Skip to content

Latest commit

 

History

History
46 lines (34 loc) · 2.39 KB

memory-usage.md

File metadata and controls

46 lines (34 loc) · 2.39 KB

Memory Metrics

Memory Usage

Is free memory really free? What's the difference between cached memory and buffers? Why can't I dump all the caches? Let's get our hands dirty and find out...

Task M1: Memory usage, Caches and Buffers

  1. Fire up top, and write down how much free memory you have (keep it running for the rest of this module)
  2. Start the memory hog scripts/memory/hog.sh, let it run until it gets killed (if it hangs- use Ctrl+c)
  3. Compare that to the number you wrote. Are they (almost) the same? If not, why?
  4. Read about the buffer and cached Mem values in man 5 proc (under meminfo)
    1. Run the memory hog scripts/memory/hog.sh
    2. Write down the buffer size
    3. Now run the buffer balloon sudo scripts/memory/buffer.sh
    4. Check the buffer size again
    5. Read the script, and see if you can make sense of the results...
    6. Repeat the 5 steps above with the cached Mem value and sudo scripts/memory/cache.sh
  5. Let's see how cached Mem affects application performance
    1. As root, drop the cache using echo 3 > /proc/sys/vm/drop_caches
    2. Time a dummy Python application time python -c 'print "Hello World"' (you can repeat these 2 steps multiple times)
    3. Now re-run our dummy Python application, but this time without flushing the cached memory. Can you see the difference?
  6. Run the memory/dentry.py script and observe the memory usage using free. What is using the memory? How does it effect performance? What tools can show you kernel memory usage?
  7. Run the memory/dentry2.py script and try dropping the caches. Does it make a difference? what's the difference between dentry.py and dentry2.py?

Discussion

  • Why wasn't our memory hog able to grab all the cached memory?

  • What will happen if we remove the following line from scripts/memory/hog.sh (Try it!)? Why?

     tmp[0] = 0;

  • Assuming a server has some amount of free memory, can we assume it has enough memory to support it's current workload? If not, why?

Tools

  • Most tools use /proc/meminfo to fetch memory usage information.
    • A simple example is the free utility
  • To get usage information over some period, use sar -r <delay> <count>
    • Here you can also see how many dirty pages you have (try running sync while sar is running)
    • The %commit field is also interesting, especially if it's larger than 100...

Next: IO Usage