From owner-freebsd-hackers Thu May 6 12:20:16 1999 Delivered-To: freebsd-hackers@freebsd.org Received: from sol (cs1-gw.cs.binghamton.edu [128.226.171.72]) by hub.freebsd.org (Postfix) with SMTP id 5F9B114DE3 for ; Thu, 6 May 1999 12:20:08 -0700 (PDT) (envelope-from zzhang@cs.binghamton.edu) Received: from localhost (zzhang@localhost) by sol (SMI-8.6/8.6.9) with SMTP id PAA15291 for ; Thu, 6 May 1999 15:09:30 -0400 Date: Thu, 6 May 1999 15:09:30 -0400 (EDT) From: Zhihui Zhang To: freebsd-hackers@freebsd.org Subject: Copies of superblocks in FFS Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG After reading the FFS source code, I can not help wondering why we keep N+1 superblocks in a file system, where N is the number of cylinder groups in the file system. It seems to me that only the second superblock in cylinder group 0 is actually used (cylinder group 0 contains two copies of superblocks). All other superblocks are never touched. If the primary superblock (the second copy in cylinder group 0, at offset 8192+8192) is updated and other superblocks are not updated at the same time, how can any of other copies be used to restore file system in case that the primary copy is corrupted somehow? If so, the performance will be degraded. Also, except for the root filesystem (/), all other filesystems (/var, /usr, etc.) do not have the (boot code + disklabel) installed, these space are also wasted (8192 bytes for each non-root filesystem). BTW, the hard disks are more stable nowadays and any bad sectors may have been hidden by the disk controllers (the filesystem does not have to deal with them). Any enlightment or correction is appreciated. -------------------------------------------------- Zhihui Zhang. Please visit http://www.freebsd.org -------------------------------------------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message