Fwd: panic(9) vs. RB_NOSYNC

Matthew Dillon dillon at apollo.backplane.com
Thu Aug 8 23:14:01 PDT 2013


:Thanks.  At least I read code's intention right. :)
:
:>     Most modern filesystems don't corrupt any more if left unsynchronized,
:>     but the real problem here is that filesystem do so much caching in
:>     memory, particularly when files are mmap()'d RW, that not at least
:>     trying to sync can result in disarray (rather than corruption) on boot
:
:Do you mean that "corruption" == "file metadata corruption" and "disarray" ==
:"file data-only corruption"?

    No filesystem that I know of guarantees operational order inbetween
    fsyncs, particularly if files are mmap()'d shared/RW.

					-Matt
					Matthew Dillon 
					<dillon at backplane.com>



More information about the Kernel mailing list