geo-replication is underperforming when the brick logs contain errors about .xtime xattrs

Issue

• The following errors gets logged on the geo-replication slave:

[2012-11-27 11:48:44.302652] E [posix.c:2582:posix_getxattr] 0-vol1-posix: getxattr failed on /bricks/vg_rhesb1_01/lv_vol1/FileStore/tenant_hua/39: trusted.glusterfs.ff100dd8-7241-4b80-a98a-08e03a9edacf.xtime (No data available)
[2012-11-27 11:48:44.434753] E [posix.c:993:posix_mkdir] 0-vol1-posix: setting xattrs on /bricks/vg_rhesb1_01/lv_vol1/FileStore/tenant_hua/39/85/2b failed (Operation not supported)

• The following errors are being logged often:

$ grep "\] E \[" *.log|grep "setting xattr" |wc -l
39778
$ grep "\] E \[" *.log|grep "getxattr" |wc -l
402108