Spotted 25% of CMS transfer from Echo to Antares failing. ATLAS had the same. Only IPv4 transfers were failing and not all of them. IPv6 was 100% successful. At first guessed it was related to Antares upgrade/new hardware, but then asked Jyothish and he quickly proved it was missing IPv4 routes to Antares from the new 23 tranche of Echo gateways. These are now blacklisted and Jyothish is following up with DI. In the meantime the failures stopped. 

There is clearly Antares data moving using IPv6. But because the current MGM does not have a publicly routable IPv6 address (quote: George) the CMS SAM tests remain in warning state. 

There have been some issues with the new ETF tests (token) for ARC-CEs, which were briefly included in the Site Status. These have now been removed from Site Status until better stability is achieved. Errors have been 'corrected' by CMS. 

The possible network issue stopping jobs being submitted from Schedds at FNAL seems to have been solved. So far as we know nothing was changed on the RAL or the FNAL side. But we can now do traceroute etc. to the relevant machine(s).