> the ext3 mini-howto refers to "certain > rare hardware failure cases (e.g. hard > drive failures)" that would require a > filesystem check, but doesn't go into > details. Hardware problems (not necessarily failures) and bugs in the ext3 code (the reason why drivers are never really finished as far as development goes; new versions always coming out to fix some issue). Do you need to run fsck nearly as often as with ext2? No. Should you completely disable it? No. You should fsck now and then just in case something strange happened along the way which you didn't see (most blips with any filesystem aren't known until an fs checker combs the system). Right now, the default (on redhat systems) seems to be 20 mounts or 6 months with ext2. If you mount often (ie. for systems that aren't up for weeks or months on end) I'd set the max mount count much higher. If your system runs for long periods of time, then maybe setting the time interval would be better.. 6 months for any system though is long enough. Also, you could disable the fsck'ing by setting the numbers to 0 (I think *might* work), but I'd still run an fsck every so often on each partition (in your spare time of course) just as a precaution. No code and no hardware is perfect... everything's affected by something called physics :) Sorry if I sounded like I was rambling a little at times (I'm half asleep) :) -Statux
