ksbsrv: No startup acknowledgement from forked process after 30 seconds

Applies to:

Oracle Server - Enterprise Edition - Version: 8.1.7.4 to 11.1.0.6 - Release: 8.1.7 to 11.1
Information in this document applies to any platform.

Goal

From time to time, messages like:
ksbsrv: No startup acknowledgement from forked process after 30 seconds
appear in the database. The timeout can have various values.

Solution

1. When the message appears at startup and the database instance cannot be started, the solution is to remove the .blocked files that reside in the $ORACLE_HOME/database directories.
2. If the message is left in the alert log by a starting process (in most situations this is done by a slave process, such as jxxx or by a shared server process):
Basically, this means that a process has started but it didn't provide an acknowledgment for the success of the operation. If this ack signal is not given, the process dies after 30 (or other interval, depending on the process type and oracle release) seconds.
The process did not hit a resource limit at startup, because when that happens, the message is different, ie: "skgpspawn failed category".
This actually means that the process was delayed in delivering the acknowledge message due to a performance problem, such as high CPU usage.  In most situations the message would just not reoccur, since the performance problem was particular to that moment.
When the message still reoccurs, to improve the situation one could improve the overall performance and try to detect the bottleneck.
+ increase the available memory for the oracle process by enabling the 3 GB switch
+ tune the system IO
+ monitor the CPU usage and find out the thread that is causing the situation (use the qslice tool for that).
Generally, this is not a reason for concern when it's only sporadically occurring.
However, when occurring too often, it is recommended to monitor and tune the system, as the cause of the problem is slow performance, most likely due to CPU overloading.