (this is just the L3 tag, we're not branching) | 19:24 |
| cp16net | dhellmann: i understand i'll make sure SlickNik knows | 19:25 |
| dhellmann | that's a lot of features landing late | 19:25 |
| cp16net | yeah the gate has been broken for us for one reason of another for the last week | 19:25 |
| peterstac | dhellmann: We've been fighting gate issues for well over a week now | 19:26 |
| cp16net | when the features were complete | 19:26 |
| peterstac | most of this stuff should have landed last week | 19:26 |
| dhellmann | peterstac: ack | 19:26 |
| openstackgerrit | amrith proposed openstack/trove: Revert change set 217881 https://review.openstack.org/220288 | 19:27 |
| amrith | cp16net, I've pushed up https://review.openstack.org/#/c/220288/1 with WF=-2 | 19:28 |
| amrith | peterstac, vkmc, edmondk, SlickNik ^^ | 19:28 |
| saurabhs | amrith: you need 220280 if you want to test too enable patch against mariadb | 19:29 |
| cp16net | amrith: nice | 19:32 |
| *** imandhan has quit IRC | 19:33 | |
| *** mattgriffin has quit IRC | 19:37 | |
| *** mattgriffin has joined #openstack-trove | 19:38 | |
| cp16net | amrith: looks like all the patches are updated | 19:50 |
| cp16net | i guess SlickNik with have to have a chat tho given the time its taking in the gate. | 19:51 |
| cp16net | amrith: did you work out this one? https://review.openstack.org/#/c/196913/ | 19:52 |
| cp16net | i think thats the only outstanding patch from the priority list | 19:57 |
| amrith | cp16net, I am close on that one | 20:19 |
| amrith | I'll look at why jenkins doesn't like it | 20:19 |
| amrith | if it is benign I'm inclined to push it over. | 20:20 |
| amrith | we can fix nits later. | 20:20 |
| amrith | have to hit the road now so I don't get stuff in gridlock | 20:20 |
| amrith | back online in 45m | 20:20 |
| cp16net | ok | 20:27 |
| *** IanGovett has quit IRC | 20:29 | |
| *** amrith is now known as _amrith_ | 20:29 | |
| *** imandhan has joined #openstack-trove | 20:30 | |
| openstackgerrit | Petr Malik proposed openstack/python-troveclient: Use more appropriate exceptions for validation https://review.openstack.org/210004 | 20:37 |
| *** sharikapongubala has joined #openstack-trove | 20:40 | |
| openstackgerrit | Petr Malik proposed openstack/python-troveclient: Use more appropriate exceptions for validation https://review.openstack.org/210004 | 20:41 |
| saurabhs1 | cp16net what is the correct dependson gerrit id to get through the gate, I am kind of lost. https://review.openstack.org/#/c/196913/ is failing consistently I think I need some of your fixes for the gate. | 20:42 |
| saurabhs1 | give me the or gertit id and I will update it on the correct patches | 20:42 |
| cp16net | i think its the novaclient but let me know | 20:44 |
| cp16net | s/know/look | 20:44 |
| saurabhs1 | thanks | 20:44 |
| cp16net | http://logs.openstack.org/13/196913/32/check/gate-trove-functional-dsvm-mysql/7ba3c28/logs/screen-tr-tmgr.txt.gz#_2015-09-03_20_13_05_330 | 20:44 |
| cp16net | yup | 20:44 |
| saurabhs1 | which is the patch which fixes this issues? I will make all my patches dependent on that one | 20:45 |
| saurabhs1 | cp16net ^ | 20:45 |
| saurabhs1 | also I think its ok to have more than 1 DependsOn entry in the commit message? is that correct? | 20:45 |
| cp16net | yeah you can but you cant have circular depends between patches | 20:46 |
| cp16net | this forces everything to pass. https://review.openstack.org/#/c/219967/ | 20:48 |
| cp16net | oh i see it | 20:49 |
| cp16net | look slike that patch is based on https://review.openstack.org/#/c/208073/23 | 20:50 |
| cp16net | and it should be 24 now | 20:50 |
| cp16net | that was something i just changed about hour or 2 ago | 20:50 |
| *** rjaiswal has quit IRC | 20:50 | |
| cp16net | if you follow the depends on link you notice your patch goes to a vesrion thats not current | 20:51 |
