On Mon, 13 Jan 2003, Martin Stricker wrote: > "Justin F. Kuo" wrote: > > > I ran that command as: > > > > fsck /usr > > You didn't unmout the filesystem before? BAD idea! File writes could > have been in disk cache, thus detroying more files. Always unmount (or, > if you absolutely have to mount it, like /, mount it read-only) the > filesystem before running fsck! > > > and saw numerous messages silimar to these: > > > > /usr contains a file syste with error, check forced. > > Pass1: Checking inodes, block sizes > > Pass2: Checking directory structures > > Missing ".." in inode 2260993 > > Fin<y>? yes > > So your filesystem was severely damaged. Now you need to find the > reason. Since for now no more errors show up, things like loose or > oxidated cable connectors or damaged IDE or SCSI controllers are not > very likely. So look in /var/log/messages for anything unusual at the > beginning of the problems (did you power off the computer without > shutting down Linux before? Power outage?) and run the command badblocks > which looks for bad blocks on your drive. > > If it reports much bad blocks, you are likely experiencing harddisk > degradation. The best you can do then is to make a backup of all your > data and get a new harddisk (reclaim the harddisk if it's still under > warranty!). When bad blocks show up they usually increase in number > rather fast, and you will more and more data. > > If there are no (or a lot less than I/O errors were in your initial ls > command) bad blocks, the harddisk might be healthy, but keep an eye on > it (and fsck the other partitions on that drive after unmounting them). > To reinstall the missing programs, verify your RPMs like suggested by > Michael Schwendt, or, probably easier, make a backup of your data and > simply reinstall. Mount? Unmount? Looks like I have a lot to learn about Linux. I checked and found no badblocks on the /dev/hda2 filesystem. I'll check the rest of the filesystems. # /sbin/badblocks -o bb-hda.txt -v /dev/hda2 Checking for bad blocks in read-only mode From block 0 to 41496840 Pass completed, 0 bad blocks found. I looked atthe /var/log/messages and found unusual activity on Jan 7. We lost power at 8PM. The message logs show several "Clearing orphaned inode" messages: Jan 7 19:59:38 mango fsck: /home: recovering journal Jan 7 19:59:44 mango fsck: /home: Clearing orphaned inode 1520756 (uid=500, gid=500, mode=0100600, size=10682) Jan 7 19:59:44 mango fsck: /home: Clearing orphaned inode 1520755 (uid=500, gid=500, mode=0100600, size=83042) Jan 7 19:59:44 mango fsck: /home: Clearing orphaned inode 1520754 (uid=500, gid=500, mode=0100600, size=12766) [snip] Jan 7 19:59:44 mango fsck: /home: Clearing orphaned inode 1520743 (uid=500, gid=500, mode=0100600, size=255806) Jan 7 19:59:44 mango fsck: /home: Clearing orphaned inode 1520742 (uid=500, gid=500, mode=0100600, size=3537) Jan 7 19:59:44 mango fsck: /home: Clearing orphaned inode 1520741 (uid=500, gid=500, mode=0100600, size=1063) Jan 7 19:59:44 mango fsck: /home: clean, 851/1896832 files, 240621/3791340 blocks Jan 7 19:59:45 mango fsck: /usr: recovering journal Jan 7 19:59:51 mango fsck: /usr: clean, 197117/5193728 files, 1170879/10374210 blocks Since the /var/log/messages are too large for the newsgroup, I posted them a web site. The errors shown here may be found at: http://www.usatfne.org/~jkuo/messages.1 I thought the boot process automatically cleared the disk problems and that all was well at the time. Did the power failure cause my problems? I'll add a UPS. What can I do in the if the Linux box looses power in the future? Thank you. -- Justin -- Psyche-list mailing list Psyche-list@redhat.com https://listman.redhat.com/mailman/listinfo/psyche-list
