From 0502383024cbf7e4776816e0a992dccc484a3cf2 Mon Sep 17 00:00:00 2001

From: Ravishankar N <ravishankar@redhat.com>

Date: Tue, 8 Dec 2020 17:23:22 +0530

Subject: [PATCH 488/511] glusterd/afr: enable granular-entry-heal by default

XXXXXXXXXXXXXXXXXXX

    IMPORTANT:

XXXXXXXXXXXXXXXXXXXX

I see that for rhgs-3.5.3, GD_OP_VERSION_MAX is GD_OP_VERSION_7_0. Since

this patch should only act on new volumes in rhgs-3.5.4, I am bumping

the op-version to GD_OP_VERSION_7_1. In glusterfs upstream, the patch

acts only if op-version >= GD_OP_VERSION_9_0 as seen in the commit

messae below.

Upstream patch details:

/------------------------------------------------------------------------------/

1. The option has been enabled and tested for quite some time now in RHHI-V

downstream and I think it is safe to make it 'on' by default. Since it

is not possible to simply change it from 'off' to 'on' without breaking

rolling upgrades, old clients etc., I have made it default only for new volumes

starting from op-verison GD_OP_VERSION_9_0.

Note: If you do a volume reset, the option will be turned back off.

This is okay as the dir's gfid will be captured in 'xattrop' folder  and heals

will proceed. There might be stale entries inside entry-changes' folder,

which will be removed when we enable the option again.

2. I encountered a cust. issue where entry heal was pending on a dir. with

236436 files in it and the glustershd.log output was just stuck at

"performing entry selfheal", so I have added logs to give us

more info in DEBUG level about whether entry heal and data heal are

progressing (metadata heal doesn't take much time). That way, we have a

quick visual indication to say things are not 'stuck' if we briefly

enable debug logs, instead of taking statedumps or checking profile info

etc.

>Fixes: #1483

>Change-Id: I4f116f8c92f8cd33f209b758ff14f3c7e1981422

>Signed-off-by: Ravishankar N <ravishankar@redhat.com>

Upstream Patch: https://github.com/gluster/glusterfs/pull/1621

/------------------------------------------------------------------------------/

BUG: 1890506

Change-Id: If449a1e873633616cfc508d74b5c22eb434b55ae

Signed-off-by: Ravishankar N <ravishankar@redhat.com>

Reviewed-on: https://code.engineering.redhat.com/gerrit/220555

Tested-by: Sunil Kumar Heggodu Gopala Acharya <sheggodu@redhat.com>

Reviewed-by: Sunil Kumar Heggodu Gopala Acharya <sheggodu@redhat.com>

---

 libglusterfs/src/glusterfs/globals.h               |   4 +-

 libglusterfs/src/syncop-utils.c                    |   4 +-

 tests/basic/afr/add-brick-self-heal-non-granular.t |  75 +++++++++++++

 tests/basic/afr/add-brick-self-heal.t              |   4 +-

 tests/basic/afr/bug-1130892-non-granular.t         |  77 ++++++++++++++

 .../basic/afr/bug-1493415-gfid-heal-non-granular.t |  79 ++++++++++++++

 ...507-type-mismatch-error-handling-non-granular.t | 117 +++++++++++++++++++++

 ...1749322-entry-heal-not-happening-non-granular.t |  90 ++++++++++++++++

 .../afr/replace-brick-self-heal-non-granular.t     |  65 ++++++++++++

 tests/basic/afr/replace-brick-self-heal.t          |   2 +-

 tests/bugs/replicate/bug-1130892.t                 |   2 +-

 tests/bugs/replicate/bug-1493415-gfid-heal.t       |   2 +-

 .../bug-1722507-type-mismatch-error-handling.t     |  26 +++--

 .../bug-1749322-entry-heal-not-happening.t         |   7 +-

 xlators/cluster/afr/src/afr-self-heal-common.c     |   5 +

 xlators/cluster/afr/src/afr-self-heal-data.c       |   3 +

 xlators/cluster/afr/src/afr-self-heal-entry.c      |   7 +-

 xlators/mgmt/glusterd/src/glusterd-utils.c         |  13 +++

 18 files changed, 558 insertions(+), 24 deletions(-)

 create mode 100644 tests/basic/afr/add-brick-self-heal-non-granular.t

 create mode 100644 tests/basic/afr/bug-1130892-non-granular.t

 create mode 100644 tests/basic/afr/bug-1493415-gfid-heal-non-granular.t

 create mode 100644 tests/basic/afr/bug-1722507-type-mismatch-error-handling-non-granular.t

 create mode 100644 tests/basic/afr/bug-1749322-entry-heal-not-happening-non-granular.t

 create mode 100644 tests/basic/afr/replace-brick-self-heal-non-granular.t

diff --git a/libglusterfs/src/glusterfs/globals.h b/libglusterfs/src/glusterfs/globals.h

index 31717ed..cc145cd 100644

--- a/libglusterfs/src/glusterfs/globals.h

+++ b/libglusterfs/src/glusterfs/globals.h

@@ -50,7 +50,7 @@

     1 /* MIN is the fresh start op-version, mostly                             \

          should not change */

 #define GD_OP_VERSION_MAX                                                      \

-    GD_OP_VERSION_7_0 /* MAX VERSION is the maximum                            \