| cp16net | saurabhs1: ^^ see what i mean? | 20:51 |
| saurabhs1 | so I should just do rebase and I will be good then | 20:51 |
| saurabhs1 | ? | 20:51 |
| cp16net | i think sol | 20:51 |
| cp16net | so* | 20:51 |
| saurabhs1 | let me try | 20:51 |
| cp16net | k | 20:51 |
| openstackgerrit | Saurabh Surana proposed openstack/trove: Root enablement for Vertica clusters/instances https://review.openstack.org/196913 | 20:52 |
| cp16net | had to follow all the dependencies to find that | 20:52 |
| cp16net | looks better when i click through | 20:53 |
| saurabhs1 | cp16net thanks for figuring that out. I tried clicking on links to figure out why I was not getting the dependancies, but didn't notice that betwee 23 and 24 you had removed one of the dependsOn. You are super awsome | 20:53 |
| saurabhs1 | I will keep an eye on zuul for this change | 20:54 |
| * cp16net keeps learning new things about the gerrit system like that :-P | 20:55 | |
| cp16net | saurabhs1: its possible now that all the other patches that are dependent on that patch will need to be updated as well | 20:58 |
| saurabhs1 | yup they will be, but they are not in the critical path so I am not going to mess with them for now | 20:59 |
| saurabhs1 | I need to get this one fixed through | 20:59 |
| saurabhs1 | https://review.openstack.org/#/c/218548/ | 20:59 |
| saurabhs1 | should I just remove its depends on, I think that should be it isn't it? | 20:59 |
| saurabhs1 | cp16net ^ | 20:59 |
| openstackgerrit | Saurabh Surana proposed openstack/python-troveclient: Root enablement for Vertica clusters/instances https://review.openstack.org/218548 | 21:00 |
| *** crobertsrh is now known as _crobertsrh | 21:05 | |
| openstackgerrit | Saurabh Surana proposed openstack/trove: Add support for extended_properties for clusters https://review.openstack.org/206931 | 21:11 |
| cp16net | saurabhs1: should be fine | 21:28 |
| *** sharikapongubala has quit IRC | 21:30 | |
| *** tosky has quit IRC | 21:41 | |
| *** tosky has joined #openstack-trove | 21:45 | |
| *** tosky has quit IRC | 21:50 | |
| *** mattgriffin has quit IRC | 21:53 | |
| *** haomaiwa_ has joined #openstack-trove | 21:54 | |
| *** haomaiw__ has quit IRC | 21:56 | |
| *** sharikapongubala has joined #openstack-trove | 21:57 | |
| *** tosky has joined #openstack-trove | 22:02 | |
| *** tosky has quit IRC | 22:07 | |
| *** tosky has joined #openstack-trove | 22:11 | |
| *** tosky has quit IRC | 22:16 | |
| *** sharikapongubala has quit IRC | 22:18 | |
| *** tosky has joined #openstack-trove | 22:30 | |
| SlickNik | hi folks - just an FYI — we're tagging L3 with what's currently merged. For issues that are in flight / still in the gate, we've got exceptions so that these can land in the next few hours. | 22:30 |
| vkmc | makes sense | 22:31 |
| vkmc | thanks SlickNik | 22:31 |
| SlickNik | I've moved the bps that are still in flight to the rc1 page | 22:31 |
| *** sharikapongubala has joined #openstack-trove | 22:34 | |
| cp16net | alright sounds good | 22:35 |
| *** tosky has quit IRC | 22:35 | |
| saurabhs1 | amrith cp16net the jenkins cleanred on https://review.openstack.org/#/c/218548/ https://review.openstack.org/#/c/196913/ | 23:01 |
| *** csoukup has quit IRC | 23:17 | |
| *** tosky___ has joined #openstack-trove | 23:22 | |
| openstackgerrit | Merged openstack/trove: Test instance name can not have special charactrers in it now. https://review.openstack.org/219950 | 23:26 |
| openstackgerrit | Merged openstack/trove-integration: Fix ubuntu image builds due to revert in tarball names https://review.openstack.org/219548 | 23:26 |
| *** saurabhs1 has left #openstack-trove | 23:31 | |
| *** saurabhs1 has joined #openstack-trove | 23:31 | |
| *** sabeen3 has quit IRC | 23:49 | |
| *** ashleighfarnham has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
