09:14:46 #startmeeting check-in_20170120 09:14:47 Meeting started Fri Jan 20 09:14:46 2017 UTC. The chair is brainfunked. Information about MeetBot at http://wiki.debian.org/MeetBot. 09:14:47 Useful Commands: #action #agreed #halp #info #idea #link #topic. 09:14:47 The meeting name has been set to 'check-in_20170120' 09:15:02 #topic build status 09:15:16 k4n0, what's the status on the framework changes we'd discussed yesterday evening? 09:15:53 brainfunked, checkpoints? 09:16:03 k4n0, ? 09:16:22 brainfunked, are you asking about namespaces? 09:16:27 k4n0, namespacing, definitions, commons etc. 09:17:07 brainfunked, namespacing basics is done, we are in progress with converting all objects via this way 09:17:48 are ndarshan, nishanth, Shu6h3ndu and anmol all in sync with the changes? 09:17:53 brainfunked, I have not been able to get all your definition ideas adopted to our definitions schema, but the ones required for import are adopted to Tendrl Definitions Schema, 09:18:12 k4n0, the idea is for those changes to be brought in piece by piece as we develop features 09:18:21 brainfunked, ack 09:18:21 bring in the changes required for a particular feature only 09:18:54 brainfunked, Commons: there will be additions to commons which will be common namespacing code which already exists in Node agent 09:19:32 Not sure if nishanth Shu6h3ndu are completely up to date, but I did respond to Shu6h3ndu's comments on the namespacing PR 09:19:55 they'll probably be able to sync up once done with the training today 09:20:12 brainfunked, yea 09:20:13 k4n0, so what's the exact status on the changes and what still needs to be done? 09:20:55 brainfunked, namespaces needs to be tested for all repos with all objects in those repose 09:21:19 brainfunked, once above is green, we test the generic importcluster flow via the node-agent 09:21:40 ok, so the changes are complete or still being made? 09:21:45 brainfunked, once that is green, we test the integration child flows which will be invoked by the generic flow 09:22:10 brainfunked, that along with the logging work gowtham has been working on 09:23:17 brainfunked, generic import cluster flow needs to be coded, its basically the current import cluster minus the hard coded integration details. The atoms in this flows are already in place, but as i said, namespaces need to be tested before anything 09:23:59 k4n0, hold on, generic meaning the base ImportCluster flow that I'd suggested, which would lead to a system specific flow being invoked? 09:24:09 brainfunked, yep 09:24:39 brainfunked, we already have that flow implemented, but as of now it contains integration related details too, which are to be removed 09:24:50 ok 09:25:08 so do we have an ETA on the namespace changes being tested? 09:26:45 brainfunked, firstly, all Tendrl objects need to be adopted to the namespaces, ndarshan and me are working on the Node agent objects, will be done adopting those today, Shu6h3ndu and nishanth had to cover the integration objects (ceph, gluster), and anmol was to look at perf an alerting objects 09:26:58 ack 09:27:31 I am Incoporating changes as per https://github.com/Tendrl/node-agent/pull/147/ into https://github.com/Tendrl/performance-monitoring/pull/28 for config object and etcd_orm will try other changes only once this works(I started following the patch only today as I was on PTO yesterday so it took me some time). But looks like there is a new update on https://github.com/Tendrl/node-agent/pull/147/ need to follow it. 09:27:56 anmol, thanks for the update 09:27:57 anmol, please raise questions on the PR as required 09:28:15 brainfunked, same for you 09:28:22 and as discussed yesterday alerting is not a priority for us.. So, I am mainly looking into performance_monitoring related changes 09:28:40 anmol, yeah, performance monitoring needs to be solid right now 09:28:56 brainfunked, ack 09:28:58 ndarshan, your status on the namespacing related stuff? 