Backup and access list question

[John F Haugh II]

In article <steve.1279 at trashy.UUCP> steve at trashy.UUCP (Steve Talmage) writes:
>Does anyone know exactly when AIX 3.1 access lists are or are not backed up
>when backing up files?	I would think the answer is "just the backup command
>itself", but I don't want to make that assumption.

The backup command is, so far as I know, the only command that backs up
a copy of the inode extensions when copying a file.  The "tar" and "cpio"
commands do not have the space needed to make a copy of that extra data,
so there is no reason to put it there since only AIX would be able to
read it back (and tar and cpio are both intended to be used for data
interchange ...)

>Also, do other commands (such as cp) copy the access lists too?

The answer should be "it depends".  Very few commands require that the
ACL's be preserved - "mv" comes to mind as about the only one.  If you
really want to copy an ACL, the pipe "aclget <source> | aclput <target>"
works just fine.

>Yes, I could try a bunch of tests, but this question got stuck in my head
>and has kept me up thinking about it, and maybe typing it in here would let
>me rest easier.  And, yes, I'm at home, nowhere near my 6000s, and I'm
>rambling terribly.  Thanks for any help.

Is there some reason that you think ACL's should be copied?  I used to
believe that any program which makes a copy of the file should also make
a copy of the ACL.  For reasons I won't go into her, I no longer feel
this way.
-- 
John F. Haugh II        | Distribution to  | UUCP: ...!cs.utexas.edu!rpp386!jfh
Ma Bell: (512) 255-8251 | GEnie PROHIBITED :-) |  Domain: jfh at rpp386.cactus.org
"UNIX signals are not interrupts.  Worse, SIGCHLD/SIGCLD is not even a UNIX
 signal, it's an abomination."  -- Doug Gwyn