Default to ON for compression

A simple change, but so many tests break with it,
and those are the majority of this.

Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rich Ercolani <rincebrain@gmail.com>
Closes #13078

contrib/pyzfs/libzfs_core/test/test_libzfs_core.py

@@ -4132,7 +4132,8 @@ class _TempPool(object):
             cachefile = 'none'
         self._zpool_create = [
             'zpool', 'create', '-o', 'cachefile=' + cachefile,
-            '-O', 'mountpoint=legacy', self._pool_name, self._pool_file_path]
+            '-O', 'mountpoint=legacy', '-O', 'compression=off',
+            self._pool_name, self._pool_file_path]
         try:
             os.ftruncate(fd, size)
             os.close(fd)

include/sys/zio.h

@@ -125,7 +125,7 @@ enum zio_checksum {
 #define	ZIO_COMPRESS_LEGACY_ON_VALUE	ZIO_COMPRESS_LZJB
 #define	ZIO_COMPRESS_LZ4_ON_VALUE	ZIO_COMPRESS_LZ4
 
-#define	ZIO_COMPRESS_DEFAULT		ZIO_COMPRESS_OFF
+#define	ZIO_COMPRESS_DEFAULT		ZIO_COMPRESS_ON
 
 #define	BOOTFS_COMPRESS_VALID(compress)			\
 	((compress) == ZIO_COMPRESS_LZJB ||		\

man/man7/zfsprops.7

@@ -784,9 +784,9 @@ Changing this property affects only newly-written data.
 .Xc
 Controls the compression algorithm used for this dataset.
 .Pp
-Setting compression to
+When set to
 .Sy on
-indicates that the current default compression algorithm should be used.
+(the default), indicates that the current default compression algorithm should be used.
 The default balances compression and decompression speed, with compression ratio
 and is expected to work well on a wide variety of workloads.
 Unlike all other settings for this property,

tests/zfs-tests/include/libtest.shlib

@@ -448,7 +448,7 @@ function create_recv_clone
 	datasetexists $recvfs && log_fail "Recv filesystem must not exist."
 	datasetexists $sendfs && log_fail "Send filesystem must not exist."
 
-	log_must zfs create -o mountpoint="$mountpoint" $sendfs
+	log_must zfs create -o compression=off -o mountpoint="$mountpoint" $sendfs
 	log_must zfs snapshot $snap
 	log_must eval "zfs send $snap | zfs recv -u $recvfs"
 	log_must mkfile 1m "$mountpoint/data"
@@ -3567,16 +3567,18 @@ function wait_replacing #pool
 	done
 }
 
-#
 # Wait for a pool to be scrubbed
 #
 # $1 pool name
+# $2 timeout
 #
-function wait_scrubbed
+function wait_scrubbed #pool timeout
 {
+       typeset timeout=${2:-300}
        typeset pool=${1:-$TESTPOOL}
-	while ! is_pool_scrubbed $pool ; do
+       for (( timer = 0; timer < $timeout; timer++ )); do
-		sleep 1
+               is_pool_scrubbed $pool && break;
+               sleep 1;
        done
 }
 
@@ -3671,6 +3673,21 @@ function zed_cleanup
 	rmdir $ZEDLET_DIR
 }
 
+#
+# Check if ZED is currently running; if so, returns PIDs
+#
+function zed_check
+{
+	if ! is_linux; then
+		return
+	fi
+	zedpids="$(pgrep -x zed)"
+#	ret1=$?
+	zedpids2="$(pgrep -x lt-zed)"
+#	ret2=$?
+	echo ${zedpids} ${zedpids2}
+}
+
 #
 # Check if ZED is currently running, if not start ZED.
 #
@@ -3686,9 +3703,14 @@ function zed_start
 	fi
 
