[10330] in Athena Bugs
Re: decmips mountd dumps core
daemon@ATHENA.MIT.EDU (Tom Yu)
Mon Mar 1 06:02:17 1993
Date: Mon, 1 Mar 93 06:02:11 -0500
To: epeisach@Athena.MIT.EDU
Cc: bugs@Athena.MIT.EDU
In-Reply-To: epeisach@Athena.MIT.EDU's message of Mon, 01 Mar 93 04:38:42 EST <9303010938.AA07461@paris>
From: Tom Yu <tlyu@MIT.EDU>
Reply-To: <tlyu@MIT.EDU>
It also seems that the mapping request makes it through, and an ack gets
sent back to the client. However, the mountd dies somewhere between
acking the mapping request and continuing the main serving loop. After
looking at the source (mountd.c), it seems that it is dying somewhere in
kuidops(), outside of the switch statement, since an unmap operation also
kills mountd. This seems to indicate that either svc_sendreply() or
svc_freeargs() is losing.
---Tom
(with my tickets)
podge% nfsid dragon
nfsid: DRAGONS-LAIR.MIT.EDU mapped
podge% nfsid dragon
nfsid: timeout while contacting mount daemon on DRAGONS-LAIR.MIT.EDU (for filesystem nfsid)
(restarted mountd)
podge% nfsid -u dragon
nfsid: DRAGONS-LAIR.MIT.EDU unmapped
podge% nfsid -u dragon
nfsid: timeout while contacting mount daemon on DRAGONS-LAIR.MIT.EDU (for filesystem nfsid)
(with systest tickets, after restarting mountd; systest is not in the
credentials file)
podge% nfsid dragon
nfsid: Authentication failed to host DRAGONS-LAIR.MIT.EDU for filesystem nfsid
podge% nfsid dragon
nfsid: timeout while contacting mount daemon on DRAGONS-LAIR.MIT.EDU (for filesystem nfsid)
