Explanation
This is a multi-line message that is issued to
report statistical information about an unresponsive name server,
or about a name server that was previously unresponsive but is now
considered to be responsive. The statistics represent data that is
specific to the name server. The data was accumulated by the resolver
during the most recent monitoring interval.
In the message
text:
- ipaddress
- The IPv4 or IPv6 network address of the name server.
- totalsent
- The count of resolver queries that were directed to the name server
during the most recent monitoring interval.
- totalfail
- The count of instances when the name server did not respond to
resolver queries that were directed to the name server during the
most recent monitoring interval.
- percent
- The percentage of resolver queries to which the name server did
not respond during the most recent monitoring interval.
System action
Processing continues. The resolver resets the
statistics to zero before starting the next monitor interval.
Operator response
Contact the system programmer.
System programmer response
If the statistics were reported when
a name server became responsive again, compare the
percent value
to the unresponsiveness threshold value.
- If the failure rate is close to the threshold, but your network
is operating at a satisfactory level, increase the UNRESPONSIVETHRESHOLD
setting to eliminate unnecessary notifications that the name server
is unresponsive at this threshold level.
- If the totalsent value is zero, then the name
server might be responsive only because no requests were sent to the
name server during the most recent monitoring interval. There might
still be issues with the name server, so monitor for any new EZZ9308E
messages that ndicate that this name server is unresponsive again.
If the statistics that were reported are for an unresponsive
name server and they are similar to previous interval statistics
for this name server, or if they represent an improvement in the
responsiveness of the name server, continue monitoring the statistics.
If the statistics that were reported are for an unresponsive
name server and they are significantly worse than previous intervals,
determine the reason for the change.
- If the name server is unresponsive due to a network condition
that prevents a larger percentage of resolver requests or name server
responses from reaching the correct destination, correct the network
condition. When requests can successfully reach the name server and
are being responded to, the resolver will detect the change in responsiveness
at the next monitor interval and will clear the message from the
operator console. The resolver will also issue message EZZ9309I to
indicate that the server is now responsive.
- If you eliminated a network condition as the reason for the message,
determine if the name server is unresponsive due to a configuration
error. Use resolver diagnostic tools such as MODIFY RESOLVER,DISPLAY
or Trace Resolver output to determine which TCPIP.DATA data sets are
being used in the failing resolver queries. Investigate the following
are possibilities:
- If the UNRESPONSIVETHRESHOLD value specifies a lower percentage
of errors than your network can tolerate during typical operations,
increase the threshold percentage in the resolver setup file. For
example, if the threshold percentage is 25%, but your network can
tolerate 40%, increase the threshold to 40%-45%.
- If the IP address is no longer valid as a destination name server,
remove the IP address from the list of name servers that were coded
by using the NSINTERADDR or NAMESERVER configuration statements.
- If the RESOLVERTIMEOUT value is so small that the response is
unable to return from the name server within the specified time value,
increase the RESOLVERTIMEOUT setting to a value that permits a larger
percentage of responses to arrive within the timeout window.
When the configuration error is corrected, have the operator
do one of the following to cause the resolver to use the updated configuration
information:
- If you modified the setting of UNRESPONSIVETHRESHOLD, the operator
should issue the MODIFY RESOLVER,REFRESH,SETUP=setup_file_name
command.
- If you modified RESOLVERTIMEOUT, or you removed an IP address
from the list of name servers that were coded by using NSINTERADDR
or NAMESERVER statement, the operator should issue the MODIFY
RESOLVER,REFRESH command.
- If you eliminated network conditions or configuration errors as
the reason for the message, then the resolver might be generating
the message for a temporary condition that might resolve itself. For
example, the name server might be having maintenance applied, or the
name server might have a very high percentage of failures because
there were few queries sent to the name server during the monitoring
interval, so even a short network interruption would severely impact
the calculations. If this situation repeats itself, an UNRESPONSIVETHRESHOLD
value that specifies a lower percentage of errors than your network
can tolerate during typical operations might be contributing to the
situation. Consider increasing the setting for UNRESPONSIVETHRESHOLD
in the resolver setup file, and then have the operator issue the MODIFY
RESOLVER,REFRESH,SETUP=setup_file_name command
to make the resolver less sensitive to name server response failures.
User response
Problem determination
Source
z/OS® Communications
Server TCP/IP: System Resolver
Module
Routing code
Descriptor code
Automation
This message is a good candidate for automation
if you are using automation to monitor message EZZ9308E to collect
statistics on name server responsiveness.
Example
EZZ9310I NAME SERVER 10.45.23.200
TOTAL NUMBER OF QUERIES SENT 6574
TOTAL NUMBER OF FAILURES 2957
PERCENTAGE 45%