 	# Verify the ZED is not already running.
-	pgrep -x zed > /dev/null
+	zedpids=$(zed_check)
-	if (($? == 0)); then
+	if [ -n "$zedpids" ]; then
-		log_note "ZED already running"
+		# We never, ever, really want it to just keep going if zed
+		# is already running - usually this implies our test cases
+		# will break very strangely because whatever we wanted to
+		# configure zed for won't be listening to our changes in the
+		# tmpdir
+		log_fail "ZED already running - ${zedpids}"
 	else
 		log_note "Starting ZED"
 		# run ZED in the background and redirect foreground logging
@@ -3707,13 +3729,13 @@ function zed_start
 function zed_stop
 {
 	if ! is_linux; then
-		return
+		return ""
 	fi
 
 	log_note "Stopping ZED"
 	while true; do
-		zedpids="$(pgrep -x zed)"
+		zedpids=$(zed_check)
-		[ "$?" -ne 0 ] && break
+		[ ! -n "$zedpids" ] && break
 
 		log_must kill $zedpids
 		sleep 1

tests/zfs-tests/tests/functional/cli_root/zfs_copies/zfs_copies_002_pos.ksh

@@ -59,7 +59,7 @@ log_assert "Verify that the space used by multiple copies is charged correctly."
 log_onexit cleanup
 
 for val in 1 2 3; do
-	log_must zfs create -o copies=$val $TESTPOOL/fs_$val
+	log_must zfs create -o compression=off -o copies=$val $TESTPOOL/fs_$val
 
 	log_must mkfile $FILESIZE /$TESTPOOL/fs_$val/$FILE
 done

tests/zfs-tests/tests/functional/cli_root/zfs_destroy/setup.ksh

@@ -30,3 +30,4 @@
 DISK=${DISKS%% *}
 
 default_setup $DISK
+zfs set compression=off $TESTPOOL/$TESTFS

tests/zfs-tests/tests/functional/cli_root/zfs_destroy/zfs_clone_livelist_condense_and_disable.ksh

@@ -39,6 +39,7 @@ function cleanup
 	set_tunable64 LIVELIST_MAX_ENTRIES $ORIGINAL_MAX
 	# reset the minimum percent shared to 75
 	set_tunable32 LIVELIST_MIN_PERCENT_SHARED $ORIGINAL_MIN
+	log_must zfs inherit compression $TESTPOOL
 }
 
 function check_ll_len
@@ -116,6 +117,11 @@ ORIGINAL_MAX=$(get_tunable LIVELIST_MAX_ENTRIES)
 ORIGINAL_MIN=$(get_tunable LIVELIST_MIN_PERCENT_SHARED)
 
 log_onexit cleanup
+# You might think that setting compression=off for $TESTFS1 would be
+# sufficient. You would be mistaken.
+# You need compression=off for whatever the parent of $TESTFS1 is,
+# and $TESTFS1.
+log_must zfs set compression=off $TESTPOOL
 log_must zfs create $TESTPOOL/$TESTFS1
 log_must mkfile 5m /$TESTPOOL/$TESTFS1/atestfile
 log_must zfs snapshot $TESTPOOL/$TESTFS1@snap

tests/zfs-tests/tests/functional/cli_root/zfs_destroy/zfs_clone_livelist_condense_races.ksh

@@ -42,6 +42,7 @@ function cleanup
 	# reset the condense tests to 0
 	set_tunable32 LIVELIST_CONDENSE_ZTHR_PAUSE 0
 	set_tunable32 LIVELIST_CONDENSE_SYNC_PAUSE 0
+	log_must zfs inherit compression $TESTPOOL
 }
 
 function delete_race
@@ -93,6 +94,11 @@ ORIGINAL_MAX=$(get_tunable LIVELIST_MAX_ENTRIES)
 
 log_onexit cleanup
 
+# You might think that setting compression=off for $TESTFS1 would be
+# sufficient. You would be mistaken.
+# You need compression=off for whatever the parent of $TESTFS1 is,
+# and $TESTFS1.
+log_must zfs set compression=off $TESTPOOL
 log_must zfs create $TESTPOOL/$TESTFS1
 log_must mkfile 100m /$TESTPOOL/$TESTFS1/atestfile
 sync_pool $TESTPOOL

tests/zfs-tests/tests/functional/cli_root/zfs_destroy/zfs_clone_livelist_dedup.ksh

