Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Weird results #29

Open
pydima opened this issue Nov 8, 2022 · 5 comments
Open

Weird results #29

pydima opened this issue Nov 8, 2022 · 5 comments

Comments

@pydima
Copy link

pydima commented Nov 8, 2022

Hi, thanks a lot for developing this tool, I think that's a fantastic idea. I've tried to run it but got really weird results genereated (screen attached). Any ideas on how to better debug this or why I've got this result? Thanks.
Screenshot from 2022-11-08 08-54-54

@avikivity
Copy link
Member

What kind of disk are you testing?

If it's a rotating disk, maybe the tool estimated read iops incorrectly. You can try with --max-read-iops 200 (adjust according to what you think the disk can bear).

@avikivity
Copy link
Member

Actually, it looks like SSD since you got 2 GB/s. What's the make and model of the disk?

Please upload the results directory with the json files somewhere.

@pydima
Copy link
Author

pydima commented Nov 8, 2022

Yes, it's SSD disk from Intel - https://ark.intel.com/content/www/us/en/ark/products/series/122570/intel-ssd-dc-p4510-series.html. I tried to run diskplorer with default iops/write bandwith (auto-discover one) and with explicitly providing the values from the disk spec - the result is the same. Here is the link to result file: https://drive.google.com/file/d/1dEASw8X7Z0Xny1xWo8xWM7ZHMUGsq_5d/view?usp=sharing

@avikivity
Copy link
Member

      "jobname": "job(r_idx=1,w_idx=0,write_bw=0,r_iops=31825)",
      "groupid": 0,
      "error": 0,
      "eta": 0,
      "elapsed": 37,
      "job options": {
        "rw": "randread",
        "bs": "4096",
        "iodepth": "1000",
        "rate_iops": "397",
        "rate_process": "poisson"
      },
      "read": {
        "io_bytes": 3534966784,
        "io_kbytes": 3452116,
        "bw_bytes": 117761569,
        "bw": 115001,
        "iops": 28748.017856,
        "runtime": 30018,

We commanded 80*397=31827 (do you have 80 threads?) and only got 28748.017856 iops. So diskplorer disqualified (diskualified?) this result.

Please pick the .fio directory corresponding to this job (r_idx=1,w_idx=0) and rerun it under fio directly, look at top/blktrace etc to see what went wrong.

@pydima
Copy link
Author

pydima commented Nov 8, 2022

There are 80 cores, so I think 80 threads should be fine. Thanks for the pointer, will try to run it!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants