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

doc: pg_num should always be a power of two #29364

Merged
merged 3 commits into from Aug 1, 2019

Conversation

@capri1989
Copy link
Contributor

commented Jul 26, 2019

Hi,

I updated the pg_num section in the docs just a little to be more strict.
I think we should make it crystal clear that pg num should be a power of two and only if there's a real good reason and someone knows exactly why they need a different value they can proceed.

I think this is the first step of several to make this clearer to the end user. Next step would be to show a Health WARN and also to flag the pool creation red in the dashboard if someone would like to add a pool and the pg count isn't a power of two.

Why do I would like to be more strict about it? We've seen more and more clusters were the osd variation was really high. The fix to that was always to set the pg_num + balancer. Also the pg autoscaler only uses power of two so we shouldn't even encourage people to use something different right from the beginning or to phrase i differently, we should protect the end-user right away.

FIxes: https://tracker.ceph.com/issues/41004
Signed-off-by: Kai Wagner kwagner@suse.com

@capri1989 capri1989 requested review from jan--f, l-mb and liewegas Jul 26, 2019

@capri1989 capri1989 force-pushed the capri1989:doc-pg-num branch from 6d1630c to c2a188c Jul 26, 2019

@capri1989

This comment has been minimized.

Copy link
Contributor Author

commented Jul 26, 2019

@liewegas Thanks for your feedback. Addressed both comments

@liewegas
Copy link
Member

left a comment

otherwise lgtm

@capri1989 capri1989 force-pushed the capri1989:doc-pg-num branch from c2a188c to 471e008 Jul 26, 2019

@jan--f

jan--f approved these changes Jul 29, 2019

Copy link
Member

left a comment

👍 for @l-mb's comments, otherwise lgmt.

@jan--f

This comment has been minimized.

Copy link
Member

commented Jul 29, 2019

We should probably backport this as well? If so please open a tracker ticket and mention it in your commit @capri1989.

@capri1989 capri1989 force-pushed the capri1989:doc-pg-num branch from 471e008 to 8a5e31a Jul 30, 2019

doc: pg_num should always be a power of two
Signed-off-by: Kai Wagner <kwagner@suse.com>

@capri1989 capri1989 force-pushed the capri1989:doc-pg-num branch from 8a5e31a to 3660e17 Jul 30, 2019

@capri1989

This comment has been minimized.

Copy link
Contributor Author

commented Jul 30, 2019

@l-mb comments addressed

l-mb added some commits Jul 31, 2019

doc: adjust examples to use 2^n pg_num
The examples used pg_num set to 300 or 1000. This misled some users.

Signed-off-by: Lars Marowsky-Bree <lmb@suse.com>
doc: Address further comments on choosing pg_num
Signed-off-by: Lars Marowsky-Bree <lmb@suse.com>
@theanalyst

This comment has been minimized.

Copy link
Member

commented Aug 1, 2019

jenkins render docs

@theanalyst

This comment has been minimized.

Copy link
Member

commented Aug 1, 2019

jenkins test docs

@theanalyst theanalyst merged commit 067db76 into ceph:master Aug 1, 2019

4 of 6 checks passed

Docs: render build Docs: render failed with errors
Details
make check (arm64) make check failed
Details
Docs: build check OK - docs built
Details
Signed-off-by all commits in this PR are signed
Details
Unmodified Submodules submodules for project are unmodified
Details
make check make check succeeded
Details
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
6 participants
You can’t perform that action at this time.