Description of problem: If ypserv crashes, ypbind processes silently disappear instead of patiently waiting for ypserv to come back. Version-Release number of selected component (if applicable): ypbind-1.17.2-3.i386 How reproducible: Quite often, but not always. This time, in a cluster of about 10 machines, only 1 ypbind was missing. All the others survived and reconnected to ypserv when it was restarted. ypbind on the NIS server machine seems to never disappear. Steps to Reproduce: Take an NIS server with a few NIS clients. Kill ypserv (the infamous "svc_run: - poll failed: No child processes" problem in RHL9). Wait for, say, 12 hours, restart ypserv. Actual results: On some clients ypbinds are gone and have to be restarted manually. Expected results: All ypbinds should survive the NIS outage and reconnect to ypserv when it becomes available. Additional info: We see the same problem in Fedora 3. RHL9 was fine. No data for Fedora 1,2 or 4. K.O.
Forgot to say what I see in the system log: "ypbind[3317]: broadcast: RPC: Timed out." while ypserv is down. Then these messages stop and ypbind is not running. K.O.
I believe this is the same problem that caused bz #167061. Please try the ypbind-1.17.2-8 rpm and let me know if it solves your problem. The ypbind-1.17.2-8 rpm can be found here: http://people.redhat.com/cfeist/ypbind/
No response to NEEDINFO since 9/6/06, closing as a dup. *** This bug has been marked as a duplicate of 167061 ***