A management node can end up with a web server that is not the one xcatconfig configured -- a stock nginx holding port 80 in front of the httpd whose xcat.conf carries the aliases. It answers 404 to every path a node is ever given, so P-22's kernel fetch failed in a way that read like a broken boot script, and the whole HTTP stage would have failed nine times for one reason. The fixture now probes the aliased directories before asserting against them. P-21 and P-22 are separate scenarios, so the script is still asserted where the kernel cannot be, and the HTTP stage skips with the reason rather than failing.
xCAT-test
Unit tests that run from the source checkout are split by implementation language:
| Test type | Location | Runner |
|---|---|---|
| Perl unit tests | xCAT-test/unit/*.t |
prove -r xCAT-test/unit |
| Shell unit tests | xCAT-test/bats/*.bats |
bats -r xCAT-test/bats |
| DHCP wire tests | xCAT-test/dhcptest/ |
python3 -m unittest discover -s tests, then dhcptest run -i <nic> <conf> |
| CLI functional tests | xCAT-test/autotest/testcase/ and xCAT-test/autotest/bundle/ |
xcattest -f <cluster.conf> -t <case> or xcattest -f <cluster.conf> -b <bundle> |
Use Perl .t tests for Perl modules, Perl scripts, templates, and repository
artifacts. Use BATS tests for shell-script behavior that can be exercised from
the checkout by sourcing a shell library or script and shadowing external
commands.
Shell behavior should not be tested by Perl tests that grep shell source. Put
those tests under xCAT-test/bats instead.
dhcptest is the odd one out: its Python unit tests and dhcptest validate
run offline like the others, but its point is the third column -- driving real
DHCP transactions on a provisioning NIC and asserting on what came back. It
talks to the server only over the wire, so it is agnostic to which DHCP
implementation is answering and to xCAT itself.
See unit/README.md, bats/README.md and dhcptest/README.md for the
detailed rules for each suite.