fix(testing): keep the bare-metal runner past a failing Testinfra sweep

run-test-baremetal.sh runs under set -e, and the P-D migration left run_testinfra_validation called bare. It returns pytest's rc, so a failing sweep aborted the script before it wrote the report or the TEST FAILED summary. The old run_all_validations returned 0 and never tripped set -e. Mirror the VM runner: set +e around the call, capture the rc, and derive TEST_PASSED from it so the report and summary always run.

1 files changed, 7 insertions, 4 deletions

diff --git a/scripts/testing/run-test-baremetal.sh b/scripts/testing/run-test-baremetal.sh
index d040768..d22c424 100755
--- a/scripts/testing/run-test-baremetal.sh
+++ b/scripts/testing/run-test-baremetal.sh
@@ -265,9 +265,12 @@ fi
 
 # Run validations. Testinfra is the authoritative validator (same as the VM
 # runner); its ZFS-conditional pytest checks cover what validate_zfs_services
-# used to. It surfaces pass/fail through the VALIDATION_* counters the
-# TEST_PASSED check below reads, and connects over the key authorized above.
+# used to, and it connects over the key authorized above. set +e because it
+# returns pytest's rc (non-zero on failures) and the report + summary below must
+# still run; the verdict is carried by testinfra_rc and the exit code at the end.
+set +e
 run_testinfra_validation "$TEST_RESULTS_DIR"
+testinfra_rc=$?
 
 # Analyze logs if we ran archsetup
 if ! $VALIDATE_ONLY; then
@@ -277,8 +280,8 @@ fi
 # Generate reports
 generate_issue_report "$TEST_RESULTS_DIR" "$ARCHZFS_INBOX"
 
-# Set validation result (TEST_PASSED is the boolean; VALIDATION_PASSED stays the counter)
-if [ "$VALIDATION_FAILED" -eq 0 ]; then
+# The run passes only if the Testinfra sweep passed.
+if [ "$testinfra_rc" -eq 0 ]; then
     TEST_PASSED=true
 else
     TEST_PASSED=false