This last week we’ve been firing on all reward-program-cylinders, and hopefully will have something to show you all very soon. Feedback loops in place, leaderboards being assembled, and a whole world of fund being loaded into the Beta cannon.
To that end, @roland has been working on integrating the initial audit node into our testnet infrastructure. Alongside some fixes for the launchpad (MaxRepaymentsError 
), he and @mazzi have worked on improvements to the launchpad (simplifying some of the resource allocation process).
@chris has been in about the launchpad and node manager, fixing some windows bugs, adding options and re-enabling the integration tests that were proving overly problematic in the last week.
@anselme has further optimised DAG audit code, both in the audit-node and the client code. He and @qi_ma have worked hard to bring the simplified reward-collection to life, which will form the backbone of the beta-rewards process.
@mazzi has also started work on a discord bot to facilitate returning stats to users in a private fashion, and to help with managing the various beta processes.
@joshuef has done some work on key handling in the code, along with @anselme and @qi_ma, getting us setup so that we have decent defaults for local development and testing, which we can easily override where we need keys to remain secret (hello foundation!) so we can start integrating that into our processes.
@benno has continued pushing the auto-NAT and UPnP detection processes, such that we can automate this and ensure that nodes are started with the optimum arguments.
@qi_ma finalised the encryption of discord users as part of reward forwarding. This should let us keep the reward quantities private, while still being able to provide some interesting and fun stats to everyone participating.
@shu has been testing out the load of the network dashboards, successfully bringing memory usage for 2k nodes from ~60gb down to 2.5gb with some optimisations and tweaks to metric collection and generation. That puts us in a great place to begin monitoring not just internal networks, but to also cope with metrics generated from community nodes in the upcoming beta phases.
Alongside this, @shu set up some NAT-blocked nodes to test out the --home-network flag and was able to reproduce many of the errors which folks have been seeing. This is good as it gives us something to test against, though it does seem to confirm what many suspected, which is that --home-network is not the silver bullet we were hoping it was yet. It may well be causing more harm than good so far. Now we can measure this, we’re hopeful to be able to make some progress there in the coming weeks.
demo, under the strain of which my laptop gave up it’s ghost and I had to wait weeks for a new one… today I inflicted it on several willing, even enthusiastic Guinea Pigs. Thank you all.
for having to go through that manual port forwarding stuff for so long, when it was there to be had all this time.