09:30:50 k4n0, btw, logging related question 09:31:15 brainfunked, yes 09:31:23 k4n0, it seems to me that it would useful to have the Message class and the eventing classes in commons to be used by all the modules 09:31:30 there are two distinct requirements 09:31:48 everyone other than the node agent should be able to log well formed messages to the message socket 09:32:07 node agent then needs to take the messages to the approapriate output channel (etcd/syslog) 09:32:27 the Message class will always be the same, before posting to the socket and after reading from the socket 09:32:54 json structure from that class essentially serves as marshaling and unmarshaling 09:33:08 the eventing class needs to be used by other components to log to the node agent sock 09:33:11 socket* 09:33:29 and the logger class will be used by the node agent to actually dump the messages in the appropriate location 09:33:36 k4n0, gowtham, make sense? 09:33:48 brainfunked, I have a PR (https://github.com/Tendrl/node-agent/pull/145)to fix the commons issues before the name-spacing patch(with this single node gluster import worked). Today morning I tried to apply my fix PR mentioned before, after pulling Rohans name-space changes resolved conflicts and now trying to run that. Ones it starts successfully will coordinate with rohan to pick up some objects in node_agent for adding to namespace 09:34:07 brainfunked, ya ok 09:34:09 brainfunked, It does, can we get this on the spec? I have implementation doubts 09:34:33 Message and Logger class relates stuff I'd commented on the spec 09:34:45 the event class not yet, I'll comment 09:35:01 gowtham, if you've understood the point, you could update the spec with these details directly 09:35:26 brainfunked, ok sure 09:35:54 brainfunked, for local logs, lets stick to python logging framework, Its gonna take substantial time to get your comments implemented 09:35:59 ndarshan, ack, I've been keeping an eye on that PR, and it looks fine, but once the commons changes `freeze', let's have a final sync on that PR before we merge it 09:36:16 brainfunked, for operationals and alerts/notifications, the socket makes sense 09:36:25 k4n0, the logger class can be python standard, no problem 09:36:34 k4n0, it already allows syslog output 09:36:49 k4n0, the socket already exists thanks to the alerting code, we've just expanded it's scope 09:36:54 brainfunked, yes, I am just unsure of the integration time 09:37:01 *implementation 09:37:44 k4n0, ok, let's do it this way, let's finish off the integration without these logging patches being merged right now 09:37:57 ensure the stability of all the components first 09:38:15 brainfunked, ack 09:38:17 that gives us a build without the properly formatted logs 09:38:28 and on the side look to merge the logging changes 09:38:40 the logging changes won't block any of the existing features being merged 09:38:55 as it is, those need to be properly stabilised first 09:39:22 and we need an instance up today 09:40:46 brainfunked, regarding the instance, youll need the instance with the namespaced objects? 09:40:55 yes of course 09:41:07 more than me, noopq needs one 09:41:31 brainfunked, thats not doable today looking at the current progress of all objects being namespaced 09:41:55 brainfunked, but the node-agent will be doable as i said 09:41:58 k4n0, I need a instance with the latest changes, it could be anything from definitions to changing jobs log lookup path 09:42:28 k4n0, can we get things merged to master one by one and start pushing the changes to the instance? 09:42:52 k4n0, it also allows us to test the use case where we're ensuring that the node agent is able to push the changes properly 09:43:05 brainfunked, I am doing that for the node-agent, we cant have some node-agent objects inside a namespace and others without one 09:45:37 brainfunked, id suggest we ask gowtham to test out these node-agent changes parallely and report any issues? 09:46:03 brainfunked, while me and ndarshan ensure all node-agent objects are namespaced 09:56:07 sorry, had to attend to a delivery 09:56:22 works for me 09:56:31 ndarshan, gowtham, works for you guys? 09:57:06 brainfunked, okay 09:57:16 brainfunked, ok 09:57:23 ok then 09:58:00 #action k4n0 and ndarshan to complete the object namespacing changes, push them to the etcd instance and test things out 09:58:50 #action gowtham to keep working on the logging specification https://github.com/Tendrl/specifications/pull/94/ to make the changes requested from the reviews and ensuer that they align with the commons changes 09:58:50 brb 09:58:59 noopq, what's your status? 