[11394] in Athena Bugs
sun4 7.6Q: AFS
daemon@ATHENA.MIT.EDU (Dorothy Bowe)
Wed Nov 17 14:03:19 1993
To: bugs@MIT.EDU
Date: Wed, 17 Nov 93 14:03:10 EST
From: Dorothy Bowe <dot@MIT.EDU>
System name: pianoforte
Type and version: SPARC/Classic 7.6Q
Display type: cgthree
[also the DECstation running 7.6]
What were you trying to do?
Have a mountpoint to a backup volume in a replicated volume.
What's wrong:
The copy which propogates to the RO volumes is definitely
strange. I don't know if it's really a bug or not, but it's not right.
This problem first appeared in the form off an infinite OldFiles
subdirectory in a replicated volume. On closer examination, it appears
that each incantation of OldFiles reports the same inode number, and you
can really confuse yourself by typing these commands:
athena% cd /afs/athena/software/autocad_r12
athena% ls -lid . ..
1307181150 drwxrwxrwx 13 root root 2048 Oct 1 13:52 .
33161232 drwxrwxrwx 2 root root 4096 Sep 14 18:32 ..
athena% ls -lid OldFiles/OldFiles OldFiles
560989404 drwxrwxrwx 13 root root 2048 Oct 1 13:52 OldFiles
560989404 drwxrwxrwx 13 root root 2048 Oct 1 13:52 OldFiles/OldFiles
athena% ls -lid . ..
560989404 drwxrwxrwx 13 root root 2048 Oct 1 13:52 .
560989404 drwxrwxrwx 13 root root 2048 Oct 1 13:52 ..
athena% pwd
/
which makes sense because now . and .. are the same - only they
shouldn't be. Maybe we shouldn't have mount points to backup volumes in
replicated volumes? Is this related to the problem we saw a while back
with . and .. being the same in the maple RO volume?
What should have happened:
. and .. shouldn't change.
Please describe any relevant documentation references:
See the autocad locker. This only happens in the RO volume;
they RW volume is fine
