Bug 2021738 - [RADOS] Have default autoscaler profile as scale-up in RHCS 5.1
Summary: [RADOS] Have default autoscaler profile as scale-up in RHCS 5.1
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Ceph Storage
Classification: Red Hat Storage
Component: RADOS
Version: 5.1
Hardware: Unspecified
OS: Unspecified
unspecified
high
Target Milestone: ---
: 5.1
Assignee: Kamoltat (Junior) Sirivadhna
QA Contact: Pawan
Ranjini M N
URL:
Whiteboard:
Depends On:
Blocks: 2031073
TreeView+ depends on / blocked
 
Reported: 2021-11-10 05:24 UTC by Pawan
Modified: 2022-04-04 10:22 UTC (History)
17 users (show)

Fixed In Version: ceph-16.2.7-4.el8cp
Doc Type: Bug Fix
Doc Text:
.The autoscaler is set to scale-up by default in {storage-product} 5.1 With this release, the autoscaler is set to scale-up by default where it starts out each pool with minimal placement groups (PGs) and scales up PGs when there is more usage in each pool. There is also a scale-down profile, where each pool starts with ideal full-capacity PGs and only scales down when the usage ratio across the pools is not even. In a scale-down profile, the autoscaler identifies any overlapping roots and prevents the pools with such roots from scaling because overlapping roots can cause problems with the scaling process.
Clone Of:
Environment:
Last Closed: 2022-04-04 10:22:28 UTC
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Ceph Project Bug Tracker 53309 0 None None None 2021-11-19 18:30:28 UTC
Ceph Project Bug Tracker 53348 0 None None None 2021-12-07 17:17:07 UTC
Github ceph ceph pull 44032 0 None Merged pacific: pybind/mgr/pg_autoscale: revert to default profile scale-up 2021-11-24 18:13:47 UTC
Red Hat Issue Tracker RHCEPH-2226 0 None None None 2021-11-10 05:27:40 UTC
Red Hat Product Errata RHSA-2022:1174 0 None None None 2022-04-04 10:22:44 UTC

Comment 2 Kamoltat (Junior) Sirivadhna 2021-11-18 14:49:37 UTC
Hi all,

So me Neha and Josh have discussed and come to a conclusion that is better for the autoscaler to start out with a scale-up profile by default. The reason for this is that scale-down was introduced to provide a better out-of-the-box experience. However, without a feature that allows us to impose a limit on the maximum number of PGs in the device_health_metrics or .mgr pool, scale-down mode can sometimes scale the PGs too large and might run into issues such as failed pool creation due to exceeding mon_max_pg_per_osd limit and etc, e.g., https://bugzilla.redhat.com/show_bug.cgi?id=2023171.

Therefore, I have created a PR that will make scale-up the default profile for the autoscaler: https://github.com/ceph/ceph/pull/43999

Comment 3 Vikhyat Umrao 2021-11-19 17:58:14 UTC
(In reply to ksirivad from comment #2)
> Hi all,
> 
> So me Neha and Josh have discussed and come to a conclusion that is better
> for the autoscaler to start out with a scale-up profile by default. The
> reason for this is that scale-down was introduced to provide a better
> out-of-the-box experience. However, without a feature that allows us to
> impose a limit on the maximum number of PGs in the device_health_metrics or
> .mgr pool, scale-down mode can sometimes scale the PGs too large and might
> run into issues such as failed pool creation due to exceeding
> mon_max_pg_per_osd limit and etc, e.g.,
> https://bugzilla.redhat.com/show_bug.cgi?id=2023171.
> 
> Therefore, I have created a PR that will make scale-up the default profile
> for the autoscaler: https://github.com/ceph/ceph/pull/43999

Thanks, Junior. I have renamed the bug title, and also added an update in bz2023171.

Comment 11 Kamoltat (Junior) Sirivadhna 2022-01-27 04:53:35 UTC
Hey Aron,

Just added, let me know if you want me to change anything or add anything more.

Comment 13 Kamoltat (Junior) Sirivadhna 2022-02-16 15:06:00 UTC
Hi,

I would like to change the part ``starts with compliments of PGS`` -->``starts with ideal full-capacity of PGS``. Since the word `compliments` might be confusing to some people.

Thank you,

Comment 15 errata-xmlrpc 2022-04-04 10:22:28 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (Moderate: Red Hat Ceph Storage 5.1 Security, Enhancement, and Bug Fix update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHSA-2022:1174


Note You need to log in before you can comment on or make changes to this bug.