@@ -80,9 +80,16 @@ function test_dedup
 ORIGINAL_MIN_SHARED=$(get_tunable LIVELIST_MIN_PERCENT_SHARED)
 
 log_onexit cleanup
+# You might think that setting compression=off for $TESTFS1 would be
+# sufficient. You would be mistaken.
+# You need compression=off for whatever the parent of $TESTFS1 is,
+# and $TESTFS1.
+log_must zfs set compression=off $TESTPOOL
 log_must zfs create $TESTPOOL/$TESTFS1
 log_must mkfile 5m /$TESTPOOL/$TESTFS1/atestfile
 log_must zfs snapshot $TESTPOOL/$TESTFS1@snap
 test_dedup
 
+log_must zfs inherit compression $TESTPOOL
+
 log_pass "Clone's livelist processes dedup blocks as expected."

tests/zfs-tests/tests/functional/cli_root/zfs_destroy/zfs_destroy_clone_livelist.ksh

@@ -42,6 +42,7 @@ function cleanup
 	datasetexists $TESTPOOL/$TESTFS1 && destroy_dataset $TESTPOOL/$TESTFS1 -R
 	# reset the livelist sublist size to its original value
 	set_tunable64 LIVELIST_MAX_ENTRIES $ORIGINAL_MAX
+	log_must zfs inherit compression $TESTPOOL
 }
 
 function clone_write_file
@@ -146,6 +147,11 @@ function test_clone_clone_promote
 ORIGINAL_MAX=$(get_tunable LIVELIST_MAX_ENTRIES)
 
 log_onexit cleanup
+# You might think that setting compression=off for $TESTFS1 would be
+# sufficient. You would be mistaken.
+# You need compression=off for whatever the parent of $TESTFS1 is,
+# and $TESTFS1.
+log_must zfs set compression=off $TESTPOOL
 log_must zfs create $TESTPOOL/$TESTFS1
 log_must mkfile 20m /$TESTPOOL/$TESTFS1/atestfile
 log_must zfs snapshot $TESTPOOL/$TESTFS1@snap

tests/zfs-tests/tests/functional/cli_root/zfs_destroy/zfs_destroy_dev_removal_condense.ksh

@@ -53,7 +53,7 @@ VIRTUAL_DISK2=$TEST_BASE_DIR/disk2
 log_must truncate -s $(($MINVDEVSIZE * 8)) $VIRTUAL_DISK1
 log_must truncate -s $(($MINVDEVSIZE * 16)) $VIRTUAL_DISK2
 
-log_must zpool create $TESTPOOL2 $VIRTUAL_DISK1
+log_must zpool create -O compression=off $TESTPOOL2 $VIRTUAL_DISK1
 log_must poolexists $TESTPOOL2
 
 log_must zfs create $TESTPOOL2/$TESTFS

tests/zfs-tests/tests/functional/cli_root/zfs_receive/receive-o-x_props_aliases.ksh

@@ -129,7 +129,7 @@ log_must eval "check_prop_inherit $destsub copies $dest"
 log_must eval "check_prop_inherit $destsub atime $dest"
 log_must eval "check_prop_inherit $destsub checksum $dest"
 log_must eval "check_prop_source $destsub quota 0 default"
-log_must eval "check_prop_source $destsub compression off default"
+log_must eval "check_prop_source $destsub compression on default"
 # Cleanup
 log_must zfs destroy -r -f $orig
 log_must zfs destroy -r -f $dest

tests/zfs-tests/tests/functional/cli_root/zfs_receive/receive-o-x_props_override.ksh

@@ -135,7 +135,7 @@ log_must eval "check_prop_inherit $destsub atime $dest"
 log_must eval "check_prop_inherit $destsub checksum $dest"
 log_must eval "check_prop_inherit $destsub '$userprop:dest2' $dest"
 log_must eval "check_prop_source $destsub quota 0 default"
-log_must eval "check_prop_source $destsub compression off default"
+log_must eval "check_prop_source $destsub compression on default"
 log_must eval "check_prop_missing $dest '$userprop:orig'"
 log_must eval "check_prop_missing $destsub '$userprop:orig'"
 log_must eval "check_prop_source " \

