[2111] in Kerberos_V5_Development
Re: Krb5-1.0 and t_kdb - It works on HP
daemon@ATHENA.MIT.EDU (Mark Eichin)
Tue Dec 10 20:52:25 1996
To: Ezra Peisach <epeisach@MIT.EDU>
Cc: "Theodore Y. Ts'o" <tytso@MIT.EDU>, Ken Hornstein <kenh@cmf.nrl.navy.mil>,
deengert@anl.gov, krbdev@MIT.EDU, Paul A Vixie <paul@vix.com>,
Andrew Hobson <ahobson@mindspring.com>,
Dave McGuire <mcguire@neurotica.com>, Joe Gross <jgross@uiuc.edu>
From: Mark Eichin <eichin@MIT.EDU>
Date: 10 Dec 1996 20:50:09 -0500
In-Reply-To: Ezra Peisach's message of Tue, 10 Dec 1996 07:23:41 EST
> The real problem is that for database locking - you need a filesystem with
> more specific characteristics - which may be more difficult to determine.
Note that the ccache code attempts one type of locking, and if it
fails *in a certain way* attempts another (this is specifically to
handle ccache's in a sunos tmpfs.) The other locking code in the tree
is independent, and doesn't use that hack (and probably shouldn't...)
However, determining the characteristics is simple -- just try the
lock, or maybe two of them. The hard part is that at least in some
cases (NFS to a linux (and aix also?) server, for example) the failure
is that the client (sun anyway) hangs forever on the lock...
