پروژه لینوکس(20629))=
چهار سال پیش منتشر شده
تعداد بازدید: 231
کد پروژه: 251435
شرح پروژه
همیارپروژه:
#پروژه_جدید
#کد_پروژه: 20629
موضوع: پروژه لینوکس =
سلام من میخوام بهم کمک کنید برای جواب به این سوالات که در زمینه لینوکس هست ،
Task 2.
Learn how to use the performance monitoring tools available in Linux
Linux commands such as sar , top and ps allow you to monitor the system performance
counters (or performance metrics), the state of the processes running on the system, and the
number of resources used by the processes running on the system.
Use the above-mentioned tools to answer the following questions (write your answers in a
report and for each answer also provide an example documented with e.g., screenshots or
textual output from Linux commands):
1. What are the performance metrics/counters that you can use to detect memory leaks?
1.1. What are the commands and related flags to enable the collection of that
metrics/counters?
2. What are the performance metrics/counters that you can use to analyze the CPU
utilization?
2.1. What are the commands and related flags to enable the collection of that
metrics/counters?
3. What are the performance metrics/counters that you can use to analyze disk throughput?
3.1. What are the commands and related flags to enable the collection of that
metrics/counters?
4. How can you sample performance counters each 2 seconds? (give an answer for all the
three tools)
5. How can you automatically stop the sampling after 3 minutes? (give an answer for all
the three tools)
6. Is it possible to save the sampled data into a file and load it again for further analysis?
(give an answer for all the three tools. If a command option is not available, propose a
solution)
27. How can you sample statistics for a specific processor?
8. How can you sample statistics for a process/thread or group of processes/threads? (the
use of “ | grep <process name> ” appended to a monitoring command can turn
useful)
9. How can you track a specific process and related threads?
Task 3.
Practice with measuring and analyzing the performance and scalability of
your system
stress-ng is a tool designed to stress a computer system (HW) in various selectable ways.
We will use that tool as a workload generator of stress, in a controllable way, on different
components of our system (CPU, memory, disk) to learn:
•
•
•
How to monitor the performance of the system.
How to analyze and interpret the results of the monitoring tools and of the
benchmarking tool.
How to present and explain the obtained results.
3.1. CPU performance (estimated time to collect results 8min)
Use the stress-ng options --cpu and --cpu-load to stress the CPUs in a controllable
way, and specifically to load the system at the 25%, 50%, 75%, and 100%. (when you use
--cpu-load set the number of stressors equal to the number of available CPUs).
Use sar and top to verify the load generated and to collect the performance counters for
further analysis. Also, use the stress-ng options --metrics-brief and --perf to print
the stress-ng results and statistics.
An example: stress-ng --cpu 2 --metrics-brief --perf
Run the test for a time interval >= 60s, collect the performance metrics (with a sampling
interval of 1 second) for the CPU utilization for all the CPUs and for the total CPU
utilization. Then:
•
•
•
•
Plot the time series of the CPU utilization for all the CPUs and the total CPU
utilization.
Compute the average values and the 90,95, 99 percentiles.
Do you observe any difference between the two options used to load the system?
When you use the --cpu-load option, try to sample the system load metrics every
2 seconds. Do you see any difference? What is the impact of the sampling interval
on the results that you obtain?
Use the option --backoff N to progressively load the system from (about) 0% to 100%
of the CPU utilization ( N is in microseconds, 1sec = 1000000 microseconds)
•
Try different values of N . Plot the time series and comment on the results.
3.2. Memory consumption (estimated time to collect results 15 min)
Use the stress-ng options --brk , --stack and --bigheap with the purpose of stressing
the main memory concurrently using the heap (by increasing the heap breakpoint,
consumption of the available heap) and the stack (by expanding the stack size).
•
•
Run stress-ng --brk N --stack N --bigheap N --metrics-brief --
perf -t 60s with N ranging from 2 to 8 ( 2, 4, 8 ) and collect the global memory
statistics with sar and the per-thread memory statistics with top ). Execute at least
5 runs for each test case to ensure that the numbers you get are reliable. Comment
on the results you obtained using plots to better explain your conclusions.
Run stress-ng --bigheap 2 --bigheap-growth N --metrics-brief --
perf -t 60s while increasing N from 4KB to 1MB ( 4K, 8K, 16K, 32K, 64K,
128K, 256K, 512K, 1024K ). Comment on the behavior of the two threads and
on the memory usage. Use plots to explain your conclusions better.
3.3. Disk throughput
Use the stress-ng option --hdd to analyze the write throughput of the system.
•
•
Run stress-ng --hdd N --hdd-opts wr-rnd --metrics-brief -t 20s
while increasing the value of N (from 1 to 16 ) to identify the maximum achievable
throughput. Analyze the throughput and the queue length of the disk and comment
on the scalability. Use plots to better explain your conclusions
Run stress-ng --hdd N --hdd-opts wr-rnd, sync --metrics-brief -
t 20s while increasing the value of N (from 1 to 16 ). What is the difference with
the previous test? Analyze the throughput and the queue length of the disk and
comment on the scalability. Use plots to explain your conclusions better.
دو تا تسکه یک و دو
لطفا هر کدومو جدا بهم هزینه اشو بفرمایید
من تا یک هفته زمان دارم
و اینکه توضیح مختصر با اسکرین شات از هر قسمت میخوام
مهارت ها و تخصص های مورد نیاز
مهلت برای انجام
7روز
وضعیت مناقصه
بسته
درباره کارفرما
عضویت دوازده سال پیش
