[282] in testers
Re: attach (6.3, 6.4)
daemon@ATHENA.MIT.EDU (daemon@ATHENA.MIT.EDU)
Mon Nov 27 01:03:36 1989
Date: Mon, 27 Nov 89 01:03:07 -0500
From: Theodore Ts'o <tytso@ATHENA.MIT.EDU>
To: John Carr <jfc@ATHENA.MIT.EDU>
Cc: testers@ATHENA.MIT.EDU, rel-eng@ATHENA.MIT.EDU
In-Reply-To: John Carr's message of Mon, 27 Nov 89 00:16:31 EST,
Reply-To: tytso@ATHENA.MIT.EDU
Date: Mon, 27 Nov 89 00:16:31 EST
From: John Carr <jfc@ATHENA.MIT.EDU>
The release and testers versions of "attach" both refuse to remove an
AFS "filesystem" from attachtab if the mountpoint symlink does not
exist.
Yikes! I thought I had gotten around to fixing this --- it was on my
buglist, but I obviously forgot..... having this bug be fixed is
somewhat important, since users have often come into the SIPB office
complaining that "attach sipb" said "already attached", and "detach
sipb" said "not detached, lose." (Repeat forever) Current fix is to
walk over to W20, go to root, and edit attachtab by hand. Lose, lose.
The fix is very simple --- you just need to add an if statement to the
error checking of afs_detach(). Is it too late to slip a change into
6.4? Sorry for not fixing it until now.....
- Ted
*** /tmp/,RCSt1000946 Mon Nov 27 00:50:31 1989
--- afs.c Mon Nov 27 00:50:21 1989
***************
*** 157,162 ****
--- 157,168 ----
}
if (unlink(at->mntpt)) {
+ if (errno == ENOENT) {
+ fprintf(stderr,
+ "%s: already detached, continuing....\n",
+ at->hesiodname);
+ return(SUCCESS);
+ }
perror(at->mntpt);
fprintf(stderr, "Couldn't detach afs filesystem %s.\n",
at->hesiodname);
