Anna M, Steve W, Marc M, Robert I, Brandon W, Vladimir P. Yuyi G, Margherita VW.
Reviewing tasks for version 3.1.0
1) Reviewing task for version 3.1.0
STEVE: going over status of open tasks.
19504: about dependencies and SAM metadata: It’s done, need little more testing.
Put in integration soon so people can test.
We have a test suite, Who runs it?
Running it, mainly after making big changes.
Need to work on design the Wizard which need to go along with the Work Flow Editor.
Also get involved in documentation besides Margherita. Yuyi is working on project.py.
About web pages, maybe improve some of the names of the links in the side menu, like,
instead of ‘Compose Launch Template’ ‘ View/Compose…’ since that’s the place people need to go to find existing templates..
Campaign Stages running very slow, there was an incident opened.. We need to understand what is causing the slowness again.
Have some statistics information from FIFEMON, example,how many jobs are running now? Have FIFEMON pages to show how
many jobs are running for POMS and an experiment. Also have some timing stats, need to know how long it takes for the different steps.
We have some information, memory stats and average response time.
We already requested to have the name for the campaign stages.. we will request to add more info.
As far as slowness, maybe we do need to revive the idea of archiving. Need monitoring so we could see what’s going on in the past few days.
Viewing the pages, taking very long time.
Instead of viewing all, we can use the check boxes to select which campaign statististics we want to view..
There might be a problem with sessions, we are losing the set role when navigating through the pages. Need to create an issue.
Maybe we need to get rid of sessions..
Keep session cookies in the browser.
Need to see if we are loosing information through the sessions.. Maybe we need one process with multiple threads.
Looking into slowness and looking into the db side.. it looks like there are locks held. Select for updates will lock itself.
This is confusing, we should not be doing any locks.
Could it be that the connection pool is full?
Can we use ‘select for share’ instead of ‘select for update’ ?
We need more monitoring to follow what’s going on and to see where we lose performance.