> >> I think we need to reconsider the above change. The bug is real and > >> needs a fix, but maybe we append the xlator name to the end of the > >> provided filename and dump the stats into that, than unwind from the > >> first instance of io-stats. > > > > I assume you mean the first instance of io-stats that recognizes its > > own name in the argument, or else the setattr might still get turned > > around before it gets to where it's intended to take effect. Another > > option would be to put the desired io-stats level in the xattr name > > instead of its value. I'd use levels instead of names because names > > are made up internally and not readily/safely guessable by the user. > > This would require that we pattern-match on the xattr name instead > > of doing a direct string comparison, but that seems easy enough. > > > > No Jeff, I want it passed to all instances of io-stats in the graph ...which "unwind from the first instance" won't do. I thought "than" was a typo, but it looks like you were using it to mean "instead of". It makes more sense that way. > the filename argument can be specialized by adding *something* to it Seems easy enough. That would require only reverting the current patch and making a slight modification to how io-stats formats the file name for sys_open. We could also process a %s in the xattr value, but that seems more error-prone. _______________________________________________ Gluster-devel mailing list Gluster-devel@xxxxxxxxxxx http://www.gluster.org/mailman/listinfo/gluster-devel
