Bug 1911669 - [FFU OSP13 TO 16.1] Filestore to bluestore migration is not working
Summary: [FFU OSP13 TO 16.1] Filestore to bluestore migration is not working
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Red Hat OpenStack
Classification: Red Hat
Component: openstack-tripleo-heat-templates
Version: 16.1 (Train)
Hardware: Unspecified
OS: Unspecified
high
high
Target Milestone: ---
: ---
Assignee: RHOS Maint
QA Contact: Joe H. Rahme
URL:
Whiteboard:
Depends On: 1886175 1895756
Blocks:
TreeView+ depends on / blocked
 
Reported: 2020-12-30 16:11 UTC by Ravi Singh
Modified: 2021-01-20 11:29 UTC (History)
5 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2021-01-20 11:29:03 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description Ravi Singh 2020-12-30 16:11:03 UTC
Description of problem:

There are multiple BZs on this issue & some of them are fixed but the issue still persists which is mentioned in [1] but should be fixed as per [2].

THT version in my env

~~~
(undercloud) [stack@undercloud ~]$ rpm -qa | grep  openstack-tripleo-heat-templates
openstack-tripleo-heat-templates-11.3.2-1.20200914170177.el8ost.noarch

#ceph-ansible

[stack@undercloud ~]$ sudo rpm -qa | grep -i ceph-ansible
ceph-ansible-4.0.31-1.el8cp.noarch
~~~

Followed WA mentioned in our doc[3] & discussed in BZ -https://bugzilla.redhat.com/show_bug.cgi?id=1902153 but issue persists.

Steps followed:

1. Ran overcloud deploy with --stack-only option with "osd_objectstore: filestore"

~~~
cat /home/stack/templates/ceph-config.yaml
parameter_defaults:
  CephAnsibleRepo: rhceph-4-tools-for-rhel-8-x86_64-rpms
  CephAnsibleExtraConfig:
    osd_objectstore: bluestore

2. In overcloud plan filestore is set

~~~
[stack@undercloud plan]$ grep -ir objectstore .

./plan-environment.yaml:    osd_objectstore: filestore
~~~

3. Ran migration command, the command succeeds but no change is objectstore, it's still showing as 'filestore'.
   As explained in other BZ it simply skipped that task since it sees bluestore is already set.

~~~
/var/lib/mistral/b8ba0cb4-da83-4d3f-9239-2a9e4344093f/ceph-ansible/ceph_ansible_command.log
2020-12-30 10:17:15,625 p=208169 u=root n=ansible | Using /usr/share/ceph-ansible/ansible.cfg as config file
2020-12-30 10:17:16,398 p=208169 u=root n=ansible | PLAY [osds] ********************************************************************
2020-12-30 10:17:16,426 p=208169 u=root n=ansible | TASK [Gathering Facts] *********************************************************
2020-12-30 10:17:16,426 p=208169 u=root n=ansible | Wednesday 30 December 2020  10:17:16 -0500 (0:00:00.086)       0:00:00.086 **** 
2020-12-30 10:17:18,922 p=208169 u=root n=ansible | ok: [overcloud-computehci-0]
2020-12-30 10:17:19,189 p=208169 u=root n=ansible | TASK [gather and delegate facts] ***********************************************
2020-12-30 10:17:19,189 p=208169 u=root n=ansible | Wednesday 30 December 2020  10:17:19 -0500 (0:00:02.763)       0:00:02.849 **** 
2020-12-30 10:17:20,520 p=208169 u=root n=ansible | ok: [overcloud-computehci-0 -> 172.16.0.23] => (item=overcloud-controller-0)
2020-12-30 10:17:21,901 p=208169 u=root n=ansible | ok: [overcloud-computehci-0 -> 172.16.0.39] => (item=overcloud-controller-1)
2020-12-30 10:17:23,179 p=208169 u=root n=ansible | ok: [overcloud-computehci-0 -> 172.16.0.32] => (item=overcloud-controller-2)
2020-12-30 10:17:23,294 p=208169 u=root n=ansible | TASK [set_fact current_objectstore] ********************************************
2020-12-30 10:17:23,294 p=208169 u=root n=ansible | Wednesday 30 December 2020  10:17:23 -0500 (0:00:04.105)       0:00:06.954 **** 
2020-12-30 10:17:23,318 p=208169 u=root n=ansible | ok: [overcloud-computehci-0] => {"ansible_facts": {"current_objectstore": "bluestore"}, "changed": false}
2020-12-30 10:17:23,359 p=208169 u=root n=ansible | TASK [warn user about osd already using bluestore] *****************************
2020-12-30 10:17:23,359 p=208169 u=root n=ansible | Wednesday 30 December 2020  10:17:23 -0500 (0:00:00.064)       0:00:07.019 **** 
2020-12-30 10:17:23,386 p=208169 u=root n=ansible | ok: [overcloud-computehci-0] => {
    "msg": "WARNING: overcloud-computehci-0 is already using bluestore. Skipping all tasks."
}
2020-12-30 10:17:23,426 p=208169 u=root n=ansible | TASK [ceph-facts : include facts.yml] ******************************************

}

[root@overcloud-controller-0 ~]# podman exec -it ceph-mon-overcloud-controller-0 sh -c "ceph -f json osd metadata" | jq -c '.[] | select(.hostname == "overcloud-computehci-0") |["host", .hostname, "osd_id", .id, "objectstore", .osd_objectstore]'
["host","overcloud-computehci-0","osd_id",2,"objectstore","filestore"]
["host","overcloud-computehci-0","osd_id",5,"objectstore","filestore"]
~~~
4. remove osd_objectstore from template hence it should be default falls to 'bluestore' but same issue.

