9/23/2019 Agenda & Notes

Displayed the release view for 1.2. We have been focused on regression testing and we have a few bugs that were causing the regression we have talked about before. (Mark)
The regressions were caused by the number of Rust to Python FFI calls. Several changes have gone in on Friday and we now have some stable running network for PoET and some new issues with PBFT we are investigating. We are hoping to start the official LR7 runs for the release this week. (James)
Can you elaborate on the issue? (William)

We are in the process of rewriting the validator into rust from python. We included a new component called the BlockManager written in rust, and the expense of calling into the BlockManager from the python code is high, so we needed to add more caching on the python side for checks that happen quite frequently around existent checks. We also needed to evaluate when we release the python GIL.

It was not incorrect, it was just falling apart under load that we expected it to be able to handle based on 1.1 (James)

Ping the people shown as reviewers to get in the remaining PR review. I will also contact people to assist with this. (Shawn)

In PBFT the default location is memory instead of stored in disk, like in sawtooth. I think we should switch to making disk default at the same default location as Sawtooth (Arun)
- That sounds fine, please make a PR for this. (James)
The deadline for submission of topics to the Global Hyperledger Meeting in March is this week. Was anyone on the call planning to provide a Sawtooth topic and attend? (Arun)
- A good idea is to start a HL Sawtooth List topic with this question to start a thread. Also mention it on the Rocket Chat #sawtooth channels. (Mark)