[5280] in Athena Bugs
vax 7.0F: NFS attach/kernel
daemon@ATHENA.MIT.EDU (Jonathan I. Kamens)
Mon Jun 25 17:12:18 1990
Date: Mon, 25 Jun 90 17:12:05 -0400
From: "Jonathan I. Kamens" <jik@pit-manager.MIT.EDU>
To: cfields@ATHENA.MIT.EDU
Cc: bugs@ATHENA.MIT.EDU
In-Reply-To: bugs[5131]
From: cfields@ATHENA.MIT.EDU
Date: Sat, 09 Jun 90 22:59:08 EDT
There is an NFS bug which sometimes causes there to appear to be a
null or two in your file when there really isn't; I was using the
detach hack that jik just told me about to straighten NFS out.
Yes, this is a bug. I'll file it.
The hack is to detach the filesystem while you are cd'd into it; since
it's busy, the detach will fail. But, while the detach fails, it does
purge your NFS mappings. Here is a sample:
fries 10:46pm 98> cat > dog
sdfkljdsk
fries 10:47pm 99> rm dog
fries 10:47pm 100> detach calendar
calendar: Unable to unmount /mit/calendar: Device busy
fries 10:47pm 101> cat > dog
dog: Permission denied
This isn't a bug. When you detach a filesystem, you are signalling to
the system that you no longer plan to use that filesystem. At that
point, detach is *supposed* to get rid of your mappings, even if it
can't unmount the filesystem for some reason.
Apparently this is only for the VAX; I couldn't reproduce it on my RT.
I could, so you were doing something wrong :-)
Jonathan Kamens
Project Athena Quality Assurance
