Computer Class Homework Help Needed (XP vs Fedora Core)

[Larry_OHF]

Today's assignment is a lab where we are to look into what Disk Thrashing is and what to do about it. The class that I am in studies the similarities and differences with WinXP and Fedora Core 4. Below, I have my write-up but I am missing some important information to finish.

I need to know three techniques that can be used to solve the problem and also what the walk-through would be in Core 4 (I have it at work, but wanted to do this today).

For the techniques, I assume one would get more memory, increase Virtual memory (page filing) and/or end those processes that are hogs. Those don't sound like "techniques" but is there more to it than that? Anyway, let me know what you think.

Disk Thrashing

According to the website www.computerhope.com/jargon/t/thrash.htm”, thrashing or disk thrashing is the popular term given to the result of a hard drive being overworked as virtual memory is read and wrote excessively. This occurs primarily when the computer’s physical memory modules are not enough to handle the work load. However, if the user has not established proper settings in the system properties/advanced/system performance area, then problems could come from that as well. The two indicators that a user would first notice would be system performance issues as well as the red led on the front of the computer excessively flashing or nearly staying on.

According to http://www.watchingthenet.com/identi...n-windows.html, if the user is on a Windows OS, then he or she can open the Task Manager (CTRL+ALT+DEL) and click the Processes tab. Under “View”, choose “Select Columns”. Then find the options I/O Reads, I/O Writes, I/O Read Bytes and I/O Write Bytes. You can then view the work load that the hard drive is being put through and which processes are the culprits. From there, one can make the determination on what needs to be done.

[Elif Godson]

msconfig . exe from the run prompt can be your friend in this case as well, it allows to bring up the start up menu without going into the registry and removing some of those pesky start up processes that are not needed as well as have a look at the services that are running.

[Larry_OHF]

BTW...I have just learned that the command ltag in Fedora is vmstat (virtual memory statistics), and I think this is for Core4.
Does the TOP command in Fedora have anything in common with Task manager, or is it more like Command Line?

[Felix The Assassin]

felix@Steiner:~$ vmstat
procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
2 0 0 529040 14632 212316 0 0 248 39 288 845 11 3 78 8


felix@Steiner:~$ TOP
bash: TOP: command not found

Ubuntu 8.10.2 base with Xubuntu 8.10 desktop installed.

[Larry_OHF]

This is what the instructions say.

Determine which tabs in the Windows Task Manager and which fields in the output of the Top command help you diagnose disk thrashing.
Find at least three techniques for correcting the situation.

There is no Top command in Windows, and since this class works with Fedora Core 4 half the time, I have to assume that the instructor meant that Top is in Core 4.

I'll keep looking on the net.

[Bungleau]

For reference, Felix, the command in Linux would be top, not TOP. It exists on my old Redhat box, and I'll wager it's there on yours as well.

Task manager shows the same kind of information that top does, Larry.

As for your techniques, I'd generally agree. Increase virtual memory / swap space, kill / fix the offending processes, and add more memory (which really just buffers you a bit more).

I also tend to set the virtual memory space to a fixed value, not a variable one. It doesn't impact my performance, but to my mind, it should impact overall disk fragmentation. Variable VM means that Windows will expand VM as needed, grabbing more VM space from disk as needed. It then gives that back later, which in my mind will lead to increased fragmentation during and after usage.

[Felix The Assassin]

Ok, TOP does not equal top!: Since I moved to the Xubuntu desktop, I have become degraded with the terminal. Should have thought about the 'caps' on your post, but a bit rusty.

top= "Provide information (frequently refreshed) about the most CPU-intensive processes currently running."

the -d10 behind the command is delay update by 10, thereby giving me enough time to capture the first read.

Originally Posted by felix@Steiner:~$ top -d10 top - 09:29:04 up 38 min, 2 users, load average: 0.09, 0.09, 0.11 Tasks: 97 total, 1 running, 96 sleeping, 0 stopped, 0 zombie Cpu(s): 9.1%us, 2.5%sy, 0.3%ni, 84.0%id, 4.1%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 1033580k total, 575936k used, 457644k free, 44748k buffers Swap: 3020180k total, 0k used, 3020180k free, 227924k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 5447 felix 20 0 207m 93m 25m S 4.0 9.2 3:00.71 firefox 1 root 20 0 3056 1888 564 S 0.0 0.2 0:01.42 init 2 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kthreadd 3 root RT -5 0 0 0 S 0.0 0.0 0:00.00 migration/0 4 root 15 -5 0 0 0 S 0.0 0.0 0:00.02 ksoftirqd/0 5 root RT -5 0 0 0 S 0.0 0.0 0:00.00 watchdog/0 6 root 15 -5 0 0 0 S 0.0 0.0 0:00.02 events/0 7 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 khelper 46 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kintegrityd/0 48 root 15 -5 0 0 0 S 0.0 0.0 0:00.04 kblockd/0 50 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kacpid 51 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kacpi_notify 145 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 cqueue 149 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kseriod 190 root 20 0 0 0 0 S 0.0 0.0 0:00.00 pdflush 191 root 20 0 0 0 0 S 0.0 0.0 0:00.02 pdflush 192 root 15 -5 0 0 0 S 0.0 0.0 0:00.00 kswapd0

BTW: Fedora Core 4 is like ancient! Why not use something a little newer?

[Felix The Assassin]

Time check, the atomic clocks is reading: 09:51:38

and the post time is reading 09:57. WOW! That explains the lag.

[Bungleau]

Unix is so picky about that capitalization stuff...

[MagiK]

Originally Posted by Bungleau Unix is so picky about that capitalization stuff...

Yeah ... you like, you know...have to actually KNOW what you are doing in UNIX

oh and LINUX isn't really UNIX...and in AIX "top" is known as "topas" cause IBM just HAS to be different.
Larry feel free to drop me email if you have UNIX issues, I can help out usually

When using top I would skip the delay parameter, never had to use it myself.
If you watch the percentage of disk usage and you see any devices at 100%
you can guess that there is a disk bottleneck....not working today so dont have screenies for ya. sorry.

In modern SAN environments though you arent actually writing to "Disks" everything is being written to cache memory so the disk devices showing 100% are actually pointing to a bandwidth issue on the interfaces....but I digress.