Hello All. Since I'm using the SerNet packages from Enterprise Samba dot org, please feel free to say "you should be asking the folks from Enterprise Samba about that".. Now with that said, on to my issue - for which i'd be very grateful if someene can help me understand if there is an issue with my configuration (either OS or samba) or perhaps if I've stumbled across a bug. I'm running the SerNet Samba Packages (3.6.6-44) on two-node CTDB cluster. My OS is CentOS 6.2. My cluster stack is Pacemaker and my filesystem is OCFS2. And just to make sure I've got everything covered, I'm using kernel 2.6.39-200.24.1 from the Oracle public yum servers so I can use some of the latest OCFS2 features on my filesystem and all of this is running on XCP (open source XenServer) As far as hardware details go, this is running on a PowerEdge R515 with 128GB of memory and 12 600GB 15K SAS drives with a H700 controller. I've been doing some basic performance testing and so far, I'd say my performance is exceptional. From two Windows VMs on a completely different host, I'm able to saturate two gibabit links when copying large files to and from the share, but there is a problem when I attempt to make copies of the files on the same share (e.g. copy \\server\share\file1 \\server\share\file1.backup) - and the behavior appears to be filesystem agnostic. I've tried with ext3 also and the behavior is the same. The Windows copy progress indicator will go back and forth and then finally it tells me I'm getting about several hundred KB p/s. And I know it's not the filesystem because on my OCFS2 volume, it takes me less than 5 seconds to copy and 480MB ISO back to the same filesystem. So that is my issue. Here are some files in case you have any thoughts. Please let me know if you'd like more information. Thanks in advance.. [smb.conf] http://pastebin.com/raw.php?i=4M10bbmF [sysctl.conf] http://pastebin.com/raw.php?i=gn84vYYb