Xrootd

• Added documentation about using xrootd proxy as a DTN.
  • Difference will be how the copy program will be started
  • University in Oklaoma is already using it but BNL deployment is delayed
• Authentication and authorization decision is between option 4 and option 5 we talked about few weeks ago
  • 4 is delegation and 5 is TLS and tokens.
  • For native xrootd 4 is easier to implement and has a ETA of about a month.
  • They have started to look at 5 too, but there is no plan yet.
  • dcache cannot say how long it will take to implement 4 without protocol documentation and would prefer 5
  • Writing the protocol documentation should be a priority to allow dcache to make a plan
    • Action: Wei push the xrootd team to produce this documentation

Http

• StoRM there was a problem with the post for a macaroon which returns the entire file. This shouldn't of course happen Andrea will have to look into it
• Tim thinks he can start testing echo S3 next week after some problems RAL had in the past couple of weeks have been solved.
• echo end points have been added to the twiki today.
• In the UK there are transfers tests ongoing for DUNE between dcache and dynafed. They should be made part of the group activities and not run in parallel.
• Wei asks if a http DTN style node for sites with firewalls is in the plans. It can certainly be done but it will not be packaged in containers. The documentation for this needs to be written.

Rucio

• rucio test instance is now up and running. There have been few teething problems and some bug to iron out. It is however integrated with FTS and has already few end points configured. Bonus we have also the monitoring.
• Thomas will investigate the possibility to get a semaphore matrix out of the monitoring so it doesn't have to be compiled manually.
• Wei asks if this is the new rucio instance he has to use for testing

Next meeting in two weeks: 26 September 2018