Replay cache file sharing

[Rick Macklem]

In a previous post, I tried to ask this, but I realized I worded it very
poorly (aka, outright wrong:-).

I am writing a server for the accept side of GSS authentication, that runs as
several processes for the same service. When I got the server credentials
after forking, the replay cache file (rc_nfs_N) would get corrupted. It hasn't
happened since I changed it to get the server's credentials before forking off
the N daemons.

Does anyone know if there are other problems w.r.t. sharing a replay cache
file between several servers? (I am using libraries built from MIT's KerberosV
sources.)

Thanks for any info, rick