Created attachment 815490 [details] screenshot Description of problem: Discover queue can't display large inventory (75 agents x 25 EAP6 instances) Version-Release number of selected component (if applicable): ER3 How reproducible: 100% Steps to Reproduce: 1. Install 25 EAP6 instances 2. Start 75 Agents via agentcopy script Actual results: Firefox shows spinning circle/loading data... timeout exception in server.log
Created attachment 815491 [details] server.log
I'll take a swing at this an try to pull data incrementally. Not with paging but rather fetching data as needed to handle expansion. Basically, increase the number of DB round trips.
ETA = now. release/jon3.2.x commit 85a41a0bc2cfc3512f13cc852def22ae5ea2cdef Author: Jay Shaughnessy <jshaughn> Date: Wed Nov 6 09:56:11 2013 -0500 The issue is that we attempt to pull all the ADQ resources at one time. Which overwhelms the view if there are many discovered resources. For example, when many new agents monitoring many servers all start up and report discovery results at basically the same time. This commit changes the approach such that we pull only the relevant platforms first. And then the server resources as needed, such as on tree expand. Things are complicated somewhat by the fact that we allow child server selection without platform expansion and also but supporting "Select All". Select All will pull the child servers for each server in a separate DB request. Whether this is the best (or an acceptable) approach remains to be seen. Finally, assuming we now can display a large inventory of uncommitted resources it remains to be seen whether we can select all and import them all successfully, as a quick look at the code reveals that we do it all in one transaction. It may make sense to break up the import work and process one platform at a time. Cherry-Pick Master : 2bc5aab6ec08055c1c7c981aabaded8f05fb44f4
I'm looking at some follow-up work now, to perhaps prevent a large import failure now that we should be able to actually perform a large import request...
master commit fd18ea22f83297bb806deba47365e86ebe23d573 Author: Jay Shaughnessy <jshaughn> Date: Wed Nov 6 15:02:57 2013 -0500 Related to the work on this BZ. Given that we can now support a large inventory in this view (hopefully) we may have a downstream issue trying to import the inventory in one large import request. So, this commit breaks it up into one import request per platform (well, per relevant platform that has selected uncommitted resources) *** note This is in master only as it doesn't immediately affect the BZ in question. But if desired, or certainly if there is a discovered issue importing a large inventory, this commit could move to the release branch. ***
Moving to ON_QA as available for testing with new brew build.
Mass moving all of these from ER6 to target milestone ER07 since the ER6 build was bad and QE was halted for the same reason.