Hello Sergey, On Wed, Jun 04, 2008, Sergey POZNYAKOFF wrote: >> It seems that an archive written in copy out mode with cpio (GNU >> cpio) 2.6 is corrupt, The command used was simply: >> >> $ find . -depth -print | cpio -ov >backup.cpio >> [...] >> home/sharedsys/newfiles/test.ps >> home/sharedsys/newfiles >> home/sharedsys >> home/ >> 824190054 blocks >> >Did cpio -t (on the same machine) show the contents of this archive >without errors? > Both '-t' and '-i' options have the same text output, so the answer is no.
>> ...just as suggested by O'Reilly: >> >> http://www.onlamp.com/pub/a/bsd/2002/07/11/FreeBSD_Basics.html >> >> When restoring in copy out mode with the identical cpio(1) binary >> as used for the backup: > >That is to say, you were using another cpio binary for restoring, right? >If so, what version of cpio was it? > The system cpio(1) was used to back up the files, and then the operating system was updated. The backup was needed because during the OS update a new RAID type was used, causing the hard drives to be wiped clean. When the new OS was installed, a newer version of cpio(1) called (GNU cpio) 2.9 came with it. This is the version used at first. When problems arose, the exact original binary from the previous OS was used. It is called cpio (GNU cpio) 2.6. By the way, after using pax(1) it seems that most of the data has been restored. At least a few important directories were lost however, which is not a very good track record for cpio(1). I'm still wondering where it all went wrong. Regards, Christopher _______________________________________________ Bug-cpio mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-cpio
