#13: Network Failure Recovery Script Update
Date: 3-29-2005
Security Status: N/A
Bugs: Repeated network
link failures were sometimes leaving zombie processes from the
automatic recovery scripts.
Explanation:
When network segments on the Phalanx
become unreachable,
there is an automatic procedure to reconnect them or an alternate
connection to replace them. During this process some connection
types were leaving stale processes on the system. This was not
causing problems, but theoretically if a connection bounces enough
times it could fill the process table with junk. This would take
many days of untreated problems, and would result in the watchdog
restarting the Phalanx.
Files:
/usr/local/sbin/netwatch
back
to list