We've had long-standing issues where randomly we'll get -672 no access errors in trace during modify operations which then causes the attribute changes to fail. The odd thing is that these are all done with a service ID as part of an application and it's able to modify the vast majority of things without any problems. I have trace running with LDAP, NMAS and AUTH and it doesn't really tell me much. The only thing that I can think of is that the volume on this server is so high that perhaps it's not able to accurately calculate the access rights that the ID has. Does that sound like a valid possibility? I haven't been able to replicate it in the lab, and it happens on random objects, so it's been annoying to try to troubleshoot. I'm really just trying to get ideas on what else I can try to do to find a root cause for these, so any thoughts are appreciated.
↧