RE: Analysis of the Late Gridcoin Superblock and Suggested Solutions
@Dutch, thanks for explaining. I really appreciate the details as a user of the GRC community.
Regarding the proposed solutions
Excluding any CPIDs less than an hour old from the BOINC project servers' data before carrying out the hash. This will reduce the likelihood of a hung quorum.
Adding a DPOR weight requirement to be selected for the quorum. This ensures that once the quorum does reach consensus, there is a relatively high likelihood of a member to stake and generate the new superblock.
I support these -although @nuda1 had a good suggestion as well regarding beacon age rather than DPOR weight...but that might be how we have to implement the first bullet. It would be great to see if Rob or one of the other active developers would be willing to implement this in the code so we can get a new update pushed out. I think this is going to be critical as well if we're going to eventually remove the team requirement (which I also support doing).
I would like to ask though: what does a lot of new users mean? What is the scale we're talking here? 10s, 100s, 1000s?
Would it not seem legit that maybe the issue also lays in the fact we just merged Rob's code with the #gridcoin community dev's code and merged the two code tree's. Since then we have had 1 if not 2 blackswans and instead of forking we are lucky enough to not get a superblock.. I would think since we were fine for months , and we had the same influx of users the past few weeks/months that seems a little weird to blame it on the number of new user beacons into the nn and the newbie block etc. IMO and no offense to the dev team , roll back the code pre merge and go from there back to 1 coder vs 5-6 additional. Yes this happens etc etc , but also Rob has in the past put in things like code time bombs and forgotten about them himself and caused us to fork , soooo this could be due to the merge of code either with the " consensus or quorum " issue being a result of something deeper. I am no dev , but when something breaks right after you change things completely and merge things the best fix is to remove the modifications and changes. ( btw , i love that to community is getting a change to get involved in dev and my hypothesis is nothing negative at them , they are much appreciated ) it's just how you would deal with it , if it was physically and tangible in your hands.
This is a bit negative. Before casting judgment I do believe these gentlemen have done some fine analysis - but, I would actually like to see the information that backs it up. How many new members? What's the magnitude of the issue - if we're going to blame it on scaling, then show us the scale. I would tend to agree with you, @jamezz, when we went up from $0.004 to $0.01, we added a lot of new users too - is that number smaller than the scaling issue that supposedly exists now?