+    GD_OP_VERSION_7_1 /* MAX VERSION is the maximum                            \

                          count in VME table, should                            \

                          keep changing with                                    \

                          introduction of newer                                 \

@@ -138,6 +138,8 @@

 #define GD_OP_VERSION_7_0 70000 /* Op-version for GlusterFS 7.0 */

+#define GD_OP_VERSION_7_1 70100 /* Op-version for GlusterFS 7.1 */

+

 #include "glusterfs/xlator.h"

 #include "glusterfs/options.h"

diff --git a/libglusterfs/src/syncop-utils.c b/libglusterfs/src/syncop-utils.c

index be03527..2269c76 100644

--- a/libglusterfs/src/syncop-utils.c

+++ b/libglusterfs/src/syncop-utils.c

@@ -495,9 +495,7 @@ syncop_dir_scan(xlator_t *subvol, loc_t *loc, int pid, void *data,

             if (!strcmp(entry->d_name, ".") || !strcmp(entry->d_name, ".."))

                 continue;

-            ret = fn(subvol, entry, loc, data);

-            if (ret)

-                break;

+            ret |= fn(subvol, entry, loc, data);

         }

         gf_dirent_free(&entries);

         if (ret)

diff --git a/tests/basic/afr/add-brick-self-heal-non-granular.t b/tests/basic/afr/add-brick-self-heal-non-granular.t

new file mode 100644

index 0000000..19caf24

--- /dev/null

+++ b/tests/basic/afr/add-brick-self-heal-non-granular.t

@@ -0,0 +1,75 @@

+#!/bin/bash

+. $(dirname $0)/../../include.rc

+. $(dirname $0)/../../volume.rc

+cleanup;

+

+TEST glusterd

+TEST pidof glusterd

+TEST $CLI volume create $V0 replica 2 $H0:$B0/${V0}{0,1}

+EXPECT 'Created' volinfo_field $V0 'Status';

+TEST $CLI volume set $V0 cluster.granular-entry-heal off

+TEST $CLI volume start $V0

+EXPECT 'Started' volinfo_field $V0 'Status';

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" brick_up_status $V0 $H0 $B0/${V0}0

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" brick_up_status $V0 $H0 $B0/${V0}1

+

+TEST $CLI volume set $V0 cluster.data-self-heal off

+TEST $CLI volume set $V0 cluster.metadata-self-heal off

+TEST $CLI volume set $V0 cluster.entry-self-heal off

+TEST $CLI volume set $V0 cluster.heal-timeout 5

+

+TEST $CLI volume set $V0 self-heal-daemon off

+TEST $GFS --volfile-id=$V0 --volfile-server=$H0 $M0;

+

+# Create files

+for i in {1..5}

+do

+        echo $i > $M0/file$i.txt

+done

+

+# Metadata changes

+TEST setfattr -n user.test -v qwerty $M0/file5.txt

+

+# Add brick1

+TEST $CLI volume add-brick $V0 replica 3 $H0:$B0/${V0}2

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" brick_up_status $V0 $H0 $B0/${V0}2

+

+# New-brick should accuse the old-bricks (Simulating case for data-loss)

+TEST setfattr -n trusted.afr.$V0-client-0 -v 0x000000000000000000000001 $B0/${V0}2/

+TEST setfattr -n trusted.afr.$V0-client-1 -v 0x000000000000000000000001 $B0/${V0}2/

+

+# Check if pending xattr and dirty-xattr are set for newly-added-brick

+EXPECT "000000000000000100000001" get_hex_xattr trusted.afr.$V0-client-2 $B0/${V0}0

+EXPECT "000000000000000100000001" get_hex_xattr trusted.afr.$V0-client-2 $B0/${V0}1

+EXPECT "000000000000000000000001" get_hex_xattr trusted.afr.dirty $B0/${V0}2

+

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status $V0 0

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status $V0 1

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status $V0 2

+

+TEST $CLI volume set $V0 self-heal-daemon on

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "Y" glustershd_up_status

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 0

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 1

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 2

+TEST $CLI volume heal $V0

+

+# Wait for heal to complete

+EXPECT_WITHIN $HEAL_TIMEOUT "^0$" get_pending_heal_count $V0

+

+# Check if entry-heal has happened

+TEST diff <(ls $B0/${V0}0 | sort) <(ls $B0/${V0}2 | sort)

+TEST diff <(ls $B0/${V0}1 | sort) <(ls $B0/${V0}2 | sort)

