Project

General

Profile

Actions
Copy link

Bug #47952

open

Replicated pool creation fails Nautilus 14.2.12 build when cluster runs with filestore OSDs

Added by Prashant Tambe over 3 years ago. Updated over 3 years ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Target version:
Ceph - v14.2.12
% Done:

0%

Source:
Community (user)
Tags:
Backport:
Regression:
No
Severity:
2 - major
Reviewed:
Affected Versions:
Ceph - v14.2.12
ceph-qa-suite:
ceph-ansible
Component(RADOS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

Tried pool creation using ceph-ansibles-4.0 and replication pool failed with following error :
Build : Nautilus 14.2.12 (2f3caa3b8b3d5c5f2719a1e9d8e7deea5ae1a5c6) nautilus (stable)

2020-10-22 12:21:11,473 p=16321 u=root | TASK [create replicated pool] ******************************************
2020-10-22 12:21:11,475 p=16321 u=root | Thursday 22 October 2020 12:21:11 +0000 (0:00:00.457) 0:00:04.033 **
2020-10-22 12:21:12,156 p=16321 u=root | FAILED - RETRYING: create replicated pool (3 retries left).
2020-10-22 12:21:17,618 p=16321 u=root | FAILED - RETRYING: create replicated pool (2 retries left).
2020-10-22 12:21:23,124 p=16321 u=root | FAILED - RETRYING: create replicated pool (1 retries left).
2020-10-22 12:21:28,642 p=16321 u=root | fatal: [host-01]: FAILED! => changed=true
attempts: 3
cmd:
- ceph
- osd
- pool
- create
- rgw-pool-2
- '64'
- replicated
delta: '0:00:00.348942'
end: '2020-10-22 07:21:28.598618'
msg: non-zero return code
rc: 1
start: '2020-10-22 07:21:28.249676'
stderr: 'Error EPERM: For better initial performance on pools expected to store a large number of objects, consider supplying the expected_num_objects parameter when creating the pool. Pass --yes-i-really-mean-it to ignore it'
stderr_lines: <omitted>
stdout: ''
stdout_lines: <omitted>

Note : I am not getting this error on Nautilus 14.2.11 build (f7fdb2f52131f54b891a2ec99d8205561242cdaf) nautilus (stable)

Actions
Copy link
 #1

Updated by Neha Ojha over 3 years ago

  • Project changed from Ceph to RADOS
  • Category deleted (build)

14.2.12 introduced the following change in https://github.com/ceph/ceph/pull/37474, which is probably the case you are hitting.

 
   if (has_filestore_osd &&
        expected_num_objects == 0 &&
        cct->_conf->filestore_merge_threshold < 0) {
      int osds = osdmap.get_num_osds();
      bool sure = false;
      cmd_getval(cct, cmdmap, "yes_i_really_mean_it", sure);
      if (!sure && osds && (pg_num >= 1024 || pg_num / osds >= 100)) {
        ss << "For better initial performance on pools expected to store a " 
           << "large number of objects, consider supplying the " 
           << "expected_num_objects parameter when creating the pool." 
           << " Pass --yes-i-really-mean-it to ignore it";
        err = -EPERM;
        goto reply;
      }
    }

You can ignore this recommendation by passing --yes-i-really-mean-it to pool create.

Actions
Copy link
 #2

Updated by Prashant Tambe over 3 years ago

Neha Ojha wrote:

14.2.12 introduced the following change in https://github.com/ceph/ceph/pull/37474, which is probably the case you are hitting.

[...]

You can ignore this recommendation by passing --yes-i-really-mean-it to pool create.

Thanks for the reply Neha! But option "--yes-i-really-mean-it" is not working with 'ceph osd pool create' command. Is there any other way to ignore this ?

Actions
Copy link

Also available in: Atom PDF