You will have noticed that, in the immortal words of Cheech and Chong, Dave’s not here. Dave (the WebApp front end) would very much like to make your acquaintance, but he’s currently being refactored to use the WebRTC communication protocol rather than Websockets, as currently the WASM/WASI/Websockets/libp2p combo isn’t up to the task. Which is a shame, but rather than hammering on the same door we want to get something out pronto, and that’s one of the things we’re doing right now.
However, we do have a new node release for you, that fixes a 3-hourly CPU and memory spiking bug. This is a hotfix release meaning you don’t need to upgrade, but it would be advantageous to do so, since it removes that spike, plus it has had good results in our staging comparison.
And node runners, don’t forget to check your place on the leaderboard.
General progress
@qi_ma fixed the CPU/RAM spiking issue for the latest release, and is also refining range based searching (Sybil resistance). Plus Qi made some chunk verification improvements.
@chriso has been continuing with the renaming of the crates (libraries) and binaries from safe/sn to autonomi as well as coordinating much of the work around the latest stable release.
@rusty.spork unveiled the new leaderboard and investigated a connectivity issue that appeared to be caused by a users’ VPN. We will need to keep an eye out for issues like this, so if you are struggling with VPNs do let us know.
As you might expect, there’s a whole lot of testing going on. @anselme has been testing the CLI with bigger and bigger files. We’ve been struggling with large files, so this is obviously an urgent fix to get in. He’s smoke testing the CLI in other ways too, trying to overload it. With a local network at least this is now looking very stable.
And @mick.vandijke is putting the graphical front end (aka Dave) through its paces, latterly getting WebRTC working as a replacement for the troublesome Websockets. We are able to fetch quotes over WebRTC and GET chunks, with PUTs still a work in progress.
@bzee is also in this territory, poking at performance issues, in this case with CPU. He’s also working through issues with WebRTC in libp2p, and getting a grasp on gaps in our API and documentation.
On infrastructure, @roland is improving the deployment setup for testing including allowing externally attached volumes to be attached to our in-house node VMs, while @shu has been troubleshooting log ingestion into the ELK monitoring stack, and also adding features to our logging dashboard.
@dirvine is implementing a fix to (1) reduce node CPU usage, (2) make registers more scalable and (3) simplify generic transactions all with a single simple data type.
Finally, Ermine debugged some crashes in vault sync testcases and created an intermediate fix. He also added more unit tests to the vault CLI, with increased file structures and sizes.
This is priceless 
patience super ants