+

+# Test if data was healed

+TEST diff $B0/${V0}0/file1.txt $B0/${V0}2/file1.txt

+

+# Test if metadata was healed and exists on both the bricks

+EXPECT "qwerty" get_text_xattr user.test $B0/${V0}2/file5.txt

+EXPECT "qwerty" get_text_xattr user.test $B0/${V0}0/file5.txt

+

+EXPECT "000000000000000000000000" get_hex_xattr trusted.afr.$V0-client-2 $B0/${V0}0

+EXPECT "000000000000000000000000" get_hex_xattr trusted.afr.$V0-client-2 $B0/${V0}1

+EXPECT "000000000000000000000000" get_hex_xattr trusted.afr.dirty $B0/${V0}2

+

+cleanup;

diff --git a/tests/basic/afr/add-brick-self-heal.t b/tests/basic/afr/add-brick-self-heal.t

index c847e22..7ebf4f6 100644

--- a/tests/basic/afr/add-brick-self-heal.t

+++ b/tests/basic/afr/add-brick-self-heal.t

@@ -38,8 +38,8 @@ TEST setfattr -n trusted.afr.$V0-client-0 -v 0x000000000000000000000001 $B0/${V0

 TEST setfattr -n trusted.afr.$V0-client-1 -v 0x000000000000000000000001 $B0/${V0}2/

 # Check if pending xattr and dirty-xattr are set for newly-added-brick

-EXPECT "000000000000000100000001" get_hex_xattr trusted.afr.$V0-client-2 $B0/${V0}0

-EXPECT "000000000000000100000001" get_hex_xattr trusted.afr.$V0-client-2 $B0/${V0}1

+EXPECT "000000010000000100000001" get_hex_xattr trusted.afr.$V0-client-2 $B0/${V0}0

+EXPECT "000000010000000100000001" get_hex_xattr trusted.afr.$V0-client-2 $B0/${V0}1

 EXPECT "000000000000000000000001" get_hex_xattr trusted.afr.dirty $B0/${V0}2

 EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status $V0 0

diff --git a/tests/basic/afr/bug-1130892-non-granular.t b/tests/basic/afr/bug-1130892-non-granular.t

new file mode 100644

index 0000000..3cdbc7d

--- /dev/null

+++ b/tests/basic/afr/bug-1130892-non-granular.t

@@ -0,0 +1,77 @@

+#!/bin/bash

+. $(dirname $0)/../../include.rc

+. $(dirname $0)/../../volume.rc

+. $(dirname $0)/../../afr.rc

+

+cleanup;

+

+TEST glusterd

+TEST pidof glusterd

+TEST $CLI volume info;

+

+# Create a 1X2 replica

+TEST $CLI volume create $V0 replica 2 $H0:$B0/${V0}-{0,1}

+EXPECT 'Created' volinfo_field $V0 'Status';

+TEST $CLI volume set $V0 cluster.granular-entry-heal off

+

+# Disable self-heal daemon

+TEST gluster volume set $V0 self-heal-daemon off

+

+# Enable Client side heal

+TEST $CLI volume set $V0 cluster.data-self-heal off

+TEST $CLI volume set $V0 cluster.metadata-self-heal off

+TEST $CLI volume set $V0 cluster.entry-self-heal off

+

+# Disable all perf-xlators

+TEST $CLI volume set $V0 performance.quick-read off

+TEST $CLI volume set $V0 performance.io-cache off

+TEST $CLI volume set $V0 performance.write-behind off

+TEST $CLI volume set $V0 performance.stat-prefetch off

+TEST $CLI volume set $V0 performance.read-ahead off

+

+# Volume start

+TEST $CLI volume start $V0;

+EXPECT 'Started' volinfo_field $V0 'Status';

+

+# FUSE Mount

+TEST ${GFS} -s $H0 --volfile-id $V0 $M0

+

+# Create files and dirs

+TEST mkdir -p $M0/one/two/

+TEST `echo "Carpe diem" > $M0/one/two/three`

+

+# Simulate disk-replacement

+TEST kill_brick $V0 $H0 $B0/${V0}-1

+EXPECT_WITHIN ${PROCESS_DOWN_TIMEOUT} "^0$" afr_child_up_status $V0 1

+TEST rm -rf $B0/${V0}-1/one

+TEST rm -rf $B0/${V0}-1/.glusterfs

+

+#Ideally, disk replacement is done using reset-brick or replace-brick gluster CLI

+#which will create .glusterfs folder.

