I have a few questions... For some reason in RHEL4 I'm having nscd caching issues and in RHEL3 I'm having periodical dying after a day or 2... So I wrote a script to service nscd reload and then service nscd restart. My question is does reload just reload the database or does it 'by design' kill nscd so that the restart says 'Stopping [failed] and Starting [ok]'. I'm thinking of changing restart to start, even though this is a temporary fix, can anyone lend an ear to how I can trouble shoot why this is failing.... I started generating a /var/log/nscd.log by setup in /etc/nscd.conf but any help would be greatly appreciated. Thanks.

I'm having nscd caching issues
Interesting. Could you be more verbose/specific, maybe add an example of what you detect/troubleshooted already?


So I wrote a script to service nscd reload and then service nscd restart. My question is does reload just reload the database or does it 'by design' kill nscd
It you look at the default initscript you'll see it does a $0 stop; $0 start on restart. Looks like a kill and not a kill -HUP.

Well a fellow coworker and myself found out with running a 'nscd -d' while the other guy reloaded nscd (service nscd reload) caused a segmentation fault when removing the entries, so we definitely have a bug on our hands that we are going to pass on to RHN. This was one of 2 problems we were having the other being the periodical dying of nscd which is going to be harder to diagnose. Unfortunately I can't be too specific with the setup we use at my location but we are going to see what our TAM representative has to say. Concerning the -HUP option I'm not sure if one of our staff members changed the default but ours is showing 'killproc /.../nscd -HUP' for a reload in /etc/init.d/nscd.