Project

General

Profile

Actions

Bug #74501

open

"IOPS is not within the threshold limit" errors

Added by Yuri Weinstein about 2 months ago. Updated 2 days ago.

Status:
Pending Backport
Priority:
High
Category:
-
Target version:
% Done:

0%

Source:
Q/A
Backport:
tentacle, squid, reef
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(RADOS):
OSD
Pull request ID:
Tags (freeform):
backport_processed
Fixed In:
v20.3.0-4944-g84d5b442ed
Released In:
Upkeep Timestamp:
2026-01-23T17:18:59+00:00

Description

Run: https://pulpito.ceph.com/yuriw-2026-01-21_19:35:54-orch-reef-release-distro-default-trial/
Jobs: see 125 failed
Logs: https://pulpito.ceph.com/yuriw-2026-01-21_19:35:54-orch-reef-release-distro-default-trial/

failure_reason: '"2026-01-21T19:45:42.843782+0000 osd.0 (osd.0) 3 : cluster [WRN]
  OSD bench result of 82891.839881 IOPS is not within the threshold limit range of
  1000.000000 IOPS and 80000.000000 IOPS for osd.0. IOPS capacity is unchanged at
  21500.000000 IOPS. The recommendation is to establish the osd''s IOPS capacity using
  other benchmark tools (e.g. Fio) and then override osd_mclock_max_capacity_iops_[hdd|ssd]." 
  in cluster log'

Related issues 4 (4 open0 closed)

Copied to RADOS - Backport #74535: reef: "IOPS is not within the threshold limit" errorsIn ProgressSridhar SeshasayeeActions
Copied to RADOS - Backport #74536: squid: "IOPS is not within the threshold limit" errorsIn ProgressSridhar SeshasayeeActions
Copied to RADOS - Backport #74537: tentacle: "IOPS is not within the threshold limit" errorsIn ProgressSridhar SeshasayeeActions
Copied to RADOS - Bug #74567: "IOPS is not within the threshold limit" errors -- continuationPending BackportSridhar Seshasayee

Actions
Actions #2

Updated by Laura Flores about 2 months ago

  • Project changed from Ceph to RADOS
Actions #3

Updated by Laura Flores about 2 months ago

/a/lflores-2026-01-21_20:56:39-rados-main-distro-default-trial/11813

2026-01-22T01:02:20.231 INFO:teuthology.orchestra.run.trial057.stdout:2026-01-22T00:27:10.881464+0000 osd.2 (osd.2) 3 : cluster [WRN] OSD bench result of 88144.544124 IOPS is not within the threshold limit range of 1000.000000 IOPS and 80000.000000 IOPS for osd.2. IOPS capacity is unchanged at 21500.000000 IOPS. The recommendation is to establish the osd's IOPS capacity using other benchmark tools (e.g. Fio) and then override osd_mclock_max_capacity_iops_[hdd|ssd].

Actions #4

Updated by Laura Flores about 2 months ago

  • Assignee deleted (Laura Flores)
  • Priority changed from Normal to High
Actions #5

Updated by Neha Ojha about 2 months ago

  • Assignee set to Sridhar Seshasayee
Actions #6

Updated by Adam Kupczyk about 2 months ago

It is possible that BlueStore is lying and not really doing IOPS, so its very fast.
But I have observed in the past BlueStore doing 70k IOPS in specific scenarios.
I would assume that BlueStore is really doing 80k+ and the only thing we need to fix is move the upper limit.

Actions #7

Updated by Sridhar Seshasayee about 2 months ago

  • Status changed from New to Fix Under Review
  • Pull request ID set to 67058
  • Component(RADOS) OSD added

Until we can better understand this, I have raised https://github.com/ceph/ceph/pull/67058
to skip the benchmark tests in both rados and orch/cephadm suites.

Teuthology testing on main with the PR on both rados and orch/cephadm suites don't show this
warning. See the associated PR for more details on the teuthology runs.

Actions #8

Updated by Sridhar Seshasayee about 2 months ago

  • Backport set to tentacle, squid, reef
Actions #9

Updated by Upkeep Bot about 2 months ago

  • Status changed from Fix Under Review to Pending Backport
  • Merge Commit set to 84d5b442ed1870ea9a4f09035e0073388ab82861
  • Fixed In set to v20.3.0-4944-g84d5b442ed
  • Upkeep Timestamp set to 2026-01-23T17:18:59+00:00
Actions #10

Updated by Upkeep Bot about 2 months ago

  • Copied to Backport #74535: reef: "IOPS is not within the threshold limit" errors added
Actions #11

Updated by Upkeep Bot about 2 months ago

  • Copied to Backport #74536: squid: "IOPS is not within the threshold limit" errors added
Actions #12

Updated by Upkeep Bot about 2 months ago

  • Copied to Backport #74537: tentacle: "IOPS is not within the threshold limit" errors added
Actions #13

Updated by Upkeep Bot about 2 months ago

  • Tags (freeform) set to backport_processed
Actions #14

Updated by Radoslaw Zarzynski about 2 months ago

The merged PR is a workaround for polluting tons of QA jobs with the warning. Let's follow up per https://tracker.ceph.com/issues/74501#note-7.

Actions #15

Updated by Radoslaw Zarzynski about 2 months ago

  • Copied to Bug #74567: "IOPS is not within the threshold limit" errors -- continuation added
Actions #16

Updated by Shraddha Agrawal about 2 months ago

https://qa-proxy.ceph.com/teuthology/yuriw-2026-01-21_20:22:09-rados-wip-yuri9-testing-2026-01-21-1558-tentacle-distro-default-trial/

9 jobs: ['11686', '11770', '11655', '11537', '11576', '11616', '11681', '11693', '11733']

Actions #17

Updated by Kamoltat (Junior) Sirivadhna about 1 month ago

/a/skanta-2026-01-29_13:10:13-rados-wip-bharath6-testing-2026-01-29-0855-squid-distro-default-trial/ [25744, 25745, 25750, 25755, 25758, 25761, 25763, 25770, 25777]

Actions #18

Updated by Aishwarya Mathuria about 1 month ago

/a/yuriw-2026-02-17_20:43:43-rados-wip-yuri6-testing-2026-02-17-1732-squid-distro-default-trial/
['53867', '53882', '53891', '53887', '53876', '53909', '53901', '53859']

Actions #19

Updated by Lee Sanders 2 days ago

/a/yuriw-2026-02-26_01:51:16-rados-wip-yuri9-testing-2026-02-25-1600-squid-distro-default-trial/
['71007', '70887', '70972', '71086', '71033', '71047', '70929', '71125', '70894']

/a/yuriw-2026-02-26_16:03:22-rados-wip-yuri9-testing-2026-02-25-1600-squid-distro-default-trial/
['72548', '72564', '72545', '72553', '72533', '72551', '72559', '72540', '72534']

Actions

Also available in: Atom PDF