Project

General

Profile

Actions

Bug #47392

closed

radosgw not listening after installation

Added by yuqi hou over 3 years ago. Updated over 2 years ago.

Status:
Can't reproduce
Priority:
Normal
Assignee:
-
Target version:
% Done:

0%

Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
09/10/2020
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

I successfully deploy ceph and rgw v14.2.10 one month ago,
but fail with v14.2.11.
a strange thing is radosgw process owner is different between two versions!

 Successful v14.2.10 radosgw process owner is "ceph",
 while failed v14.2.11 radosgw process owner is "root".
I reproduce the procedure(v14.2.11):
  export CEPH_DEPLOY_REPO_URL=http://mirrors.163.com/ceph/rpm-nautilus/el7

  export CEPH_DEPLOY_GPG_URL=http://mirrors.163.com/ceph/keys/release.asc
  ceph-deploy new zx1
[ceph1@zx2 my-cluster]$ cat ceph.conf
[global]
fsid = fda8464f-c43c-44b1-ab9e-9714bb4cedd7
mon_initial_members = zx1
mon_host = 10.100.251.218
auth_cluster_required = cephx
auth_service_required = cephx
auth_client_required = cephx


  ceph-deploy install --release nautilus zx1 zx0
  ceph-deploy mon create-initial
  ceph-deploy admin zx1 zx0
  ceph-deploy mgr create zx1
  ceph-deploy osd create --data /dev/sde zx0
  ceph-deploy osd create --data /dev/sdd zx0
  ceph-deploy osd create --data /dev/sdc zx0
  ceph-deploy install --release nautilus --rgw zx0
  ceph-deploy rgw create zx0
In v14.2.10, ps -ef|grep ceph show
v14.2.10 sucessful radosgw process owner is ceph and v14.2.10 sucessful radosgw process 7480 port In v14.2.11, ps -ef|grep ceph show
v14.2.11 failed radosgw process owner is root and v14.2.11 failed radosgw log I donot know how to solve it, someone help me,thanks.

Files

suc_rgw_owner.png (71.8 KB) suc_rgw_owner.png v14.2.10 sucessful radosgw process owner is "ceph" yuqi hou, 09/10/2020 02:37 PM
suc_rgw_port.png (106 KB) suc_rgw_port.png v14.2.10 sucessful radosgw process 7480 port yuqi hou, 09/10/2020 02:37 PM
fail_rgw_owner_root.png (124 KB) fail_rgw_owner_root.png v14.2.11 failed radosgw process owner is "root" yuqi hou, 09/10/2020 02:40 PM
fail_rgw_log.png (53.9 KB) fail_rgw_log.png v14.2.11 failed radosgw log yuqi hou, 09/10/2020 02:40 PM
healthy.png (59.9 KB) healthy.png yuqi hou, 09/11/2020 02:02 AM
Actions #1

Updated by yuqi hou over 3 years ago

ceph -s is also health_ok

Actions #2

Updated by yuqi hou over 3 years ago

we had to manually change 'type host' to 'type osd' in our crushmap.

after booting rgw, I find that the owner of radosgw process is "root", not "ceph". command "ceph -s" also show that "100.000% pgs not active".

I search the clue "100.000% pgs not active", the post "https://www.cnblogs.com/boshen-hzb/p/13305560.html" tell how to solve it - change 'type host' to 'type osd', "ceph -s" show "HEALTH_OK" and the owner of radosgw process become "ceph", rgw service(7480) is in service.

Actions #3

Updated by Abhishek Lekshmanan over 3 years ago

The reason RGW is not working seems to be linked to deploying all OSDs on single host and hence no pools and pgs being created, so RGW fails to initialize, can you simply try creating pools in this cluster?

Actions #4

Updated by Casey Bodley over 3 years ago

  • Status changed from New to Need More Info
Actions #5

Updated by Casey Bodley over 3 years ago

doesn't look like a radosgw issue. are you able to create pools otherwise?

Actions #6

Updated by Casey Bodley over 2 years ago

  • Status changed from Need More Info to Can't reproduce
Actions

Also available in: Atom PDF