RE: [nfsv4] (no subject) - Owner Group attr strings.

Next message: Talpey, Thomas: "[nfsv4] Draft agenda for NFSv4 meeting in Seoul - IETF 59 (fwd)"
Previous message: Sergey Klyushin: "RE: [nfsv4] (no subject)"
Next in thread: Spencer Shepler: "Re: [nfsv4] (no subject) - Owner Group attr strings."
Reply: Spencer Shepler: "Re: [nfsv4] (no subject) - Owner Group attr strings."

scallen@us.ibm.com

Subject: RE: [nfsv4] (no subject) - Owner Group attr strings.
Message-ID: <OF6CDF0BA2.8A840D48-ON87256E47.008044B8-88256E48.000107D8@us.ibm.com>
From: Stevan Steve Allen <scallen@us.ibm.com>
Date: Fri, 27 Feb 2004 16:11:17 -0800

>>Do you mean “OWNER@” and “GROUP@” are not supported by server or 
something else?
>>Why would domain name be unknown?
>>- Sergey

I am speaking of owner and owner_group attributes 36 & 37 not being 
supported by the server for SETATTR.  The protocol states if the server 
supports owner & owner_group on SETATTR, the server promises to return the 
exact same string on a subsequent GETATTR.

If a server can not support the requirement for setting and returning the 
same owner & owner_group attributes,  it seems it must turn off support in 
the server supp_attr mask, which would also affect getattr and readdir not 
returning the strings.  The server turning owner & owner_group "on" would 
suggest the server supports a SETATTR of owner and owner_group.

Does anyone expect otherwise?  ...such as... a server which says it 
supports owner/group (attr mask) and returns internal strings as 
owner/group and does not allow setting these values (using an error code 
for setattr). ?   e.g. Is there a benefit or need to obtain the servers 
internal strings vs. nothing? 

note:  other R/W attributes have separated the read and write support by 
adding a second attr, such as cansettime.

_______________________________________________
nfsv4 mailing list
nfsv4@ietf.org
https://www1.ietf.org/mailman/listinfo/nfsv4