Slow until e2fsck does something

Slow until e2fsck does something

Post by Eric Smi » Mon, 21 Nov 1994 14:13:37

I've noticed that my system runs sluggishly until e2fsck does something.
What I mean is during a boot the sequence pauses sometimes at the e2fsck
line while it checks whatever on the disk. After the boot is complete,
the system runs noticeably faster a boot sequence when e2fsck does not
check the disk. Anyone ever see this? Is there something in my configuration
which could be set wrong.

the machine is a: 486' 50MHz, 32 MB RAM, 540 MB disk.

thanks in advance


1. am I doing something wrong with badblocks / e2fsck?

I'm using Mandrake 8.1 (e2fsprogs-1.24a-2mdk).  I have an IDE-100 hard drive, an
IBM Deskstar, which has apparently developed some kind of hardware error.  When
reading certain spots on the disk, it makes funny noises and messages about read
failures appear in the kernel error stream.  I tried running e2fsck with the -c
option to scan for bad blocks, and it supposedly marked the blocks.  Then it
went through the filesystem and supposedly corrected the directory errors that
resulted from the missing blocks.  When I remounted the filesystem, I had a few
files in the lost+found directory as expected.  However, when I tried to read
them, I found that one of them was still residing on a bad block!  Maybe it's
just my inexperience (I've only had to run fsck to fix soft errors before) but I
expected that if it was on a bad block it would've been done away with.  I
continued using the drive for a few weeks, when suddenly it started reporting
errors again when I tried to save a large amount of data.  So I decided that
perhaps the read-only scan in e2fsck missed some bad blocks.  I ran badblocks
with the option to do a non-destructive read/write test and save the block list
to a file.  Nine hours later (this is a 30-gig partition) it had indeed detected
more blocks than e2fsck had.  But when I ran e2fsck with the option to load the
bad block list from a file, it complained that the block numbers were all out of
range!  Huh?  Shouldn't the two programs be using the same numbering system?
Can anyone enlighten me?  I plan on replacing this drive eventually, but I was
hoping to get it into a stable state (assuming that the bad block problem isn't
continuously growing) until then.

2. Connectivity Help

3. SMB slow, Samba slow, or something else slow?

4. Millions of Colors

5. network speed slows until restart

6. 75/100 DPI fonts kills my X-server

7. Sol2.3 performance slows until reboot

8. having to run dhcpcd twice to get an IP

9. SCO 5.0.2 NSF is slow - VERY slow What am I doing wrong??

10. Redhat 8--Intellimouse very slow until System Settings

11. Drive I/O freezes system until done

12. Stuck on win2000 until sil3112 support is done

13. Something is slowing things down ?