On Mon, Oct 10, 2011 at 10:18:25AM -0400, Christoph Hellwig wrote: > Any chance we can get some coverage for the losts of extents cases > in xfstests? It will probably require pre-created images, but I'd > really hate not beeing able to regression test this. Test 042 creates files with a lot of extents, but we don't have coverage of extent countsin the hundreds of thousands or greater. I can put together a simple test to do it using preallocation and/or hole punching, though I'm not sure what the runtime will be. Given we assume we have at least 10GB on the scratch device, I can easily create a couple of million extent files on a 4k block size filesystem for repair/db to walk.... Cheers, Dave. -- Dave Chinner david@xxxxxxxxxxxxx _______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
