Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Monitoring System Selection

IP.com Disclosure Number: IPCOM000102593D
Original Publication Date: 1990-Dec-01
Included in the Prior Art Database: 2005-Mar-17
Document File: 2 page(s) / 79K

Publishing Venue

IBM

Related People

Swanson, MD: AUTHOR

Abstract

In a multisystem configuration, the status of participating systems must be monitored. Determining which system will monitor which other systems must be performance-sensitive, flexible, tolerate temporarily inoperative systems, and insure that no more than one system takes recovery action for a failed system.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Monitoring System Selection

       In a multisystem configuration, the status of
participating systems must be monitored.  Determining which system
will monitor which other systems must be performance-sensitive,
flexible, tolerate temporarily inoperative systems, and insure that
no more than one system takes recovery action for a failed system.

      A protocol is described for determining which system(s) are to
be monitored by an active system.  The protocol allows for dynamic
changes in the number of systems participating in the configuration
and temporary outages by participating systems.

      The proposed protocol supports a dynamically changing
configuration while enabling each participating system to
independently determine which other systems are to be monitored.
This independent operation insures a given system is always monitored
while still insuring that in most environments only one system is
responsible for monitoring a given system.  By having only one system
monitor any other system at most times, the performance impact of the
system monitor protocol is minimized.

      The system monitor routine executes as a disabled exit (DIE)
which is given control on a one-second interval.  The DIE routine is
responsible for monitoring other systems in the configuration.  The
beginning and end of the loop for monitoring systems is controlled by
a table as shown below. This table contains one entry for every
possible system in the configuration.  Entries are numbered
sequentially from zero and correspond to system numbers.  A system is
assigned a number when it is initialized into the configuration.
Entries contain information for:
      -    defined or not defined in the configuration
      -    a defined system is in a temporarily stopped
           (pause) state
sysid          sys     status
SYSA           0         Defined
1         Not Defined
SYSB           2         Defined
SYSC           3         Defined/Pause
4         Not Defined
SYSD           5         Defined
                6         Not Defined
                7         Not Defined
Note:  System Monitor Status Table

      Using the above table, the protocol described determines that:
           -    SYSA monitors SYSB
           -    SYSB monitors SYSC
           -    SYSB monitors SYSD...