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

os/bluestore: put bluefs in the middle of the shared device #14873

Merged
merged 1 commit into from May 17, 2017

Conversation

Projects
None yet
3 participants
@liewegas
Member

liewegas commented Apr 28, 2017

If this is an HDD, this will be a shorter average seek distance for an
aged device.

If it is an SSD, it doesn't matter, so don't bother with making this
conditional on is_rotational().

Signed-off-by: Sage Weil sage@redhat.com

os/bluestore: put bluefs in the middle of the shared device
If this is an HDD, this will be a shorter average seek distance for an
aged device.

If it is an SSD, it doesn't matter, so don't bother with making this
conditional on is_rotational().

Signed-off-by: Sage Weil <sage@redhat.com>

@liewegas liewegas requested a review from ifed01 Apr 30, 2017

@liewegas liewegas requested a review from markhpc Apr 30, 2017

@liewegas liewegas changed the title from os/bluestore: put bluefs in the middle of the shared device to DNM os/bluestore: put bluefs in the middle of the shared device May 2, 2017

@liewegas

This comment has been minimized.

Show comment
Hide comment
@liewegas

liewegas May 2, 2017

Member

Need to do some more performance validation on this first!

Member

liewegas commented May 2, 2017

Need to do some more performance validation on this first!

@ifed01

ifed01 approved these changes May 3, 2017

@markhpc

This comment has been minimized.

Show comment
Hide comment
@markhpc

markhpc May 3, 2017

Member

I'll take a look today!

Member

markhpc commented May 3, 2017

I'll take a look today!

@markhpc

This comment has been minimized.

Show comment
Hide comment
@markhpc

markhpc May 4, 2017

Member

So far on an empty system where we'd expect to see the greatest drop in performance for sequential IO, I see perhaps a small decrease in sequential read performance (maybe a couple of percent), but it's small enough it could just be noise too. No obvious difference for sequential writes. I haven't tested a full cluster yet, but the feeling I'm getting is that other factors are probably going to matter more than this does and we could probably go either way with certain advantages and disadvantages.

Member

markhpc commented May 4, 2017

So far on an empty system where we'd expect to see the greatest drop in performance for sequential IO, I see perhaps a small decrease in sequential read performance (maybe a couple of percent), but it's small enough it could just be noise too. No obvious difference for sequential writes. I haven't tested a full cluster yet, but the feeling I'm getting is that other factors are probably going to matter more than this does and we could probably go either way with certain advantages and disadvantages.

@liewegas liewegas changed the title from DNM os/bluestore: put bluefs in the middle of the shared device to os/bluestore: put bluefs in the middle of the shared device May 12, 2017

@liewegas liewegas added needs-qa and removed needs-review labels May 12, 2017

@liewegas

This comment has been minimized.

Show comment
Hide comment
@liewegas

liewegas May 12, 2017

Member

I see no difference in my (limited :) test. I think we should go ahead and merge it.

Member

liewegas commented May 12, 2017

I see no difference in my (limited :) test. I think we should go ahead and merge it.

@liewegas liewegas merged commit c946455 into ceph:master May 17, 2017

2 of 3 checks passed

default Build triggered. sha1 is merged.
Details
Signed-off-by all commits in this PR are signed
Details
Unmodifed Submodules submodules for project are unmodified
Details

@liewegas liewegas deleted the liewegas:wip-bluefs-position branch May 17, 2017

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment