I am far from being a power user, and I simply would like to ask you guys for your opinion here... The problems are that :
- I'm not sure if I'm having a dying hard disk or not
- Maybe this is instead Lilo that is losing its marbles
- and if it's a dying disk, what might be the nux tools to try to make sure and perhaps correct disk errors
A longer description follows
My linux (mandriva powerpack 2007) is having great trouble to boot, I sometimes have to hit "reset" 8 or 9 times before the boot ends and linux is able to work. There is not always the same error message in the text console that is displayed when I hit "escape", this is more likely either a sudden reboot as it I had hit "reset", either the boot halting on screen and never resuming. The most late hang-on there was was during the nvidia splash screen, most of them are during the beginning of the system boot.
Almost 50% of the reboots the system is in read-only mode, I don't know if it's a bug or if it's a natural system security in case something serious caused the previous reboot.
There is no specific filesystem-related error message. Once of twice I got some hard to get error message during one or another step of the boot. "Superblock last write time is in the future" might be the filesystem error that I see the most often, at least once per "let's try to make a complete boot" period.
And if the boot finally made it to his end, then linux will be working perfectly fine for hours and hours, nothing wrong will happen anymore.
Lilo is installed on my OS, and when I ask Lilo to boot windows XP instead (that happens very rarely), I once got this "ascii soup" : http://sabin.free.fr/tmp/trap06.jpg
To tell how my partitions are arranged, the computer has 5 hard disks with 2 partitions each in average, the main master bootdisk has windows XP installed on it, Lilo is installed on that partition too.
Linux has its system, swap and user files installed on another hard disk composed of 3 partitions.
And here, it seems that it's the hard disk on which windows XP and Lilo are installed is dying slowly.
BUT maybe not, maybe this is only the boot sector containing Lilo that is screwed up !
BUT, how could I make sure ? :-/
I come from the windows world, I made the transition six years ago, but so far I never needed to make disk scans for errors, nor to try to repair a boot sector. Using windows XP, I know I could have rebooted in system rescue mode and ran "chkdsk -f", and next chkdsk /fixmbr, for instance... Except that windows's way of doing would erase Lilo of course.
If you know how I could try to repair the windows disk (to make it harder, it's an NTFS partition) or how I fix Lilo if I manage to test if it's the guilty part, I'd be very grateful
My thanks if you can help,