Skip to content

crimson: add MOSDPGPCT support#62619

Merged
athanatos merged 3 commits intoceph:mainfrom
athanatos:sjust/wip-replica-read-crimson-mosdpct
Apr 22, 2025
Merged

crimson: add MOSDPGPCT support#62619
athanatos merged 3 commits intoceph:mainfrom
athanatos:sjust/wip-replica-read-crimson-mosdpct

Conversation

@athanatos
Copy link
Contributor

Show available Jenkins commands

@athanatos
Copy link
Contributor Author

@athanatos
Copy link
Contributor Author

"osd/pg_features: advertise PG_FEATURE_PCT for crimson" does modify a file in osd/, but only to update a constant used only in crimson. I don't feel that this requires a classic run.

@athanatos athanatos requested a review from Matan-B April 2, 2025 02:05
@athanatos athanatos force-pushed the sjust/wip-replica-read-crimson-mosdpct branch from e346fa6 to bc452b3 Compare April 2, 2025 02:14
@athanatos
Copy link
Contributor Author

The pool pct_update_delay option should already default to 5 in the rados task, it was just ignored before.

@athanatos
Copy link
Contributor Author

jenkins test make check

@athanatos
Copy link
Contributor Author

jenkins test make check arm64

@tchaikov
Copy link
Contributor

tchaikov commented Apr 2, 2025

jenkins test make check

@tchaikov
Copy link
Contributor

tchaikov commented Apr 2, 2025

jenkins test make check arm64

@athanatos athanatos force-pushed the sjust/wip-replica-read-crimson-mosdpct branch from bc452b3 to bfc72f7 Compare April 3, 2025 23:29
@athanatos
Copy link
Contributor Author

jenkins test api

* schedule/cancel
*/
class intrusive_timer_t {
using clock_t = ceph::coarse_real_clock;
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is it better to use seastar native seastar::lowres_system_clock ?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why can't we use seastar::timer<seastar::lowres_clock> instead? It allows canceling and rearming a callback, are there any other advantages to intrusive_timer_t which I'm missing?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nope, that works perfectly -- I just missed it. I've repushed with seastar::timer and am running another test set.

* schedule/cancel
*/
class intrusive_timer_t {
using clock_t = ceph::coarse_real_clock;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why can't we use seastar::timer<seastar::lowres_clock> instead? It allows canceling and rearming a callback, are there any other advantages to intrusive_timer_t which I'm missing?

@Matan-B Matan-B added this to Crimson Apr 7, 2025
@Matan-B Matan-B moved this to In Progress in Crimson Apr 7, 2025
@athanatos athanatos force-pushed the sjust/wip-replica-read-crimson-mosdpct branch from bfc72f7 to 1df6b82 Compare April 8, 2025 02:38
@athanatos
Copy link
Contributor Author

Repushed using seastar::timer as @Matan-B suggests, retesting.

@Matan-B Matan-B moved this from In Progress to Needs QA in Crimson Apr 8, 2025
@Matan-B
Copy link
Contributor

Matan-B commented Apr 8, 2025

jenkins test make check arm64

@Matan-B
Copy link
Contributor

Matan-B commented Apr 8, 2025

@athanatos
Copy link
Contributor Author

@athanatos
Copy link
Contributor Author

jenkins test make check arm64

Signed-off-by: Samuel Just <sjust@redhat.com>
Signed-off-by: Samuel Just <sjust@redhat.com>
Signed-off-by: Samuel Just <sjust@redhat.com>
@athanatos athanatos force-pushed the sjust/wip-replica-read-crimson-mosdpct branch from 1df6b82 to 666cafc Compare April 21, 2025 16:31
@athanatos
Copy link
Contributor Author

athanatos commented Apr 21, 2025

https://pulpito.ceph.com/sjust-2025-04-21_20:00:18-crimson-rados-wip-sjust-crimson-testing-2025-04-21-1745253568-distro-default-gibba/

2/51 failures:

@athanatos
Copy link
Contributor Author

jenkins test api

@Matan-B
Copy link
Contributor

Matan-B commented Apr 22, 2025

https://pulpito.ceph.com/sjust-2025-04-21_20:00:18-crimson-rados-wip-sjust-crimson-testing-2025-04-21-1745253568-distro-default-gibba/

2/51 failures:

8250534 seems like a duplicate of the backfill stuck issue - IIUC the backfill fixes were part of this run as well?
I've updated the tracker.

@athanatos athanatos merged commit cd82bfc into ceph:main Apr 22, 2025
12 checks passed
@Matan-B Matan-B moved this from Needs QA to Merged in Crimson Apr 23, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

Status: Merged (Pre Tentacle Freeze)

Development

Successfully merging this pull request may close these issues.

4 participants