Bug 27457

When updating the NIS databases on the slaves by running yppush
in /var/yp on the server, the clients get updated properly. However,
the calling yppush hangs:

root@server:/var/yp>yppush -vv -h spoolfix -h printfix amd.host 
yppush: initiating transfer: amd.host -> pr (transid = 990293135)
yppush: pr has been called
yppush: checking return status: transaction ID: 990293135
yppush: transfer of map amd.host to server pr failed
yppush: status returned by ypxfr: Master's version not newer
yppush: initiating transfer: amd.host -> spoolfix (transid = 990293136)
yppush: spoolfix has been called
yppush: all jobs dispatched
yppush: spoolfix has not responded
yppush: pr has responded
yppush: 1 transfer still pending

[wait 1 to 2 minutes here]

yppush: timed out
yppush: warning: exiting with transfer to spoolfix (transid = 990293136) still pending


In the logfile of one client we see:
May 19 19:25:43 <daemon.notice> spoolfix ypxfr[4428]: callback failed: RPC: Timed out
May 19 19:26:15 <daemon.notice> spoolfix ypxfr[4429]: callback failed: RPC: Timed out

It seems that somehow the callback of the second slave doesn't
make it through to the server.

The problem already appeared as described in PR# 12496. It then
went away until the recent changes to libexec/rtld-elf/rtld.c
made appear it again.

Fix: The following patch is no fix. It circumvents the problem but
I can't explain why :-)
How-To-Repeat: 
Use the newest FreeBSD 4.3-STABLE sources. Put two slave servers into
ypservers. Run yppush as above (with any map).