On Fri, Dec 06, 2024 at 12:43:03PM +0000, Mitta Sai Chaithanya wrote: > Thanks Dave Chinner for taking time and explaining in detail, we > are exposing XFS files through SPDK mechanism and as you pointed I have no idea what a "SPDK mechanism" is - there isn't a single hit in the kernel tree on "SPDK".... > for having low latencies we are writing to the file asynchronously > (using uring as default configuration). I have one follow up > question "Will be there any journal updates for future IOs when > entire file is explicitly zeroe'd and synced for future IOs"? Yes: mtime updates. These are asynchronous transactions, though, and if you are using O_DSYNC/RWF_DSYNC will not directly trigger journal flushes. If you use fsync/O_SYNC/RWF_SYNC, then mtime update will trigger journal flushes. You can use the lazytime mount option to avoid transactional mtime updates if necessary (only syncs mtime updates from write() calls when the file is otherwise modified or evicted from cache). -Dave. -- Dave Chinner david@xxxxxxxxxxxxx
