I'm running Slackware 13 with a custom kernel based off of 2.6.32.3. I tend to leave my system on 24/7, as well as my web browser. Originally it was Firefox and now it is Google's Chrome. Usually about a day of leaving the web browser open my HD activity spikes so high that I can barely do anything on the system until I kill the web browser. This has been happening with both Firefox AND Chrome! As soon as the browser processes are killed, the system returns back to normal. Anyone have any suggestions on this?

Limit memory, cpu and swap usage of the browser... nice

what addons do you have?

That doesn't seem normal at all. First of all, browsers don't normally use the HDD a whole lot. Second, even if they were to use it so some extent, you should still be able to use the rest of your system normally.

So,

First, I would eliminate the chances of there being any kind of virus or malware on your system. I know that Linux tends not to have problems with these, but you never know. So run a virus scanner clamav, as well rootkit finders like chkrootkit and rkhunter. Also, look for anything suspicious, check 'netstat', 'nmap localhost', 'lsof', etc. Make sure you have a firewall running.

Second, what filesystem are you using ? In some cases it may have to do with filesystem and I/O scheduler, but it's more rare.

Also, check the RAM usage, make sure it isn't some type of memory leak, which tends to exert its effects after leaving the system on for extended periods of time. See the output of 'top' as it may help.

It just happened again, this time with less than a day's worth of the browser being open. I looked at the stats in top and did not see anything out of the ordinary! RAM and swap usage did not spike, nor did the amount of memory used for the chrome processes listed in top. Only addon I have for chrome is the adblock to block ads. I'm running on an ext4 filesystem that originally was an ext3 filesystem. Is there any way to find out what is accessing the disk at any given time in Linux?

iotop would probably do. Works like top - has batch option and count. Stick it in a loop with a sleep, and write it out to a file - I like to use "date" each iteration as well.

try atop as well
http://www.atoptool.nl/downloadatop.php

I keep my browsers open 24/7, with a current uptime of 19 days, x86_64, 2.6.32.8, ext4. Firefox exhibits this behavior because of the (un)awesome bar - but only when typing in the address bar. Once your history gets so large, Firefox becomes a fat bloated pig. You can fix this by tuning some options in about:config or limiting your history. But I doubt this is the problem because you said it happens with Chrome as well.

I'd look to see if it happens at the same time everyday and check for a cron job, or some other service running, or an application doing chores in the background. My email application is known to hit my system pretty hard with all the mailing lists, news groups, rss feeds and email accounts it retrieves and filters 4 times a day.

There is a memory leak with KDE 4 that causes plasma-desktop to chew away at ram over time. After 6-7 days, I need to kill and restart plasma-desktop. I've seen the application use up to 5GiB of ram. Currently after less than 48hours since the last time I killed it, it has ballooned to a 425MiB footprint, compared to the normal ~35MiB footprint.

Another thought is plugins. Remove all plugins, and add them back one by one to see if the culpurt is found. The only plugins I have are flash, java, and the default mozilla svg plugin. Gxine, and the improperly located skype plugin get nuked right off the bat.

Still attempting to figure out what exactly is causing the high disk usage...

disturbed1, do you know if KDE 4.2.4 still has this memory leak issue with plasma? How does one easily check how much memory a specific process/application is using? With top I rarely see plasma in there and when I do, it flashes by too quickly.

i use ctrl esc to open System Activity when my machine is being a pain can sort by cpu or memory then and find whats causing the problems.

What data did the two tools suggested above provide ?. How was it inadequate ?.

EDIT: didn't read top post carefully enough.