Paul Archer schrieb: > 11:17am, Tomasz Chmielewski wrote: > >> Why should any filesystem perform seeks better (when writing) than any >> other filesystem? >> >> I imagine it could be true only if: >> >> - kernel would cache a large amount of writes >> - kernel would commit these writes not in a FIFO manner, but whenever it >> sees that the blocks on the underlying device are close to each other >> >> Can ZFS do it? >> > zfs should be faster on writes because it never overwrites any data in > place; instead, it finds an available bit of disk, writes its data, then > updates the metablocks. Still unless a fs is very poorly written, you > shouldn't see more than a few percent different between filesystems. The > physical disk will almost always be your bottleneck.
For me, it seems, one of the problems is metadata updating. When BackupPC_trashClean removes directories, it doesn't write any real data, but rather only updates metablocks. As I see, 10 BackupPC_dump processes add new directories faster than 1 BackupPC_trashClean process can remove. Right now, I have over 100 directories in "trash", a few hours ago it was ~90. And I was running additional "rm -rf" processes just a couple of days ago. -- Tomasz Chmielewski http://wpkg.org ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ BackupPC-users mailing list [email protected] List: https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki: http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
