Terrell Russell

oh oh!

This might already be fixed with the user/zone fixes in 4.2.8-4.2.11. Will test to confirm.

Is there anything interesting/related in either the remoteZone rodsLog or the localZone rodsLog? I would have expected `CAT_INVALID_USER` before `CAT_NO_ACCESS_PERMISSION` for an undefined user.

So, this is definitely... correct? `CAT_INVALID_CLIENT_USER`? Since... you tried to get in, and the server doesn't know who you are, so, no, you cannot get in. --- You expected the...

I like the flexibility of note 3 above... perhaps consider `SERVER_LIST=server1.irods.org:/lustre_mount/resource1` (using a colon instead of a comma... like we do for `iadmin mkresc`. Would also need to consider default...

> > I like the flexibility of note 3 above... perhaps consider `SERVER_LIST=server1.irods.org:/lustre_mount/resource1` (using a colon instead of a comma... like we do for `iadmin mkresc`. Would also need to...

it's just a context string setting, so it can be set/manipulated after creation time (same as S3, and every other resource context string).

> Q. Are there use-cases for not specifying a vault path while in detached mode? yes, we want a subset of hosts, (aka HOST_LIST), but they will each be provided...

clang-format has some opinions.

The registrations issue is very subtle... do we just disallow non-vault registrations into unixfilesystem resources configured as `detached`? That seems... a brittle solution. Or... do we allow it to happen...