Thursday-Friday we saw SAM test and transfer failures into Antares due to missing router ACLs(?) on the new EOS front-end. 

Observed the problem again that number of running cores is very different in Vande compared to CMS monit. This discrepancy seems to be much reduced today. Doing some connection tests with FNAL to see if there is a problem (again) connecting to the schedulers based at FNAL. 

Testing transfers at CNAF today and yesterday. Using RAL as a destination for reads from CNAF. Investigating some errors seen at RAL, whereas other CMS T1 destinations used show much lower (or zero) error rates. One error found just existed for 2 minutes - possible network glitch? Tom Birkett contacted DI. 

Overall good performance of jobs relative to other CMS T1s.