- The technology we are using is not very good at updating. Since we are mainly focused on inserting and retrieving data, this should not be a big problem. However, if someone updates, for example, lat/long of a node, the repository will maintain both the old and new information. There is no way (at this moment) to distinguish which one is the current real data. This can and should be fixed.
- When I was writing the program to populate the RDF database for the name of the Measurements, I accidentally wrote ConditionalLossProability (note the missing b). Hence, the instance name for an individual of the class MD:ConditionalLossProbabilityMeasurement is SimpleMeasurement-ConditionalLossProabilityMeasurement. There is no known easy way to correct it using regular Sparql Update Query. Although it is not a big problem, we need to study a way to properly fix it.
- The list of remote hosts is created by dbpac.pl. It uses pingtable.pl last 60 days with average_rtt. The 365 days LOD analysis uses pingtable.pl and gets data going back 365 days. It is probable that there will be nodes for which there is data for some of the last 365 days but none of the last 60 days. This will result in a discrepancy in the statistics reported at the bottom of pingtable.pl and LOD.
- There are some odd files being created in Tomcat in Wanmon. Such as Snap.20130812.205008.5300.0013.trc, heapdump.20130802.193643.12532.0011.phd, ... These files take much more space than the usual ones. What is being done at this moment is deleting those files when identified.
Overview
Content Tools