[KLUG Members] RH 7.3, rpm sefaults when installing, upgrading, erasing

[Adam Williams]

>>Is it not set to start at boot, or did it crash/die?  If it goes away again
>>check to see if /var/run/.nscd_socket exists,  it should not if nscd is not
>>running.  But if nscd abends the socket may remain,  and thusly glibc thinks
>>nscd is still there to answer queries.... there awaits havoc and mayhem.
>I set it (nscd) to not start at boot... Teach me to not read up on it
>first:)  I've never really had problems with it crashing on other
>systems(or this one).  I guess I'm a little surprised at just how
>important it actually is, but the manpage doesn't really say so:)

It doesn't need to be running,  but glibc needs to be clear whether it
is running or not (The presence of .nscd_socket).  Unless your strapped
for RAM or have VERY few users and VERY simple group assignment you WANT
it running.  It is probably a bad idea to shut it down (if running)
while a machine is busy.

>>Make sure you have the latest nscd as the stability of your entire system rests
>>upon it,  and until recently it has been the soft spot on the underbelly of the
>>Linux dragon.  I've seen big beautiful IBM boxes go careening down to earth
>>because nscd got mumbly-jumbly;  and a crashed IBM is flat out criminal.
>>If nscd seems hung and you still have an operational console (GO SERIAL
>>CONSOLES!) you sometimes can yank /var/run/.nscd_socket away with "rm -f" as
>>root;  definitely a Lazarus class moment.
>Hehe... definitely... but everything worked fine that I saw, except
>rpm:)

Interesting,  makes one wonder if RPM is statically linked to something,
etc....