+mkdir $B0/${V0}-1/.glusterfs && chmod 600 $B0/${V0}-1/.glusterfs

+

+# Start force

+TEST $CLI volume start $V0 force

+

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status $V0 1

+

+TEST stat $M0/one

+

+sleep 1

+

+# Check pending xattrs

+EXPECT "00000000" afr_get_specific_changelog_xattr $B0/${V0}-0/one trusted.afr.$V0-client-1 data

+EXPECT_NOT "00000000" afr_get_specific_changelog_xattr $B0/${V0}-0/one trusted.afr.$V0-client-1 entry

+EXPECT_NOT "00000000" afr_get_specific_changelog_xattr $B0/${V0}-0/one trusted.afr.$V0-client-1 metadata

+

+TEST gluster volume set $V0 self-heal-daemon on

+

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "Y" glustershd_up_status

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 0

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 1

+TEST $CLI volume heal $V0

+EXPECT_WITHIN $HEAL_TIMEOUT "Y" is_dir_heal_done $B0/${V0}-0 $B0/${V0}-1 one

+EXPECT_WITHIN $HEAL_TIMEOUT "Y" is_dir_heal_done $B0/${V0}-0 $B0/${V0}-1 one/two

+EXPECT_WITHIN $HEAL_TIMEOUT "Y" is_file_heal_done $B0/${V0}-0 $B0/${V0}-1 one/two/three

+

+cleanup;

diff --git a/tests/basic/afr/bug-1493415-gfid-heal-non-granular.t b/tests/basic/afr/bug-1493415-gfid-heal-non-granular.t

new file mode 100644

index 0000000..aff001c

--- /dev/null

+++ b/tests/basic/afr/bug-1493415-gfid-heal-non-granular.t

@@ -0,0 +1,79 @@

+#!/bin/bash

+. $(dirname $0)/../../include.rc

+. $(dirname $0)/../../volume.rc

+. $(dirname $0)/../../afr.rc

+cleanup;

+

+TEST glusterd

+TEST pidof glusterd

+TEST $CLI volume create $V0 replica 2 $H0:$B0/${V0}{0,1}

+TEST $CLI volume set $V0 cluster.granular-entry-heal off

+TEST $CLI volume start $V0

+

+TEST $GFS --volfile-id=$V0 --volfile-server=$H0 --attribute-timeout=0 --entry-timeout=0 $M0;

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status $V0 0

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status $V0 1

+TEST $CLI volume set $V0 self-heal-daemon off

+

+# Create base entry in indices/xattrop

+echo "Data" > $M0/FILE

+

+#------------------------------------------------------------------------------#

+TEST touch $M0/f1

+gfid_f1=$(gf_get_gfid_xattr $B0/${V0}0/f1)

+gfid_str_f1=$(gf_gfid_xattr_to_str $gfid_f1)

+

+# Remove gfid xattr and .glusterfs hard link from 2nd brick. This simulates a

+# brick crash at the point where file got created but no xattrs were set.

+TEST setfattr -x trusted.gfid $B0/${V0}1/f1

+TEST rm $B0/${V0}1/.glusterfs/${gfid_str_f1:0:2}/${gfid_str_f1:2:2}/$gfid_str_f1

+

+# storage/posix considers that a file without gfid changed less than a second

+# before doesn't exist, so we need to wait for a second to force posix to

+# consider that this is a valid file but without gfid.

+sleep 2

+

+# Assume there were no pending xattrs on parent dir due to 1st brick crashing

+# too. Then name heal from client must heal the gfid.

+EXPECT_WITHIN $UMOUNT_TIMEOUT "Y" force_umount $M0

+TEST $GFS --volfile-id=$V0 --volfile-server=$H0 --attribute-timeout=0 --entry-timeout=0 $M0;

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status $V0 0

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status $V0 1

+TEST stat $M0/f1

+EXPECT "$gfid_f1" gf_get_gfid_xattr $B0/${V0}1/f1

+TEST stat $B0/${V0}1/.glusterfs/${gfid_str_f1:0:2}/${gfid_str_f1:2:2}/$gfid_str_f1

+

+#------------------------------------------------------------------------------#

+TEST mkdir $M0/dir

+TEST touch $M0/dir/f2

+gfid_f2=$(gf_get_gfid_xattr $B0/${V0}0/dir/f2)

+gfid_str_f2=$(gf_gfid_xattr_to_str $gfid_f2)

+

+# Remove gfid xattr and .glusterfs hard link from 2nd brick. This simulates a

+# brick crash at the point where file got created but no xattrs were set.

+TEST setfattr -x trusted.gfid $B0/${V0}1/dir/f2

+TEST rm $B0/${V0}1/.glusterfs/${gfid_str_f2:0:2}/${gfid_str_f2:2:2}/$gfid_str_f2

+

+#Now simulate setting of pending entry xattr on parent dir of 1st brick.

+TEST setfattr -n trusted.afr.$V0-client-1 -v 0x000000000000000000000001 $B0/${V0}0/dir

+create_brick_xattrop_entry $B0/${V0}0 dir

+

+# storage/posix considers that a file without gfid changed less than a second

+# before doesn't exist, so we need to wait for a second to force posix to

+# consider that this is a valid file but without gfid.