~~~
Running /var/lib/mistral/ed1614f6-7c57-4607-9802-d716886853f6/ceph-ansible/ceph_ansible_command.sh
2020-12-30 11:03:19,584 p=285045 u=root n=ansible | Using /usr/share/ceph-ansible/ansible.cfg as config file
2020-12-30 11:03:20,288 p=285045 u=root n=ansible | PLAY [osds] ********************************************************************
2020-12-30 11:03:20,558 p=285045 u=root n=ansible | TASK [gather and delegate facts] ***********************************************
2020-12-30 11:03:20,558 p=285045 u=root n=ansible | Wednesday 30 December 2020  11:03:20 -0500 (0:00:00.327)       0:00:00.327 **** 
2020-12-30 11:03:22,248 p=285045 u=root n=ansible | ok: [overcloud-computehci-0 -> 172.16.0.23] => (item=overcloud-controller-0)
2020-12-30 11:03:24,127 p=285045 u=root n=ansible | ok: [overcloud-computehci-0 -> 172.16.0.39] => (item=overcloud-controller-1)
2020-12-30 11:03:25,956 p=285045 u=root n=ansible | ok: [overcloud-computehci-0 -> 172.16.0.32] => (item=overcloud-controller-2)
2020-12-30 11:03:26,069 p=285045 u=root n=ansible | TASK [set_fact current_objectstore] ********************************************
2020-12-30 11:03:26,070 p=285045 u=root n=ansible | Wednesday 30 December 2020  11:03:26 -0500 (0:00:05.511)       0:00:05.839 **** 
2020-12-30 11:03:26,097 p=285045 u=root n=ansible | ok: [overcloud-computehci-0] => {"ansible_facts": {"current_objectstore": "bluestore"}, "changed": false}
2020-12-30 11:03:26,142 p=285045 u=root n=ansible | TASK [warn user about osd already using bluestore] *****************************
2020-12-30 11:03:26,142 p=285045 u=root n=ansible | Wednesday 30 December 2020  11:03:26 -0500 (0:00:00.072)       0:00:05.911 **** 
2020-12-30 11:03:26,174 p=285045 u=root n=ansible | ok: [overcloud-computehci-0] => {
    "msg": "WARNING: overcloud-computehci-0 is already using bluestore. Skipping all tasks."
}
~~~



[1]https://bugzilla.redhat.com/show_bug.cgi?id=1853721
[2]https://bugzilla.redhat.com/show_bug.cgi?id=1733577
[3]
~~~
https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/16.1/html-single/framework_for_upgrades_13_to_16.1/index#migrating-OSDs-from-FileStore-to-BlueStore
~~~


Version-Release number of selected component (if applicable):

OSP16.1

How reproducible:
100%

Steps to Reproduce:
1.
2.
3.

Actual results:
migration not working

Expected results:
migration should work.

Additional info:

Comment 1 John Fulton 2021-01-04 14:43:09 UTC
(In reply to Ravi Singh from comment #0)
> Description of problem:
> 
> There are multiple BZs on this issue & some of them are fixed but the issue
> still persists which is mentioned in [1] but should be fixed as per [2].
...
> [1]https://bugzilla.redhat.com/show_bug.cgi?id=1853721
> [2]https://bugzilla.redhat.com/show_bug.cgi?id=1733577

It cannot be fixed as per bz 1733577 because bz 1733577 is still assigned and has no fixed-in. Per comment #12 of 1733577 ceph-ansible needs an update to better deal with a mixed-scenario.

*** This bug has been marked as a duplicate of bug 1733577 ***

Comment 2 John Fulton 2021-01-04 14:44:55 UTC
(In reply to John Fulton from comment #1)
> (In reply to Ravi Singh from comment #0)
> > Description of problem:
> > 
> > There are multiple BZs on this issue & some of them are fixed but the issue
> > still persists which is mentioned in [1] but should be fixed as per [2].
> ...
> > [1]https://bugzilla.redhat.com/show_bug.cgi?id=1853721
> > [2]https://bugzilla.redhat.com/show_bug.cgi?id=1733577
> 
> It cannot be fixed as per bz 1733577 because bz 1733577 is still assigned
> and has no fixed-in. Per comment #12 of 1733577 ceph-ansible needs an update
> to better deal with a mixed-scenario.
> 
> *** This bug has been marked as a duplicate of bug 1733577 ***

Apologies 1733577 != 1875777 and I meant to mark this as a duplicate of the 1875777

*** This bug has been marked as a duplicate of bug 1875777 ***


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