rgw: extend automatic rgw pool creation capability #4445
Conversation
alimaredia
changed the title
|
@alimaredia when you have a moment 🙂 |
alimaredia
force-pushed
the
wip-rgw-multisite-create-pools
branch
6 times, most recently
from
6f435c9
to
0663a38
Compare
alimaredia
force-pushed
the
wip-rgw-multisite-create-pools
branch
from
0663a38
to
20f2b9e
Compare
|
@guits @dsavineau: This PR should be ready for final review. I made some changes in the |
alimaredia
changed the title
| when: | ||
| - item.value.pool_type is defined | ||
| - item.value.pool_type == 'ec' | ||
| ignore_errors: yes |
There was a problem hiding this comment.
The reason we had this was if the playbook gets run again. When the playbook gets run again, it will try to rm the ec_profile and we didn't want it to fail if the ec_profile was already created and being used.
There was a problem hiding this comment.
@alimaredia If we really want to stick to ignore errors, failed_when: false should be enough
alimaredia
force-pushed
the
wip-rgw-multisite-create-pools
branch
from
20f2b9e
to
3f02b2a
Compare
| - item.value.pool_type == 'ec' | ||
|
|
||
| - name: set crush rule | ||
| command: "{{ container_exec_cmd }} ceph --connect-timeout 5 --cluster {{ cluster }} osd crush rule create-erasure {{ item.key }} {{ item.value.ec_profile }}" |
There was a problem hiding this comment.
what about this command?
If the playbook is run again, will it return 0 even though it's already created? Otherwise, we might need a failed_when: false here as well
Signed-off-by: Ali Maredia <amaredia@redhat.com>
alimaredia
force-pushed
the
wip-rgw-multisite-create-pools
branch
from
3f02b2a
to
325b2a9
Compare
|
Closing in flavor of #4688 |
Signed-off-by: Ali Maredia amaredia@redhat.com