+sleep 2

+

+#Trigger entry-heal via shd

+TEST $CLI volume set $V0 self-heal-daemon on

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "Y" glustershd_up_status

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 0

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 1

+

+TEST $CLI volume heal $V0

+EXPECT_WITHIN $HEAL_TIMEOUT "^0$" get_pending_heal_count $V0

+

+EXPECT "$gfid_f2" gf_get_gfid_xattr $B0/${V0}1/dir/f2

+TEST stat $B0/${V0}1/.glusterfs/${gfid_str_f2:0:2}/${gfid_str_f2:2:2}/$gfid_str_f2

+

+#------------------------------------------------------------------------------#

+cleanup;

diff --git a/tests/basic/afr/bug-1722507-type-mismatch-error-handling-non-granular.t b/tests/basic/afr/bug-1722507-type-mismatch-error-handling-non-granular.t

new file mode 100644

index 0000000..9079c93

--- /dev/null

+++ b/tests/basic/afr/bug-1722507-type-mismatch-error-handling-non-granular.t

@@ -0,0 +1,117 @@

+#!/bin/bash

+

+. $(dirname $0)/../../include.rc

+. $(dirname $0)/../../volume.rc

+. $(dirname $0)/../../afr.rc

+

+cleanup;

+

+## Start and create a volume

+TEST glusterd;

+TEST pidof glusterd;

+TEST $CLI volume info;

+

+TEST $CLI volume create $V0 replica 3 $H0:$B0/${V0}{0,1,2};

+TEST $CLI volume set $V0 cluster.granular-entry-heal off

+TEST $CLI volume start $V0;

+TEST $CLI volume set $V0 cluster.heal-timeout 5

+TEST $CLI volume heal $V0 disable

+EXPECT 'Started' volinfo_field $V0 'Status';

+TEST $GFS --volfile-id=/$V0 --volfile-server=$H0 $M0

+

+TEST mkdir $M0/dir

+

+##########################################################################################

+# GFID link file and the GFID is missing on one brick and all the bricks are being blamed.

+

+TEST touch $M0/dir/file

+TEST `echo append>> $M0/dir/file`

+

+#B0 and B2 must blame B1

+setfattr -n trusted.afr.$V0-client-0 -v 0x000000000000000000000001 $B0/$V0"2"/dir

+setfattr -n trusted.afr.$V0-client-1 -v 0x000000000000000000000001 $B0/$V0"0"/dir

+setfattr -n trusted.afr.$V0-client-2 -v 0x000000000000000000000001 $B0/$V0"0"/dir

+

+# Add entry to xattrop dir to trigger index heal.

+xattrop_dir0=$(afr_get_index_path $B0/$V0"0")

+base_entry_b0=`ls $xattrop_dir0`

+gfid_str=$(gf_gfid_xattr_to_str $(gf_get_gfid_xattr $B0/$V0"0"/dir/))

+ln $xattrop_dir0/$base_entry_b0 $xattrop_dir0/$gfid_str

+EXPECT "^1$" get_pending_heal_count $V0

+

+# Remove the gfid xattr and the link file on one brick.

+gfid_file=$(gf_get_gfid_xattr $B0/$V0"0"/dir/file)

+gfid_str_file=$(gf_gfid_xattr_to_str $gfid_file)

+TEST setfattr -x trusted.gfid $B0/${V0}0/dir/file

+TEST rm -f $B0/${V0}0/.glusterfs/${gfid_str_file:0:2}/${gfid_str_file:2:2}/$gfid_str_file

+

+# Launch heal

+TEST $CLI volume heal $V0 enable

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "^Y$" glustershd_up_status

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "^1$" afr_child_up_status_in_shd $V0 0

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "^1$" afr_child_up_status_in_shd $V0 1

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "^1$" afr_child_up_status_in_shd $V0 2

+

+# Wait for 2 second to force posix to consider that this is a valid file but

+# without gfid.

+sleep 2

+TEST $CLI volume heal $V0

+

+# Heal should not fail as the file is missing gfid xattr and the link file,

+# which is not actually the gfid or type mismatch.

+EXPECT_WITHIN $HEAL_TIMEOUT "^0$" get_pending_heal_count $V0

+

+EXPECT "$gfid_file" gf_get_gfid_xattr $B0/${V0}0/dir/file

+TEST stat $B0/${V0}0/.glusterfs/${gfid_str_file:0:2}/${gfid_str_file:2:2}/$gfid_str_file

+rm -f $M0/dir/file

+

+

+###########################################################################################

+# GFID link file and the GFID is missing on two bricks and all the bricks are being blamed.

+

+TEST $CLI volume heal $V0 disable

+TEST touch $M0/dir/file

+#TEST kill_brick $V0 $H0 $B0/$V0"1"

+

+#B0 and B2 must blame B1

+setfattr -n trusted.afr.$V0-client-0 -v 0x000000000000000000000001 $B0/$V0"2"/dir

+setfattr -n trusted.afr.$V0-client-1 -v 0x000000000000000000000001 $B0/$V0"0"/dir

+setfattr -n trusted.afr.$V0-client-2 -v 0x000000000000000000000001 $B0/$V0"0"/dir

+

+# Add entry to xattrop dir to trigger index heal.

