I was adjusting the bdflush parameters with the Kernel Tuning app and figured out, too late, that there's no "reset to default" function. It's so unlike me to mess with something like that without recording the initial values first, but I did.

I would greatly appreciate it if someone could tell me the default values for Redhat V9, the app is on the menu under System Tools -> More System Tools -> Kernel Tuning. Within the app, it's under Virtual Memory -> bdflush. The parameters are:

Activate bdflush when % dirty
Dirty blocks to write per wake cycle
Try to reclaim n buffers on refill
Dirty buffer threshold
Time to age normal buffer before flush
Time to age superblock before flush

Activate bdflush when % dirty = 50
Dirty blocks to write per wake cycle = 500
Try to reclaim n buffers on refill = 0
Dirty buffer threshold = 0
Time to age normal buffer before flush = 3000
Time to age superblock before flush = 80

Have fun. ;P

Thank-you!

Also, the following occurs to me:

The redhat-config-proc tool you were using makes changes to the /proc/sys directory. Proc and all of its subdirs are not actually resident on your hard disk, but are generated dynamically by the kernel in RAM. As such, any changes you make do not survive a reboot. The way changes are made persistent across reboots is to modify /etc/sysctl.conf, which is referenced at boot time by rc.sysinit. This is (I assume) what redhat-config-proc does when you tell it to save settings.

In other words, if you back up the default sysctl.conf file, any changes you make can be easily undone by restoring /etc/sysctl.conf and rebooting the system. No need for tedious recording of all the values you change.

The contents of the default /etc/sysctl.conf file are attached for your convenience.

-----------
# Kernel sysctl configuration file for Red Hat Linux
#
# For binary values, 0 is disabled, 1 is enabled. See sysctl(8) and
# sysctl.conf(5) for more details.

# Controls IP packet forwarding
net.ipv4.ip_forward = 0

# Controls source route verification
net.ipv4.conf.default.rp_filter = 1

# Controls the System Request debugging functionality of the kernel
kernel.sysrq = 0

# Controls whether core dumps will append the PID to the core filename.
# Useful for debugging multi-threaded applications.
kernel.core_uses_pid = 1

Noted.

One issue that came up when I entered the values:

Activate bdflush when % dirty = 50
Dirty blocks to write per wake cycle = 500
Try to reclaim n buffers on refill = 0
Dirty buffer threshold = 0
Time to age normal buffer before flush = 3000
Time to age superblock before flush = 80

is that the two that were zero automatically changed to 64 & 256 respectively, and the superblock flush changed from 80 to 100.

Another thing I noticed is that there are 7 entries on the bdflush
line in /etc/sysctl.conf even though the there are only 6 shown in the app.

vm.bdflush =50 500 64 256 500 3000 100

Interesting. Well, I'm afraid I'm not enough of a kernel tweaking expert to offer much more than this:

1) If you replace sysctl.conf and reboot, you'll definately come back up with the out-of-the-box defaults.

2) The reference guide (or maybe the customization guide) has a chapter that covers just about every /proc/sys/* file, what it's for and what values/syntax it accepts. Both guides are available online in the docs and support section of Red Hat's website.