sleddog
03-13-2001, 03:25 PM
Last night was a nightmare.
Sitting in my home office, I noticed the console of my Linux server fill up with error messages. The all pertained to HDD errors on one partition -- /usr -- where most software packages are installed. Binaries couldn't be read or couldn't be executed.
After 20 minutes or so of trying one thing and another, I turned to my NT workstation to find some particular Linux documentation that was saved there. In a minute the NT machine blue-screened.. crashed. I rebooted it but NT errored out and wouldn't load -- reporting errors loading kernel32.dll. I rebooted in DOS and ran scandisk. The C: partition (FAT16) had many corrupted files with incorrect directory entries in c:\winnt\system32.
To make a long story short, it took over 3 hours of rebuilding to get both machines back up and working correctly (and without re-installing either OS I might add!) In both cases the HDD partitions were physically *error-free* -- no bad blocks or sectors were detected by a full scandisk in DOS or by the badblocks program in Linux. These partition are back in use now and working fine.
What on earth could cause disk corruption like this, on two very different machines, two different OS's, at the same time? Not a virus. There are no known viri that can affect the Linux ext2 filesystem -- and certainly not one that can run on both Linux *and* NT! A power surge? Both machines are plugged into a UPS which is supposed to provide surge protection. Another machine (Win98/FAT32) was also running at the same time, not plugged into the UPS, and is just fine.
Or was it just one H*LL of a coincidence?
------------------
sleddog
[sleddog.f2s.com] (http://www.sleddog.f2s.com)
Sitting in my home office, I noticed the console of my Linux server fill up with error messages. The all pertained to HDD errors on one partition -- /usr -- where most software packages are installed. Binaries couldn't be read or couldn't be executed.
After 20 minutes or so of trying one thing and another, I turned to my NT workstation to find some particular Linux documentation that was saved there. In a minute the NT machine blue-screened.. crashed. I rebooted it but NT errored out and wouldn't load -- reporting errors loading kernel32.dll. I rebooted in DOS and ran scandisk. The C: partition (FAT16) had many corrupted files with incorrect directory entries in c:\winnt\system32.
To make a long story short, it took over 3 hours of rebuilding to get both machines back up and working correctly (and without re-installing either OS I might add!) In both cases the HDD partitions were physically *error-free* -- no bad blocks or sectors were detected by a full scandisk in DOS or by the badblocks program in Linux. These partition are back in use now and working fine.
What on earth could cause disk corruption like this, on two very different machines, two different OS's, at the same time? Not a virus. There are no known viri that can affect the Linux ext2 filesystem -- and certainly not one that can run on both Linux *and* NT! A power surge? Both machines are plugged into a UPS which is supposed to provide surge protection. Another machine (Win98/FAT32) was also running at the same time, not plugged into the UPS, and is just fine.
Or was it just one H*LL of a coincidence?
------------------
sleddog
[sleddog.f2s.com] (http://www.sleddog.f2s.com)