+xattrop_dir0=$(afr_get_index_path $B0/$V0"0")

+base_entry_b0=`ls $xattrop_dir0`

+gfid_str=$(gf_gfid_xattr_to_str $(gf_get_gfid_xattr $B0/$V0"0"/dir/))

+ln $xattrop_dir0/$base_entry_b0 $xattrop_dir0/$gfid_str

+EXPECT "^1$" get_pending_heal_count $V0

+

+# Remove the gfid xattr and the link file on two bricks.

+gfid_file=$(gf_get_gfid_xattr $B0/$V0"0"/dir/file)

+gfid_str_file=$(gf_gfid_xattr_to_str $gfid_file)

+TEST setfattr -x trusted.gfid $B0/${V0}0/dir/file

+TEST rm -f $B0/${V0}0/.glusterfs/${gfid_str_file:0:2}/${gfid_str_file:2:2}/$gfid_str_file

+TEST setfattr -x trusted.gfid $B0/${V0}1/dir/file

+TEST rm -f $B0/${V0}1/.glusterfs/${gfid_str_file:0:2}/${gfid_str_file:2:2}/$gfid_str_file

+

+# Launch heal

+TEST $CLI volume heal $V0 enable

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "^Y$" glustershd_up_status

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "^1$" afr_child_up_status_in_shd $V0 0

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "^1$" afr_child_up_status_in_shd $V0 1

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "^1$" afr_child_up_status_in_shd $V0 2

+

+# Wait for 2 second to force posix to consider that this is a valid file but

+# without gfid.

+sleep 2

+TEST $CLI volume heal $V0

+

+# Heal should not fail as the file is missing gfid xattr and the link file,

+# which is not actually the gfid or type mismatch.

+EXPECT_WITHIN $HEAL_TIMEOUT "^0$" get_pending_heal_count $V0

+

+EXPECT "$gfid_file" gf_get_gfid_xattr $B0/${V0}0/dir/file

+TEST stat $B0/${V0}0/.glusterfs/${gfid_str_file:0:2}/${gfid_str_file:2:2}/$gfid_str_file

+EXPECT "$gfid_file" gf_get_gfid_xattr $B0/${V0}1/dir/file

+TEST stat $B0/${V0}1/.glusterfs/${gfid_str_file:0:2}/${gfid_str_file:2:2}/$gfid_str_file

+

+cleanup

diff --git a/tests/basic/afr/bug-1749322-entry-heal-not-happening-non-granular.t b/tests/basic/afr/bug-1749322-entry-heal-not-happening-non-granular.t

new file mode 100644

index 0000000..4f27da4

--- /dev/null

+++ b/tests/basic/afr/bug-1749322-entry-heal-not-happening-non-granular.t

@@ -0,0 +1,90 @@

+#!/bin/bash

+

+. $(dirname $0)/../../include.rc

+. $(dirname $0)/../../volume.rc

+. $(dirname $0)/../../afr.rc

+

+cleanup

+

+function check_gfid_and_link_count

+{

+        local file=$1

+

+        file_gfid_b0=$(gf_get_gfid_xattr $B0/${V0}0/$file)

+        TEST [ ! -z $file_gfid_b0 ]

+        file_gfid_b1=$(gf_get_gfid_xattr $B0/${V0}1/$file)

+        file_gfid_b2=$(gf_get_gfid_xattr $B0/${V0}2/$file)

+        EXPECT $file_gfid_b0 echo $file_gfid_b1

+        EXPECT $file_gfid_b0 echo $file_gfid_b2

+

+        EXPECT "2" stat -c %h $B0/${V0}0/$file

+        EXPECT "2" stat -c %h $B0/${V0}1/$file

+        EXPECT "2" stat -c %h $B0/${V0}2/$file

+}

+TESTS_EXPECTED_IN_LOOP=18

+

+################################################################################

+## Start and create a volume

+TEST glusterd;

+TEST pidof glusterd;

+TEST $CLI volume info;

+

+TEST $CLI volume create $V0 replica 3 $H0:$B0/${V0}{0,1,2};

+TEST $CLI volume set $V0 cluster.granular-entry-heal off

+TEST $CLI volume start $V0;

+TEST $CLI volume set $V0 cluster.heal-timeout 5

+TEST $CLI volume heal $V0 disable

+EXPECT 'Started' volinfo_field $V0 'Status';

+TEST $GFS --volfile-id=/$V0 --volfile-server=$H0 $M0

+

+TEST mkdir $M0/dir

+TEST `echo "File 1 " > $M0/dir/file1`

+TEST touch $M0/dir/file{2..4}

+

+# Remove file2 from 1st & 3rd bricks

+TEST rm -f $B0/$V0"0"/dir/file2

+TEST rm -f $B0/$V0"2"/dir/file2

+

+# Remove file3 and the .glusterfs hardlink from 1st & 2nd bricks

+gfid_file3=$(gf_get_gfid_xattr $B0/$V0"0"/dir/file3)

+gfid_str_file3=$(gf_gfid_xattr_to_str $gfid_file3)

