Closed cpitman closed 8 years ago
@cpitman
I'll pull this in and test against my uses cases.
Yep, good catch. I've added the same logic to jboss_batch. Again, it appears to work on a simple test case.
Honestly, I'm mostly worried about whether or not the CLI is going to throw some sort of syntax we've never seen before when doing a full container configuration parsing. I already had to make a change here for '(' and ')' characters that I'm a little nervous about. The actual batching of reading is pretty straightforward.
@cpitman I'll put it into my larger test environment that I am building out and see what happens.
As a heads up though I likely wont be able to finish testing this week and I am on vacation next week, so it maybe a bit before I can get us some good results.
~Ian
@cpitman any idea what could be causing the issue I am seeing with this branch?
This is a performance improvement for the current CLI provider. Instead of doing a read-resource for each resource, it does a single recursive read-resource from the root, and then searches for each resource within the result. I've tested this locally on some simple manifests, and it appears to be working. However, it is a large change to how things work. @itewk can you test if this works for you?
I'm using the quickstart_logging test case to profile this. What I am timing is the length of a run after all the resources exist. My expectation is that the exact same amount of time will be shaved off from an initial run. Before the change, it takes 11.5 seconds. After the change it takes 2.7 seconds.
This appears to save roughly .5 seconds per resource.