If the resolver detects that a name server is not being responsive,
a series of network operator messages is issued that relate to that
name server. For example, if a name server is operating at IP address
10.42.35.200 and the UNRESPONSIVETHRESHOLD value is 100, then the
following sequence of messages might be generated by the resolver:
- At the end of a 30-second monitoring interval, the resolver determines
that the name server failed to respond to 100% of the 500 queries
that were attempted by the resolver. The resolver considers the name
server to be unresponsive and issues the following messages:
EZZ9311E STOPPED USING NAME SERVER AT IP ADDRESS 10.42.35.200
EZZ9313I NAME SERVER 10.42.35.200
TOTAL NUMBER OF QUERIES SENT 500
TOTAL NUMBER OF FAILURES 500
TOTAL NUMBER OF RESOLVER POLLS SENT 0
TOTAL NUMBER OF POLL FAILURES 0
PERCENTAGE 100%
After issuing
this message, the resolver sends only DNS polling queries to the unresponsive
name server.
- At the end of a subsequent monitoring interval, the resolver determines
that the name server failed to respond to 10% of the 10 or more DNS
polling queries during the last two monitoring intervals. This percentage
is below the 100% threshold value, so the resolver considers this
name server to be responsive again. The resolver clears message EZZ9311E
from the operator console and issues the following message:
EZZ9312I RESUMED USING NAME SERVER AT IP ADDRESS 10.42.35.200
- After the resolver issues message EZZ9312I, it resumes sending
DNS queries that are generated by an application to the name server.
- The resolver also clears message EZZ9311E from the operator console
if any of the following events occur:
- The network operator disables the monitoring function using the
MODIFY RESOLVER,REFRESH,SETUP command. For more information, see Steps for modifying the UNRESPONSIVETHRESHOLD value.
- The network operator enables the network operator notification
function using the MODIFY RESOLVER,REFRESH,SETUP command. For more
information, see Steps for modifying the UNRESPONSIVETHRESHOLD value.
- The autonomic quiescing of unresponsive name servers function
abnormally terminates because of a system error.
- The resolver is stopped.
The resolver issues individual messages for each name server that
it considers to be unresponsive at the end of a given monitoring interval.
Because the resolver calculates the responsiveness of a name server
at every interval, messages might be issued for different name servers
at different times, rather than notifications being issued about all
unresponsive name servers at the same time. The resolver does not
periodically reissue message EZZ9313I for each name server, but the
status and current failure rate for each name server specified on
an NSINTERADDR statement in the global TCPIP.DATA file is included
in the MODIFY RESOLVER,DISPLAY command output when the autonomic quiescing
function is active. The failure rate that is displayed is the percentage
calculated by the resolver for the individual name server at the end
of the last 30-second monitoring interval. For more information about
the MODIFY command for the resolver address
space, see z/OS Communications Server: IP System Administrator's
Commands.
Guideline: In most cases, the resolver
requires a sample size of at least 10 DNS queries before it declares
a name server to be either unresponsive or responsive. For example,
if only one query is received by a particular name server during a
30-second monitoring interval, and that query fails to obtain a response,
then the resolver does not consider that name server to be unresponsive,
even though the failure rate is 100% and exceeds the UNRESPONSIVETHRESHOLD
value. The resolver sends DNS polling queries to this name server
to collect a sufficient sample size of responses. In subsequent intervals,
the resolver uses the results from DNS queries that are generated
by an application and from DNS polling queries to calculate the overall
failure rate for the name server.