diff options
| author | Ashish Pandey <aspandey@redhat.com> | 2016-09-20 12:32:28 +0530 |
|---|---|---|
| committer | Xavier Hernandez <xhernandez@datalab.es> | 2016-10-11 02:29:27 -0700 |
| commit | 0fed7e7f0aad9973900c89434f736797d9ace2bd (patch) | |
| tree | b7e02dfd5c6e21a29b5fe122bf911b10c42a611c /tests/basic/ec | |
| parent | 85e959052148ec481823d55c8b91cdee36da2b43 (diff) | |
cluster/ec: Implement heal info with lock
Problem: Currently heal info command prints all
the files/directories if the index for the
file/directory is present in .glusterfs/indices folder.
After implementing patch http://review.gluster.org/#/c/13733/
indices of the file which is going through update fop
will also be present in .glusterfs/indices even
if the fop is successful on all the brick. At this time
if heal info command is being used, it will also display this
file which is actually healthy and does not require any heal.
Solution: Take lock on a file corresponding to the indices
and inspect xattrs to decide if the file needs heal or not.
Change-Id: I6361e2813ece369be12d02e74816df4eddb81cfa
BUG: 1366815
Signed-off-by: Ashish Pandey <aspandey@redhat.com>
Reviewed-on: http://review.gluster.org/15543
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
Reviewed-by: Xavier Hernandez <xhernandez@datalab.es>
CentOS-regression: Gluster Build System <jenkins@build.gluster.org>
Smoke: Gluster Build System <jenkins@build.gluster.org>
Diffstat (limited to 'tests/basic/ec')
| -rw-r--r-- | tests/basic/ec/ec-background-heals.t | 6 |
1 files changed, 4 insertions, 2 deletions
diff --git a/tests/basic/ec/ec-background-heals.t b/tests/basic/ec/ec-background-heals.t index 7ac6c0efc12..eb434908bad 100644 --- a/tests/basic/ec/ec-background-heals.t +++ b/tests/basic/ec/ec-background-heals.t @@ -23,7 +23,10 @@ EXPECT_WITHIN $CHILD_UP_TIMEOUT "3" ec_child_up_count $V0 0 EXPECT_WITHIN $CONFIG_UPDATE_TIMEOUT "0" mount_get_option_value $M0 $V0-disperse-0 background-heals EXPECT_WITHIN $CONFIG_UPDATE_TIMEOUT "0" mount_get_option_value $M0 $V0-disperse-0 heal-wait-qlength TEST touch $M0/a -EXPECT_WITHIN $CHILD_UP_TIMEOUT "0" get_pending_heal_count $V0 #One for each active brick +EXPECT_WITHIN $HEAL_TIMEOUT "^0$" count_sh_entries $B0/${V0}0 +EXPECT_WITHIN $HEAL_TIMEOUT "^0$" count_sh_entries $B0/${V0}1 +EXPECT_WITHIN $HEAL_TIMEOUT "^0$" count_sh_entries $B0/${V0}2 + TEST kill_brick $V0 $H0 $B0/${V0}2 echo abc > $M0/a EXPECT 2 get_pending_heal_count $V0 #One for each active brick @@ -31,7 +34,6 @@ $CLI volume start $V0 force EXPECT_WITHIN $CHILD_UP_TIMEOUT "3" ec_child_up_count $V0 0 #Accessing file shouldn't heal the file EXPECT "abc" cat $M0/a -sleep 3 EXPECT 2 get_pending_heal_count $V0 #One for each active brick TEST $CLI volume set $V0 disperse.background-heals 1 EXPECT_WITHIN $CONFIG_UPDATE_TIMEOUT "1" mount_get_option_value $M0 $V0-disperse-0 background-heals |