+TEST rm $B0/$V0"0"/.glusterfs/${gfid_str_file3:0:2}/${gfid_str_file3:2:2}/$gfid_str_file3

+TEST rm $B0/$V0"1"/.glusterfs/${gfid_str_file3:0:2}/${gfid_str_file3:2:2}/$gfid_str_file3

+TEST rm -f $B0/$V0"0"/dir/file3

+TEST rm -f $B0/$V0"1"/dir/file3

+

+# Remove the .glusterfs hardlink and the gfid xattr of file4 on 3rd brick

+gfid_file4=$(gf_get_gfid_xattr $B0/$V0"0"/dir/file4)

+gfid_str_file4=$(gf_gfid_xattr_to_str $gfid_file4)

+TEST rm $B0/$V0"2"/.glusterfs/${gfid_str_file4:0:2}/${gfid_str_file4:2:2}/$gfid_str_file4

+TEST setfattr -x trusted.gfid $B0/$V0"2"/dir/file4

+

+# B0 and B2 blame each other

+setfattr -n trusted.afr.$V0-client-0 -v 0x000000000000000000000001 $B0/$V0"2"/dir

+setfattr -n trusted.afr.$V0-client-2 -v 0x000000000000000000000001 $B0/$V0"0"/dir

+

+# Add entry to xattrop dir on first brick.

+xattrop_dir0=$(afr_get_index_path $B0/$V0"0")

+base_entry_b0=`ls $xattrop_dir0`

+gfid_str=$(gf_gfid_xattr_to_str $(gf_get_gfid_xattr $B0/$V0"0"/dir/))

+TEST ln $xattrop_dir0/$base_entry_b0 $xattrop_dir0/$gfid_str

+

+EXPECT "^1$" get_pending_heal_count $V0

+

+# Launch heal

+TEST $CLI volume heal $V0 enable

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "^Y$" glustershd_up_status

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "^1$" afr_child_up_status_in_shd $V0 0

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "^1$" afr_child_up_status_in_shd $V0 1

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "^1$" afr_child_up_status_in_shd $V0 2

+TEST $CLI volume heal $V0

+EXPECT_WITHIN $HEAL_TIMEOUT "^0$" get_pending_heal_count $V0

+

+# All the files must be present on all the bricks after conservative merge and

+# should have the gfid xattr and the .glusterfs hardlink.

+check_gfid_and_link_count dir/file1

+check_gfid_and_link_count dir/file2

+check_gfid_and_link_count dir/file3

+check_gfid_and_link_count dir/file4

+

+cleanup

diff --git a/tests/basic/afr/replace-brick-self-heal-non-granular.t b/tests/basic/afr/replace-brick-self-heal-non-granular.t

new file mode 100644

index 0000000..c86bff1

--- /dev/null

+++ b/tests/basic/afr/replace-brick-self-heal-non-granular.t

@@ -0,0 +1,65 @@

+#!/bin/bash

+. $(dirname $0)/../../include.rc

+. $(dirname $0)/../../volume.rc

+cleanup;

+

+TEST glusterd

+TEST pidof glusterd

+TEST $CLI volume create $V0 replica 2 $H0:$B0/${V0}{0,1}

+TEST $CLI volume set $V0 cluster.granular-entry-heal off

+TEST $CLI volume start $V0

+TEST $CLI volume set $V0 cluster.data-self-heal off

+TEST $CLI volume set $V0 cluster.metadata-self-heal off

+TEST $CLI volume set $V0 cluster.entry-self-heal off

+TEST $CLI volume set $V0 cluster.heal-timeout 5

+TEST $CLI volume set $V0 self-heal-daemon off

+TEST $GFS --volfile-id=$V0 --volfile-server=$H0 $M0;

+

+# Create files

+for i in {1..5}

+do

+        echo $i > $M0/file$i.txt

+done

+

+# Metadata changes

+TEST setfattr -n user.test -v qwerty $M0/file5.txt

+

+# Replace brick1

+TEST $CLI volume replace-brick $V0 $H0:$B0/${V0}1 $H0:$B0/${V0}1_new commit force

+

+# Replaced-brick should accuse the non-replaced-brick (Simulating case for data-loss)

+TEST setfattr -n trusted.afr.$V0-client-0 -v 0x000000000000000000000001 $B0/${V0}1_new/

+

+# Check if pending xattr and dirty-xattr are set for replaced-brick

+EXPECT "000000000000000100000001" get_hex_xattr trusted.afr.$V0-client-1 $B0/${V0}0

+EXPECT "000000000000000000000001" get_hex_xattr trusted.afr.dirty $B0/${V0}1_new

+

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status $V0 0

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status $V0 1

+

+TEST $CLI volume set $V0 self-heal-daemon on

+EXPECT_WITHIN $PROCESS_UP_TIMEOUT "Y" glustershd_up_status

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 0

+EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status_in_shd $V0 1

+TEST $CLI volume heal $V0

+

+# Wait for heal to complete

+EXPECT_WITHIN $HEAL_TIMEOUT "0" get_pending_heal_count $V0

+

+# Check if entry-heal has happened

+TEST diff <(ls $B0/${V0}0 | sort) <(ls $B0/${V0}1_new | sort)

