Description of problem: Distribute fails to heal quota xattrs after add-brick and rebalance. Following change reveals the present issue diff --git a/tests/basic/quota.t b/tests/basic/quota.t index d5a95e2..7e01d8d 100755 --- a/tests/basic/quota.t +++ b/tests/basic/quota.t @@ -168,17 +168,17 @@ done # 65 ## <Test whether quota limit crossed more than 10% of limit> ## --------------------------------------------------------- -USED_KB=`du -s $M0/$TESTDIR/dir1 | cut -f1`; +USED_KB=$(du -s $M0/$TESTDIR/dir1 | cut -f1); USED_MB=$(($USED_KB/1024)); TEST [ $USED_MB -le $((($QUOTALIMIT * 110) / 100)) ] # 66-67 ## <Test the xattrs healed to new brick> ## ------------------------------------- -TEST getfattr -d -m "trusted.glusterfs.quota.limit-set" -e hex \ +TEST getfattr -n trusted.glusterfs.quota.limit-set -e hex \ --absolute-names $B0/brick{3,4}/$TESTDIR/dir{1..10}; # Test on root. -TEST getfattr -d -m "trusted.glusterfs.quota.limit-set" -e hex \ +TEST getfattr -n trusted.glusterfs.quota.limit-set -e hex \ --absolute-names $B0/brick{3,4}; ## ------------------------------------------------- The real issue in the test case was it wasn't checking the attributes properly before so this issue seemed non-existent, when we add the proper checks - we do see that the "root" or the top-level brick doesn't properly carry the quota attributes from its neighboring bricks. Version-Release number of selected component (if applicable): mainline How reproducible: Modify the quota.t test case shown above, it is reproducible consistently.
This issue is same as bug# 1306220 and the patch has been submitted upstream: http://review.gluster.org/#/c/13422/ *** This bug has been marked as a duplicate of bug 1306220 ***