EOS WFE should report better errors
The "outer" error message from the EOS WFE engine is always "Transport endpoint is not connected" no matter the cause of the problem. This is rather confusing.
[itctabuild02] silo (master) > grep ERROR /var/log/eos/mgm/xrdlog.mgm
180320 15:49:27 time=1521557367.380794 func=SendProtoWFRequest level=ERROR logid=static.............................. unit=mgm@itctabuild02.cern.ch:1094 tid=00007f9b1d0eb700 source=WFE:2022 tident= sec=(null) uid=99 gid=99 name=- geo="" RSP_ERR_USER Cannot allocate a new archive file: Storage class eosdev:double has no archive routes
180320 15:49:27 time=1521557367.381717 func=Emsg level=ERROR logid=e35fee42-2c4d-11e8-b193-00224daae19b unit=mgm@itctabuild02.cern.ch:1094 tid=00007f9b1d0eb700 source=XrdMgmOfsFile:2710 tident=smurray.10527:77@localhost sec=krb5 uid=19227 gid=1487 name=eosuser1 geo="" Unable to open - synchronous create workflow error /eos/dev/userfiles/testdir/group_1; Transport endpoint is not connected
[itctabuild02] silo (master) >
Please could the error message either be more generic or if at all possible related to the original cause.