Normally if the system reboots after a power failure the checkfs or checkrootfs scripts do a fsck -a to check the filesystem. In some cases the following happens:

<br></br>/dev/sda1: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY<br></br><br></br>              (i.e., without -a or -p options)<br></br><br></br><br></br><br></br>A maintenance shell will now be started.<br></br><br></br>After preforming system maintenance, press Control-D<br></br><br></br>to terminate the maintenance shell and restart the system.<br></br><br></br>Give root password for maitenance<br></br><br></br>(or type Control-D to continue)<br></br>

In normal cases you enter the root password and do a fsck of the disc to fix the errors. If you are working with a remote system if might be better (not wiser) to force a fsck -y on reboot. Warning: this might cause data loss

This can be done by changing

FSCKFIX=no to FSCKFIX=yes in /etc/default/rcS

this option is best described in man rcS:

FSCKFIX<br></br><br></br>              When  the  root  and all other file systems are checked, fsck is<br></br><br></br>              invoked with the -a option which means "autorepair".   If  there<br></br><br></br>              are  major  inconsistencies then the fsck process will bail out.<br></br><br></br>              The system will print a  message  asking  the  administrator  to<br></br><br></br>              repair  the  file  system  maually and will present a root shell<br></br><br></br>              prompt (actually a sulogin prompt) on the console.  Setting this<br></br><br></br>              option  to  yes  causes  the fsck commands to be run with the -y<br></br><br></br>              option instead of the -a option.  This will tell fsck always  to<br></br><br></br>              repair the file systems without asking for permission.<br></br>