tests/zfs-tests/tests/functional/cli_root/zfs_rename/setup.ksh

@@ -42,7 +42,8 @@ if [[ -d $TESTDIR2 ]]; then
 		log_unresolved Could not remove $TESTDIR2
 	fi
 fi
-log_must zfs create $TESTPOOL/$DATAFS
+log_must zfs set compression=off $TESTPOOL/$TESTFS
+log_must zfs create -o compression=off $TESTPOOL/$DATAFS
 log_must zfs set mountpoint=$TESTDIR2 $TESTPOOL/$DATAFS
 log_must eval "dd if=$IF of=$OF bs=$BS count=$CNT >/dev/null 2>&1"
 

tests/zfs-tests/tests/functional/cli_root/zpool_trim/zpool_trim_partial.ksh

@@ -64,7 +64,7 @@ log_must set_tunable64 TRIM_EXTENT_BYTES_MIN 4096
 
 log_must mkdir "$TESTDIR"
 log_must truncate -s $LARGESIZE "$LARGEFILE"
-log_must zpool create $TESTPOOL "$LARGEFILE"
+log_must zpool create -O compression=off $TESTPOOL "$LARGEFILE"
 log_must mkfile $(( floor(LARGESIZE * 0.80) )) /$TESTPOOL/file
 sync_all_pools
 

tests/zfs-tests/tests/functional/events/cleanup.ksh

@@ -28,4 +28,6 @@
 
 zed_cleanup all-debug.sh all-syslog.sh all-dumpfds
 
+zed_stop
+
 default_cleanup

tests/zfs-tests/tests/functional/events/events_002_pos.ksh

@@ -62,7 +62,7 @@ log_must truncate -s $MINVDEVSIZE $VDEV1 $VDEV2
 # 1. Create a pool and generate some events.
 log_must truncate -s 0 $ZED_DEBUG_LOG
 log_must zpool events -c
-log_must zpool create $MPOOL mirror $VDEV1 $VDEV2
+log_must zpool create -O compression=off $MPOOL mirror $VDEV1 $VDEV2
 
 # 2. Start the ZED and verify it handles missed events.
 log_must zed_start

tests/zfs-tests/tests/functional/events/zed_fd_spill.ksh

@@ -47,8 +47,14 @@ log_onexit cleanup
 logdir="$(mktemp -d)"
 log_must ln -s "$logdir" /tmp/zts-zed_fd_spill-logdir
 
+
 self="$(readlink -f "$0")"
-log_must ln -s "${self%/*}/zed_fd_spill-zedlet" "${ZEDLET_DIR}/all-dumpfds"
+zedlet="${self%/*}/zed_fd_spill-zedlet"
+log_must ln -s $zedlet "${ZEDLET_DIR}/all-dumpfds"
+
+# zed will cry foul and refuse to run it if this isn't true
+sudo chown root $zedlet
+sudo chmod 700 $zedlet
 
 log_must zpool events -c
 log_must zed_stop

tests/zfs-tests/tests/functional/fault/decrypt_fault.ksh

@@ -40,6 +40,7 @@ function cleanup
 log_onexit cleanup
 
 default_mirror_setup_noexit $DISK1 $DISK2
+log_must zfs set compression=off $TESTPOOL
 log_must eval "echo 'password' | zfs create -o encryption=on \
 	-o keyformat=passphrase -o keylocation=prompt $TESTPOOL/fs"
 mntpt=$(get_prop mountpoint $TESTPOOL/fs)

tests/zfs-tests/tests/functional/inheritance/inherit_001_pos.ksh

@@ -388,12 +388,12 @@ set -A prop "checksum" "" \
 # above must have a corresponding entry in the two arrays below.
 #
 
-set -A def_val "on" "off" "on" \
+set -A def_val "on" "on" "on" \
 	"off" "" \
 	"hidden" \
 	"off"
 
-set -A local_val "off" "on" "off" \
+set -A local_val "off" "off" "off" \
 	"on" "" \
 	"visible" \
 	"off"

tests/zfs-tests/tests/functional/projectquota/setup.ksh

@@ -53,4 +53,6 @@ fi
 DISK=${DISKS%% *}
 default_setup_noexit $DISK
 
+zfs set compression=off $TESTPOOL
+
 log_pass

tests/zfs-tests/tests/functional/quota/setup.ksh

@@ -33,4 +33,6 @@
 
 DISK=${DISKS%% *}
 
-default_container_setup $DISK
+default_setup_noexit $DISK "true"
+log_must zfs set compression=off $TESTPOOL
+log_pass

