Bug 1477260

The section looks good Bara

Hi Bara,

In Section 1.2, Step 5.ii,

1) All mentions of ceph_osd_docker_extra_env has to be removed as it is no longer needed.

2) value of ceph_osd_docker_prepare_env should be "-e OSD_JOURNAL_SIZE={{ journal_size }}" So it can be added in example usage section as,

ceph_osd_docker_prepare_env: -e OSD_JOURNAL_SIZE={{ journal_size }}

3) As example description says 'For example, if you want to use encrypted OSDs with a separated journal, use', I think we need to provide steps/instructions to enable the variable 'dmcrypt', 'devices', 'dedicated_devices', and 'osd_scenario'.

In this case,
dmcrypt: true
osd_scenario: non-collocated

I think we can suggest users to select for appropriate as per their environment devices and dedicated_devices.
I think the word 'separated' in example description can be changed to 'dedicated' to maintain consistency between variable name and description.

REF - https://bugzilla.redhat.com/show_bug.cgi?id=1475820#c12

----------------------------------------------------------------

In Section 1.3, Step 3 needs to be removed as the variable 'rgw_containerized_deployment' has been removed from rgws.yml


Moving back to ASSIGNED, please let me know if there are any concerns.


Regards,
Vasishta

Bara, after looking at https://doc-stage.usersys.redhat.com/documentation/en-us/red_hat_ceph_storage/3/html-single/container_guide/#installing-a-red-hat-ceph-storage-cluster-in-containers

You can remove some variables:

* 'docker' has never been needed

The 3.3. Purging Clusters Deployed by Ansible section

You have to copy the file at the root of the repo. So:

[user@admin ceph-ansible]$ cp infrastructure-playbooks/purge-docker-cluster.yml . 

The rest looks good.

Hi Bara,

Sorry for the delayed update.

Here are some points that I think doc needs to be having.

1) In the list of daemons supported to be colocated with OSD, NFS is missing, can you please add it ?

2) In procedure section of 2.2, As we are talking about colocated daemons, it would be better if we mention like - 
By default, the value of ceph_<daemon_type>_docker_memory_limit and ceph_<daemon_type>_docker_cpu_limit will be 1 and the colocated daemon will be configured with default value if its not changed.

3) Can you please add an example for "How Colocation Works" ? 
Though the description for "How Colocation Works" looks good, It would be better if we can add an example of "Ansible inventory file" following the convention as in step 6 of Section 1.2 -

like -

[mons]
<monitor-host-name>

[mgrs]
<monitor-host-name>

[osds]
<monitor-host-name>
<osd-host-name-1>
<osd-host-name-2>

[rgws]
<osd-host-name-1> #RGW will be collocated with OSD 
<gateway>

Can we expect this bug come ON_QA once we get the diagram ? (Just to make sure that QE doesn't miss out checking the diagram) ?

Regards,
Vasishta

Hi Bara,

Yes, Looks good to me, Moving to VERIFIED state.

Thanks,
Vasishta