Skip to content

Conversation

@julio-santiesteban
Copy link
Member

These are not supported on some platforms

These are not supported on some platforms
@apollolemonte
Copy link
Member

Even if they are not supported, I think the direct option should be the default to measure real disk performance and not buffered IO.

@julio-santiesteban
Copy link
Member Author

julio-santiesteban commented Nov 23, 2019

measure real disk performance and not buffered IO.

direct = 0 is not buffered IO. direct = 1 is buffered. And both are supported, we're only talking about our example here (which I changed to non buffered for the sake of the e2e tests).

@julio-santiesteban
Copy link
Member Author

@apollolemonte @otto-von-bivouac One more note here: direct = false (0, non-buffered) is actually the default: https://fio.readthedocs.io/en/latest/fio_doc.html#cmdoption-arg-direct

@otto-von-bivouac
Copy link
Member

measure real disk performance and not buffered IO.

direct = 0 is not buffered IO. direct = 1 is buffered. And both are supported, we're only talking about our example here (which I changed to non buffered for the sake of the e2e tests).

The documentation states otherwise:

direct=bool
If value is true, use non-buffered I/O. This is usually O_DIRECT. Note that OpenBSD and ZFS on Solaris don’t support direct I/O. On Windows the synchronous ioengines don’t support direct I/O. Default: false.

That is, if you set direct to false we are going to measure page cache performance.

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

Successfully merging this pull request may close these issues.

3 participants