tests/zfs-tests/tests/functional/redundancy/redundancy.kshlib

@@ -139,7 +139,7 @@ function setup_test_env
 
 	log_must truncate -s $MINVDEVSIZE $vdevs
 
-	log_must zpool create -f -m $TESTDIR $pool $keyword $vdevs
+	log_must zpool create -O compression=off -f -m $TESTDIR $pool $keyword $vdevs
 
 	log_note "Filling up the filesystem ..."
 	typeset -i ret=0

tests/zfs-tests/tests/functional/redundancy/redundancy_draid.ksh

@@ -219,7 +219,7 @@ for nparity in 1 2 3; do
 	raid=draid$nparity
 	dir=$TEST_BASE_DIR
 
-	log_must zpool create -f -o cachefile=none $TESTPOOL $raid ${disks[@]}
+	log_must zpool create -O compression=off -f -o cachefile=none $TESTPOOL $raid ${disks[@]}
 	log_must zfs set primarycache=metadata $TESTPOOL
 
 	log_must zfs create $TESTPOOL/fs

tests/zfs-tests/tests/functional/redundancy/redundancy_draid_damaged.ksh

@@ -85,7 +85,9 @@ function test_sequential_resilver # <pool> <parity> <dir>
 
 	for (( i=0; i<$nparity; i=i+1 )); do
 		spare=draid${nparity}-0-$i
-		log_must zpool replace -fsw $pool $dir/dev-$i $spare
+		zpool status $pool
+		zpool replace -fsw $pool $dir/dev-$i $spare
+		zpool status $pool
 	done
 
 	log_must zpool scrub -w $pool
@@ -128,7 +130,7 @@ for nparity in 1 2 3; do
 	raid=draid${nparity}:${nparity}s
 	dir=$TEST_BASE_DIR
 
-	log_must zpool create -f -o cachefile=none $TESTPOOL $raid ${disks[@]}
+	log_must zpool create -O compression=off -f -o cachefile=none $TESTPOOL $raid ${disks[@]}
 	log_must zfs set primarycache=metadata $TESTPOOL
 
 	log_must zfs create $TESTPOOL/fs

tests/zfs-tests/tests/functional/redundancy/redundancy_raidz.ksh

@@ -219,7 +219,7 @@ for nparity in 1 2 3; do
 	raid=raidz$nparity
 	dir=$TEST_BASE_DIR
 
-	log_must zpool create -f -o cachefile=none $TESTPOOL $raid ${disks[@]}
+	log_must zpool create -O compression=off -f -o cachefile=none $TESTPOOL $raid ${disks[@]}
 	log_must zfs set primarycache=metadata $TESTPOOL
 
 	log_must zfs create $TESTPOOL/fs

tests/zfs-tests/tests/functional/refquota/setup.ksh

@@ -33,4 +33,6 @@
 
 verify_runnable "both"
 DISK=${DISKS%% *}
-default_setup $DISK
+default_setup_noexit $DISK
+log_must zfs set compression=off $TESTPOOL
+log_pass

tests/zfs-tests/tests/functional/refreserv/refreserv_raidz.ksh

@@ -98,7 +98,7 @@ for parity in 1 2 3; do
 			continue
 		fi
 
-		log_must zpool create "$TESTPOOL" "$raid" "${disks[@]}"
+		log_must zpool create -O compression=off "$TESTPOOL" "$raid" "${disks[@]}"
 
 		for bits in "${allshifts[@]}"; do
 			vbs=$((1 << bits))

tests/zfs-tests/tests/functional/refreserv/setup.ksh

@@ -33,4 +33,6 @@
 
 verify_runnable "both"
 DISK=${DISKS%% *}
-default_setup $DISK
+default_setup_noexit $DISK
+log_must zfs set compression=off $TESTPOOL
+log_pass

tests/zfs-tests/tests/functional/removal/removal_with_errors.ksh

@@ -78,6 +78,7 @@ function wait_for_removing_cancel
 
 default_setup_noexit "mirror $DISK0 $DISK1 mirror $DISK2 $DISK3"
 log_onexit cleanup
+log_must zfs set compression=off $TESTPOOL
 
 FILE_CONTENTS="Leeloo Dallas mul-ti-pass."
 

tests/zfs-tests/tests/functional/removal/removal_with_export.ksh

@@ -22,6 +22,7 @@
 . $STF_SUITE/tests/functional/removal/removal.kshlib
 
 default_setup_noexit "$DISKS"
+log_must zfs set compression=off $TESTPOOL
 log_onexit default_cleanup_noexit
 
 function callback

tests/zfs-tests/tests/functional/reservation/setup.ksh

@@ -32,4 +32,5 @@
 . $STF_SUITE/include/libtest.shlib
 
 default_setup_noexit ${DISKS%% *}
+zfs set compression=off $TESTPOOL
 log_pass

tests/zfs-tests/tests/functional/rsend/rsend_009_pos.ksh

@@ -60,8 +60,8 @@ log_onexit cleanup
 
 log_must mkfile $MINVDEVSIZE $TESTDIR/bfile
 log_must mkfile $SPA_MINDEVSIZE  $TESTDIR/sfile
-log_must zpool create bpool $TESTDIR/bfile
+log_must zpool create -O compression=off bpool $TESTDIR/bfile
-log_must zpool create spool $TESTDIR/sfile
+log_must zpool create -O compression=off spool $TESTDIR/sfile
 
 #
 # Test out of space on sub-filesystem

tests/zfs-tests/tests/functional/rsend/send_partial_dataset.ksh

@@ -46,7 +46,7 @@ function cleanup
 }
 log_onexit cleanup
 
-log_must zfs create $POOL/testfs2
+log_must zfs create -o compression=off $POOL/testfs2
 log_must zfs create $POOL/stream
 mntpnt=$(get_prop mountpoint $POOL/testfs2)
 

tests/zfs-tests/tests/functional/snapshot/snapshot_014_pos.ksh

@@ -56,11 +56,14 @@ function cleanup
 	datasetexists $TESTPOOL/$TESTCTR/$TESTFS1 && \
 		log_must zfs set quota=none $TESTPOOL/$TESTCTR/$TESTFS1
 
+	zfs inherit compression $TESTPOOL
 }
 
 log_assert "Verify creating/destroying snapshots do things clean"
 log_onexit cleanup
 
+log_must zfs set compression=off $TESTPOOL
+
 log_must zfs set quota=$FSQUOTA $TESTPOOL/$TESTCTR/$TESTFS1
 log_must mkfile $FILESIZE $TESTDIR1/$TESTFILE
 

tests/zfs-tests/tests/functional/userquota/setup.ksh

@@ -53,4 +53,6 @@ if [ $? -ne 0 ]; then
 fi
 
 DISK=${DISKS%% *}
-default_setup $DISK
+default_setup_noexit $DISK
+log_must zfs set compression=off $TESTPOOL
+log_pass

tests/zfs-tests/tests/functional/userquota/userspace_send_encrypted.ksh

@@ -52,7 +52,7 @@ log_assert "Sending raw encrypted datasets back to the source dataset succeeds."
 
 # Setup pool and create source
 truncate -s 200m $FILEDEV
-log_must zpool create -o feature@encryption=enabled $POOLNAME \
+log_must zpool create -O compression=off -o feature@encryption=enabled $POOLNAME \
 	$FILEDEV
 log_must eval "echo 'password' | zfs create -o encryption=on" \
 	"-o keyformat=passphrase -o keylocation=prompt " \

tests/zfs-tests/tests/functional/write_dirs/setup.ksh

@@ -35,4 +35,6 @@ verify_runnable "global"
 
 DISK=$TEST_BASE_DIR/disk0
 truncate -s 2G $DISK
-default_setup $DISK
+default_setup_noexit $DISK
+log_must zfs set compression=off $TESTPOOL
+log_pass

tests/zfs-tests/tests/functional/xattr/setup.ksh

@@ -52,4 +52,6 @@ echo $ZFS_USER > $TEST_BASE_DIR/zfs-xattr-test-user.txt
 echo $USES_NIS > $TEST_BASE_DIR/zfs-xattr-test-nis.txt
 
 DISK=${DISKS%% *}
-default_setup $DISK
+default_setup_noexit $DISK
+log_must zfs set compression=off $TESTPOOL
+log_pass "Setup complete"