On Tue, Mar 02, 2010 at 04:26:59PM +0100, Jiri Denemark wrote:
Instead of opening storage file with O_DSYNC, make sure data are written to a disk only before we claim allocation has finished.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com> --- src/storage/storage_backend.c | 9 ++++++++- 1 files changed, 8 insertions(+), 1 deletions(-)
diff --git a/src/storage/storage_backend.c b/src/storage/storage_backend.c index 849f01b..4cc2f93 100644 --- a/src/storage/storage_backend.c +++ b/src/storage/storage_backend.c @@ -332,6 +332,13 @@ static int createRawFileOpHook(int fd, void *data) { goto cleanup; } } + + if (fsync(fd) < 0) { + ret = errno; + virReportSystemError(errno, _("cannot sync data to file '%s'"), + hdata->vol->target.path); + goto cleanup; + } }
cleanup: @@ -357,7 +364,7 @@ virStorageBackendCreateRaw(virConnectPtr conn ATTRIBUTE_UNUSED, }
if ((createstat = virFileOperation(vol->target.path, - O_RDWR | O_CREAT | O_EXCL | O_DSYNC, + O_RDWR | O_CREAT | O_EXCL, vol->target.perms.mode, vol->target.perms.uid, vol->target.perms.gid, createRawFileOpHook, &hdata,
I would tend to prefer this patch on the basis that I would prefer a smaller change to the code at this point before the release. But the O_DSYNC was added after finding that cache effects could just break the qemu startup timeout in the test environment. I'm not 100% sure fsync() will have the same behaviour overall, it should but theory and practice may diverge :-) So I'm tempted to go with this patch but we should validate it on testing before the end of the week. Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel@veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/