suspend_resume_single: clear pool errors on fail

robn · web-flow · commit 523e3adac9c0 · 2025-02-23T15:22:00.000-05:00
If the timing is unfortunate, the pool can suspend just as we're failing because it didn't suspend. If we don't resume the pool, we hang trying to destroy it. Sponsored-by: Klara, Inc. Sponsored-by: Wasabi Technology, Inc. Reviewed-by: Alexander Motin <mav@FreeBSD.org> Signed-off-by: Rob Norris <rob.norris@klarasystems.com> Closes #17054
diff --git a/tests/zfs-tests/tests/functional/fault/suspend_resume_single.ksh b/tests/zfs-tests/tests/functional/fault/suspend_resume_single.ksh
@@ -30,6 +30,7 @@ DATAFILE="$TMPDIR/datafile"
 
 function cleanup
 {
+	zpool clear $TESTPOOL
 	destroy_pool $TESTPOOL
 	unload_scsi_debug
 	rm -f $DATA_FILE

Original file line number	Diff line number	Diff line change
`@@ -30,6 +30,7 @@ DATAFILE="$TMPDIR/datafile"`
`30`	`30`
`31`	`31`	`function cleanup`
`32`	`32`	`{`
	`33`	`+ zpool clear $TESTPOOL`
`33`	`34`	`destroy_pool $TESTPOOL`
`34`	`35`	`unload_scsi_debug`
`35`	`36`	`rm -f $DATA_FILE`