users/groups

Bruce Dubbs bdubbs at swbell.net
Mon Apr 18 21:33:34 PDT 2005


Archaic wrote:
> On Mon, Apr 18, 2005 at 08:52:05PM -0500, Bruce Dubbs wrote:
> 
>>Why do something different for those?
> 
> 
> Just trying to get a feel for the issues at hand. ISTM that to buck
> historical use because obscure or highly specialized (liveCD) fs's might
> not like them is odd. Also, the live cd doesn't actually matter anyway
> because lastlog will not be written to. Therefore the file does not even
> have to exist. So again, I'm just trying to find out what value there is
> to *not* setting it to 65535. Just trying to paint a picture here,
> Bruce. I currently have no preformed opinion other than I would like
> lfs/blfs to match uid/gid and 65535 is the historical favorite. IOW, I'm
> trying to figure out if BLFS should use 65535, or if LFS should use what
> BLFS ends up using.

It is the historic way of doing things, but IMO, the value is
depricated.  For instance in RH Enterprise Linux we have in /etc/group:

nobody:x:99:
nfsnobody:x:65534

and in /etc/passwd

nobody:x:99:99:Nobody:/:/sbin/nologin
nfsnobody:x:65534:65534:Anonymous NFS User:/var/lib/nfs:/sbin/nologin

I've looked at the nfs source code and it uses nobody if defined and -2
otherwise.  If systems are configured properly, the -2 values is never used.

I will also note that on my RHEL system we have

$ ll -h /var/log/lastlog
-r--------  1 root  root 18M Apr 18 23:17 /var/log/lastlog
$ du -h /var/log/lastlog
44K     /var/log/lastlog

It's certainly a minor thing.  The only problem I know of if a system
does not have a user or group name defined is that a ls -l produces a
number instead of a name.  If nfs servers have nobody defined, it
doesn't matter what the number is.

  -- Bruce





More information about the blfs-dev mailing list