The "check key refreshes are resumed after root servers become
available" check may trigger a false positive for the "mkeys" system
test if the second example/TXT query sent by dig is received by ns5 less
than a second after it receives a REFUSED response to the upstream query
it sends to ns1 in order to resolve the first example/TXT query sent by
dig. Since that REFUSED response from ns1 causes ns5 to return a
SERVFAIL answer to dig, example/TXT is added to the SERVFAIL cache,
which is enabled by default with a TTL of 1 second. This in turn may
cause ns5 to return a cached SERVFAIL response to the second example/TXT
query sent by dig, i.e. make ns5 not perform full query processing as
expected by the check.
Since the primary purpose of the check in question is to ensure that key
refreshes are resumed once initially unavailable root servers become
available, the optimal solution appears to be disabling SERVFAIL cache
for ns5 as doing that still allows the check to fulfill its purpose and
it is arguably more prudent than always sleeping for 1 second.
For consistency between all system tests, add missing setup.sh scripts
for tests which do not have one yet and ensure every setup.sh script
calls its respective clean.sh script.
Temporary files created by a given system test should be removed by its
clean.sh script, not its setup.sh script. Remove redundant "rm"
invocations from setup.sh scripts. Move required "rm" invocations from
setup.sh scripts to their corresponding clean.sh scripts.
If dots are not escaped in the "1.2.3.4" regular expressions used for
checking whether IP address 1.2.3.4 is present in the tested resolver's
answers, a COOKIE that matches such a regular expression will trigger a
false positive for the "resolver" system test. Properly escape dots in
the aforementioned regular expressions to prevent that from happening.
Including $SYSTEMTESTTOP/conf.sh from a system test's clean.sh script is
not needed for anything while it causes an error message to be printed
out when "./configure" is run, as "make clean" is invoked at the end.
Remove the offending line to prevent the error from occurring.
For all system tests utilizing named instances, call clean.sh from each
test's setup.sh script in a consistent way to make sure running the same
system test multiple times using run.sh does not trigger false positives
caused by stale files created by previous runs.
Ideally we would just call clean.sh from run.sh, but that would break
some quirky system tests like "rpz" or "rpzrecurse" and being consistent
for the time being does not hurt.
These tests check if a key with an unsupported algorithm in
managed-keys is ignored and when seeing an algorithm rollover to
an unsupported algorithm, the new key will be ignored too.
Illustrate the syntax for the policy options, with semicolons.
Explicitly mention the "default" policy.
Fix a few typos and remove some redundant wording.
When a mirror zone is verified, the 'ignore_kskflag' argument passed to
dns_zoneverify_dnssec() is set to false. This means that in order for
its verification to succeed, a mirror zone needs to have at least one
key with the SEP bit set configured as a trust anchor. This brings no
security benefit and prevents zones signed only using keys without the
SEP bit set from being mirrored, so change the value of the
'ignore_kskflag' argument passed to dns_zoneverify_dnssec() to true.
The "mirror" system test checks whether log messages announcing a mirror
zone coming into effect are emitted properly. However, the helper
functions responsible for waiting for zone transfers and zone loading to
complete do not wait for these exact log messages, but rather for other
ones preceding them, which introduces a possibility of false positives.
This problem cannot be addressed by just changing the log message to
look for because the test still needs to discern between transferring a
zone and loading a zone.
Add two new log messages at debug level 99 (which is what named
instances used in system tests are configured with) that are to be
emitted after the log messages announcing a mirror zone coming into
effect. Tweak the aforementioned helper functions to only return once
the log messages they originally looked for are followed by the newly
added log messages. This reliably prevents races when looking for
"mirror zone is now in use" log messages and also enables a workaround
previously put into place in the "mirror" system test to be reverted.
In the "mirror" system test, ns3 periodically sends trust anchor
telemetry queries to ns1 and ns2. It may thus happen that for some
non-recursive queries for names inside mirror zones which are not yet
loaded, ns3 will be able to synthesize a negative answer from the cached
records it obtained from trust anchor telemetry responses. In such
cases, NXDOMAIN responses will be sent with the root zone SOA in the
AUTHORITY section. Since the root zone used in the "mirror" system test
has the same serial number as ns2/verify.db.in and zone verification
checks look for the specified serial numbers anywhere in the answer, the
test could be broken if different zone names were used.
The +noauth dig option could be used to address this weakness, but that
would prevent entire responses from being stored for later inspection,
which in turn would hamper troubleshooting test failures. Instead, use
a different serial number for ns2/verify.db.in than for any other zone
used in the "mirror" system test and check the number of records in the
ANSWER section of each response.
Due to the way the "mirror" system test is set up, it is impossible for
the "verify-unsigned" and "verify-untrusted" zones to contain any serial
number other than the original one present in ns2/verify.db.in. Thus,
using presence of a different serial number in the SOA records of these
zones as an indicator of problems with mirror zone verification is
wrong. Look for the original zone serial number instead as that is the
one that will be returned by ns3 if one of the aforementioned zones is
successfully verified.
* Alphabetize the option lists in the man page and help text
* Make the synopses more consistent between the man page and help
text, in particular the number of different modes
* Group mutually exclusive options in the man page synopses, and order
options so that it is more clear which are available in every mode
* Expand the DESCRIPTION to provide an overview of the output modes
and input modes
* Improve cross-references between options
* Leave RFC citations to the SEE ALSO section, and clarify which RFC
specifies what
* Clarify list of digest algorithms in dnssec-dsfromkey and dnssec-cds
man pages
Ensure IXFR statistics are calculated correctly by dig and named, both
for incoming and outgoing transfers. Disable EDNS when using dig to
request an IXFR so that the same reference file can be used for testing
statistics calculated by both dig and named (dig uses EDNS by default
when sending transfer requests, which affects the number of bytes
transferred).
Ensure AXFR statistics are calculated correctly by dig and named, both
for incoming and outgoing transfers. Rather than employing a zone which
is already used in the "xfer" system test, create a new one whose AXFR
form spans multiple TCP messages. Disable EDNS when using dig to
request an AXFR so that the same reference file can be used for testing
statistics calculated by both dig and named (dig uses EDNS by default
when sending transfer requests, which affects the number of bytes
transferred).
the occluded-key test creates both a KEY and a DNSKEY. the second
call to dnssec-keygen calls dns_dnssec_findmatchingkeys(), which causes
a spurious warning to be printed when it sees the type KEY record.
this should be fixed in dnssec.c, but the meantime this change silences
the warning by reversing the order in which the keys are created.
