Hi,
While testing nilfs2 for crash consistency, we stumbled upon the
following workload, where a file is not persisted even after a fsync
on the file.
Workload:
creat foo
fsync(foo)
--- crash----
On recovery : File foo is missing
While we agree that fsync on the parent directory is required to
persist the entry, most filesystems (ext4, xfs, btrfs, and f2fs),
don't require an explicit fsync on the directory to do so. On all the
above mentioned filesystems, the file foo is on disk if we recover
after a crash. Is this an expected behavior with nilfs2? Does nilfs2
always require a fsync on the parent directory to persist the newly
created file?
Filesystems like xfs that have strictly ordered metadata behavior,
persist all the transactions prior to the fsync call and thereby
ensure that the directory entry is also persisted in this case. Does
nilfs2 classify itself as a filesystem that provides strict metadata
ordering?
Thanks,
Jayashree Mohan
--
To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
