Issues:

• ECHO instability last Wednesday
  • Resulted in failed transfers/job uploads/downloads, and, most annoyingly, in file loss (GGUS 683184)
    • Due to race conditions between https and root protocols
    • There should be a way to mitigate this, e.g.
      • Poison DNS entry for webdav.echo.stfc.ac.uk on WNs (for LHCb jobs at least), so that it is redirected to local gateway as well
      • Introduce some locking/protection (e.g. do not execute delete if it arrived more than N minutes ago)?