This is good enough to allow me to do a test backup and then verify it without having to dissassemble and reassemble raids.
During a real restore, I would do the disk formatting and RAID configuration with my fingers anyway because I can't imagine testing mondoarchive well enough to trust it in that scenario. Therefore, if you take this patch it closes the issue as far as I am concerned. I'll send email to control in a moment saying that it's a mindi bug and that there's a patch. -- Tim Freeman http://www.fungible.com [EMAIL PROTECTED] diff -ru orig/mindi-2.20/rootfs/sbin/init mindi-2.20/rootfs/sbin/init --- orig/mindi-2.20/rootfs/sbin/init 2006-10-03 07:12:38.000000000 -0700 +++ mindi-2.20/rootfs/sbin/init 2007-08-21 08:03:05.000000000 -0700 @@ -320,11 +320,19 @@ raidstart $i fi done - elif which mdrun > /dev/null 2> /dev/null ; then - LogIt "Running 'mdrun'" 1 - mdrun + # /usr/share/doc/mdadm/README.mdrun on a Debian Etch system, at least, + # says that mdrun is obsolete and one should do + # mdadm --assemble --scan --auto=yes + # instead. mdrun does not always work. See Debian bugs 429308 + # and 354705. + #elif which mdrun > /dev/null 2> /dev/null ; then + # LogIt "Running 'mdrun'" 1 + # mdrun + elif which mdadm > /dev/null 2> /dev/null ; then + LogIt "Using mdadm to assemble any raids" 1 + mdadm --assemble --scan --auto=yes else - LogIt "Warning: Neither 'raidstart' nor 'mdrun''found. RAID devices may not have started." 1 + LogIt "Warning: Neither 'raidstart' nor 'mdadm' were found. RAID devices may not have started." 1 fi } -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]