Part 2: Analysis of Completeness of Downloaded Votes and Consensus


Development CollecTor Mirror

According to the logs and the downloaded data there was neither a vote nor a consensus or microconsensus missing during the 185 hours of logs, i.e., the logged statistics always report x/x downloaded vs. available votes.

Main CollecTor Instance

The logs from the main CollecTor mostly report x/x downloaded vs. available votes, but twice 0/7 votes were reported. On further inspection the zero number of downloaded votes was compensated by the process of importing data from the filesystem.

Another odd ocurrance was the report of 9/8 votes, which was due to 8 votes imported from the filesystem and one downloaded vote document.


The results from the main CollecTor instance might indicate some interference of the downloading and import processes. This could simply be due unclear statistics or a deeper problem. Thus, a first step for tackling this question is to verify and improve the runtime statistics. For further discussion/work see ticket #19169

Last modified 16 months ago Last modified on Jun 6, 2016, 9:25:59 AM