andy/glusterfs
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

afr: prevent winding inodelks twice for arbiter volumes

Browse Source 
Problem:
In an arbiter volume, if there is a pending data heal of a file only on
arbiter brick, self-heal takes inodelks twice due to a code-bug but unlocks
it only once, leaving behind a stale lock on the brick. This causes
the next write to the file to hang.

Fix:
Fix the code-bug to take lock only once. This bug was introduced master
with commit eb472d82a083883335bc494b87ea175ac43471ff

Thanks to  Pranith Kumar K <pkarampu@redhat.com> for finding the RCA.

fixes: bz#1637802
Change-Id: I15ad969e10a6a3c4bd255e2948b6be6dcddc61e1
Signed-off-by: Ravishankar N <ravishankar@redhat.com>
This commit is contained in:
Ravishankar N 2018-10-10 12:18:55 +05:30 committed by Amar Tumballi
parent 4da0d93f87
commit 196b32423a
2 changed files with 45 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

44
tests/bugs/replicate/bug-1637802-arbiter-stale-data-heal-lock.t Normal file
View File

@ -0,0 +1,44 @@
#!/bin/bash
. $(dirname $0)/../../include.rc
. $(dirname $0)/../../volume.rc
. $(dirname $0)/../../afr.rc
cleanup;
# Test to check that data self-heal does not leave any stale lock.
TEST glusterd;
TEST pidof glusterd;
TEST $CLI volume create $V0 replica 3 arbiter 1 $H0:$B0/${V0}{0,1,2};
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
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" brick_up_status $V0 $H0 $B0/${V0}2
TEST $GFS --volfile-id=/$V0 --volfile-server=$H0 $M0
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
# Create base entry in indices/xattrop
echo "Data" > $M0/FILE
# Kill arbiter brick and write to FILE.
TEST kill_brick $V0 $H0 $B0/${V0}2
echo "arbiter down" >> $M0/FILE
EXPECT 2 get_pending_heal_count $V0
# Bring it back up and let heal complete.
TEST $CLI volume start $V0 force
EXPECT_WITHIN $PROCESS_UP_TIMEOUT "1" brick_up_status $V0 $H0 $B0/${V0}2
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
# write to the FILE must succeed.
echo "this must succeed" >> $M0/FILE
TEST [ $? -eq 0 ]
cleanup;

2
xlators/cluster/afr/src/afr-self-heal-data.c
View File

@ -731,7 +731,7 @@ restore_time:
afr_selfheal_restore_time(frame, this, fd->inode, source, healed_sinks,
locked_replies);
if (!is_arbiter_the_only_sink || !empty_file) {
if (!is_arbiter_the_only_sink && !empty_file) {
ret = afr_selfheal_inodelk(frame, this, fd->inode, this->name, 0, 0,
data_lock);
if (ret < priv->child_count) {