10:04:35 brainfunked, I am going to finish the API doc today. 10:04:55 and will start with Gowtham's changes from the spec for the task details 10:05:23 brainfunked, Also I have asked neha to make some changes to the endpoint names as we had discussed earlier, she said she can do it today. 10:05:30 great, the spec looks finalised in terms of the message structure and etcd schema etc. 10:05:42 the implementation details for commons are being finalised, but that shouldn't block you 10:05:55 ok 10:06:10 #action noopq to finish the API documentation today 10:06:28 brainfunked, Also I am verifying the previous API don't break, like create volume and pools 10:07:37 #action noopq to sync up the API codebase based on the changes in the logging specification and provide the job updates API details so that the UI could implement support for them 10:08:04 #action noopq working on ensuring that there are no regressions with APIs that had already been shipped 10:08:18 neha, ankush, status? 10:09:28 I have completed work for "Nested left navigation for tasks menu". 10:09:50 Sent the patch for the same https://github.com/Tendrl/dashboard/pull/105 10:10:14 brainfunked, reviewed https://github.com/Tendrl/specifications/pull/125. working on https://github.com/Tendrl/dashboard/issues/101 10:10:23 Waiting on neha and kamlesh review. 10:11:21 neha, please provide the titles 10:11:46 Task Progress page for Import cluster functionality. 10:12:22 Also will start with the url changes of list and import clusters which noopq has suggested 10:13:06 Also writing Pool features for BDD, will be completing it by today. 10:14:44 ankush, can we get the BDD repos upstream today? 10:16:50 brainfunked, yes we can. But we have to decide upon whether to put them inside frontend project or to create a separate repo for this. 10:18:46 ok, let's take that discussion post the check-in 10:19:04 brainfunked, ack 10:19:22 #accept ankush has completed the nested left navigation for tasks, and has sent a pull request 10:19:42 #action neha and kaverma to review ankush's left navigation for tasks PR https://github.com/Tendrl/dashboard/pull/105 10:22:02 #action neha is working on the task progress page for import cluster https://github.com/Tendrl/dashboard/issues/101 10:22:10 #topic testing updates 10:22:27 fbalak, mkudlej, mbukatov, ltrilety updates please 10:23:03 fbalak: go 1st, let's follow the alphabethical order 10:23:37 #action fbalak Working on https://github.com/Tendrl/usmqe-tests/issues/20 10:23:38 #action fbalak Blocked by https://github.com/Tendrl/gluster-integration/issues/108 and https://github.com/Tendrl/ceph-integration/issues/78 - with tendrl-common still listed as dependency for new packages, tendrl-commons can not be installed from packages. 10:24:50 brainfunked: I work on tendrl unrelated tasks 10:25:05 #info fbalak working on the import cluster test cases https://github.com/Tendrl/usmqe-tests/issues/20 and is blocked by unavailability of updated packages 10:25:26 brainfunked: sorry, I should step out for while 10:25:29 #info mkudlej is working on tasks not related to tendrl and has no updates 10:26:39 brainfunked: are there new packages which we can test, please? 10:27:16 mkudlej|afk, not yet, codebase changes are being integrated and tested before the packages are built 10:28:59 #info mbukatov did review of pull request #23 of usmqe-setup, #27 of usmqe-tests 10:29:06 #link https://github.com/Tendrl/usmqe-setup/pull/23 10:29:13 I am working on https://github.com/Tendrl/usmqe-tests/issues/12, still waiting on packages 10:29:14 #link https://github.com/Tendrl/usmqe-tests/pull/27 10:31:00 #info ltrilety is working on the gluster import scenario tests https://github.com/Tendrl/usmqe-tests/issues/12, awaiting updated packages 10:31:11 #info mbukatov has reviewed https://github.com/Tendrl/usmqe-setup/pull/23 https://github.com/Tendrl/usmqe-tests/pull/27 10:31:15 alright, thanks for the updates 10:31:18 #endmeeting