Run 35K nodes simulation in shadow with low traffic
Implemented constant traffic in the node
analysis:nomos:simulation-analysis
sync with Moh and Nomos
Nomos sim team now has eveything to reproduce and fix the bug: exact configs and access to full runs
Suggested improvements to the data output that will reduce both the memory and disk overload of nomos simulation by orders of magnitude. Moh and Gusto agree that this will work.
Wrapped up the nomos analysis for now: waiting for the sims team to finish fixing the bugs
analysis-shadow:vac:shadow-gossipsub-analysis
Wrote an analysis script that can read graphs generated by Shadow runs
nwaku issue found : container crashes when a message is published with malformed timestamp
rest-api-specs issue found: missing fields in the REST API schema
software-testing:waku:test-automation-nwaku
Updated last PR, missing reviewer responses.
Investigating assert false: Ivan took charge of that task; found it doesn’t happen anymore (will write notes on the issue to resume investigation later)
Picking up pace with store tests.
Investigate a PEER_DIAL_FAILURE error in store; Happens when archive isn’t mounted (unexpected); Not yet reported.
revisited documents, and worked on presentation for logos research call on GossipSub Improvements.
Based on the feedback from the logos research call, revisited nim-libp2p documentation, codebase etc.
gsub-scaling:vac:gossipsub-improvements-paper
Started revisiting the GossipSub improvement paper to reflect current work and finalize writeup (Still work in progress. will need 1-2 more days to reflect in overleaf document).
Requested the DST team for initial simulation. I intend to use the outcomes to finalize test patterns.