+

+# To make sure that files were not lost from brick0

+TEST diff <(ls $B0/${V0}0 | sort) <(ls $B0/${V0}1 | sort)

+EXPECT "000000000000000000000000" get_hex_xattr trusted.afr.$V0-client-1 $B0/${V0}0

+

+# Test if data was healed

+TEST diff $B0/${V0}0/file1.txt $B0/${V0}1_new/file1.txt

+# To make sure that data was not lost from brick0

+TEST diff $B0/${V0}0/file1.txt $B0/${V0}1/file1.txt

+

+# Test if metadata was healed and exists on both the bricks

+EXPECT "qwerty" get_text_xattr user.test $B0/${V0}1_new/file5.txt

+EXPECT "qwerty" get_text_xattr user.test $B0/${V0}0/file5.txt

+

+cleanup;

diff --git a/tests/basic/afr/replace-brick-self-heal.t b/tests/basic/afr/replace-brick-self-heal.t

index 0360db7..da31c87 100644

--- a/tests/basic/afr/replace-brick-self-heal.t

+++ b/tests/basic/afr/replace-brick-self-heal.t

@@ -30,7 +30,7 @@ TEST $CLI volume replace-brick $V0 $H0:$B0/${V0}1 $H0:$B0/${V0}1_new commit forc

 TEST setfattr -n trusted.afr.$V0-client-0 -v 0x000000000000000000000001 $B0/${V0}1_new/

 # Check if pending xattr and dirty-xattr are set for replaced-brick

-EXPECT "000000000000000100000001" get_hex_xattr trusted.afr.$V0-client-1 $B0/${V0}0

+EXPECT "000000010000000100000001" get_hex_xattr trusted.afr.$V0-client-1 $B0/${V0}0

 EXPECT "000000000000000000000001" get_hex_xattr trusted.afr.dirty $B0/${V0}1_new

 EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" afr_child_up_status $V0 0

diff --git a/tests/bugs/replicate/bug-1130892.t b/tests/bugs/replicate/bug-1130892.t

index 0f57d66..e23eb26 100644

--- a/tests/bugs/replicate/bug-1130892.t

+++ b/tests/bugs/replicate/bug-1130892.t

@@ -56,7 +56,7 @@ EXPECT_WITHIN $CHILD_UP_TIMEOUT "1" afr_child_up_status $V0 1

 TEST stat $M0/one

 # Check pending xattrs

-EXPECT "00000000" afr_get_specific_changelog_xattr $B0/${V0}-0/one trusted.afr.$V0-client-1 data

+EXPECT "00000001" afr_get_specific_changelog_xattr $B0/${V0}-0/one trusted.afr.$V0-client-1 data

 EXPECT_NOT "00000000" afr_get_specific_changelog_xattr $B0/${V0}-0/one trusted.afr.$V0-client-1 entry

 EXPECT "00000000" afr_get_specific_changelog_xattr $B0/${V0}-0/one trusted.afr.$V0-client-1 metadata

diff --git a/tests/bugs/replicate/bug-1493415-gfid-heal.t b/tests/bugs/replicate/bug-1493415-gfid-heal.t

index 125c35a..9714d5e 100644

--- a/tests/bugs/replicate/bug-1493415-gfid-heal.t

+++ b/tests/bugs/replicate/bug-1493415-gfid-heal.t

@@ -49,7 +49,7 @@ TEST setfattr -x trusted.gfid $B0/${V0}1/dir/f2

 TEST rm $B0/${V0}1/.glusterfs/${gfid_str_f2:0:2}/${gfid_str_f2:2:2}/$gfid_str_f2

 #Now simulate setting of pending entry xattr on parent dir of 1st brick.

-TEST setfattr -n trusted.afr.$V0-client-1 -v 0x000000000000000000000001 $B0/${V0}0/dir

+TEST setfattr -n trusted.afr.$V0-client-1 -v 0x000000010000000000000001 $B0/${V0}0/dir

 create_brick_xattrop_entry $B0/${V0}0 dir

 #Trigger entry-heal via shd

diff --git a/tests/bugs/replicate/bug-1722507-type-mismatch-error-handling.t b/tests/bugs/replicate/bug-1722507-type-mismatch-error-handling.t

index 0aeaaaf..1fdf7ea 100644

--- a/tests/bugs/replicate/bug-1722507-type-mismatch-error-handling.t

+++ b/tests/bugs/replicate/bug-1722507-type-mismatch-error-handling.t

@@ -23,19 +23,21 @@ TEST mkdir $M0/dir

 ##########################################################################################

 # GFID link file and the GFID is missing on one brick and all the bricks are being blamed.

-TEST touch $M0/dir/file

-#TEST kill_brick $V0 $H0 $B0/$V0"1"

+TEST `echo append>> $M0/dir/file`