An IRC user was complaining of "OSError: [Errno 12] Cannot allocate memory" error messages. I tracked them down to a Server::xtime buffer overflow. In resource.py:144 return struct.unpack('!II', Xattr.lgetxattr(path, '.'.join([cls.GX_NSPACE, uuid, 'xtime']), 8)) Apparently getattr is overflowing the fixed "8". The user stopped responding after I found that problem so I have no further diagnostic details.
I could not reproduce this in my setup. Also the bug does not have enough information (geo-rep [mount] log files, backtraces) to make some assumptions/about the error. I'm closing this with "INSUFFICIENT DATA" for now. Please reopen if someone hits this again.