idr050: modifications for OMERO 5.4 #103
Conversation
ansible/idr-centos-init.yml
Outdated
| state: present | ||
|
|
||
| - name: centos-init | Update locate db | ||
| command: updatedb |
There was a problem hiding this comment.
I hadn't added previously. @mtbc : was the behavior on idr-experimental ok for you? If so, I could drop. Otherwise, the only thing that occurs to me is to setup in cron.
There was a problem hiding this comment.
I did an updatedb manually there. Waiting for crond to run /etc/cron.daily/mlocate would be fine.
There was a problem hiding this comment.
I don't think we've set up cron anywhere else for the IDR. @manics : do you foresee any issues?
There was a problem hiding this comment.
Does updating the mlocate db have any performance impact?
There was a problem hiding this comment.
For the short window where it's running, a slight one. In general, no, I don't think locate is going to be a significant burden on any of our instances.
There was a problem hiding this comment.
mlocate's significantly better than predecessors in my limited experience.
| - name: centos-init | Install locate | ||
| become: yes | ||
| yum: | ||
| name: mlocate |
There was a problem hiding this comment.
What's the background for this? The PR ome/openmicroscopy#5684 in the description doesn't mention it.
There was a problem hiding this comment.
It's not an OMERO-specific request but an "IDR devops" request. i.e. for the likes of Mark to better find things on these servers.
There was a problem hiding this comment.
From a "devops" perspective we shouldn't really need this since it means we haven't defined/documented where everything is. Does it need to be on every server, or just the OMERO ones?
There was a problem hiding this comment.
OMERO and Database was the initial requirement.
There was a problem hiding this comment.
Can you make it specific to those then? idr-00-preinstall.yml is run by everything- publicidr, ftp, and the k8s vae.
There was a problem hiding this comment.
Added idr_debug after discussion in #idr.
Note: also dropped the cron section since @mtbc pointed out that a daily job is installed by default.
ansible/idr-omero-readonly.yml
Outdated
| config set -- omero.sessions.session_manager ome.services.sessions.InMemorySessionManagerImpl | ||
| config set -- omero.security.event_provider ome.security.basic.InMemoryEventProvider | ||
| config set -- omero.cluster.read_only true | ||
| config set -- omero.pixeldata.memoizer.dir.local /tmp/BioFormatsCache |
There was a problem hiding this comment.
Does this mean the NFS share https://github.com/IDR/deployment/blob/IDR-0.4.7-1/ansible/idr-omero-readonly.yml#L24 can either be removed or made read-only?
There was a problem hiding this comment.
It could be made read-only, but this is actually one of the places where I think we've gone too far. Specifically, I want to be able to write to the global BioFormatsCache from the read-only servers. I don't need to yet since we're not bumping our Bio-Formats version, but we will need to soon.
There was a problem hiding this comment.
@mtbc suggests in https://trello.com/c/q5L39XPZ/5-bioformatscache-is-expecting-read-write that if I omit this line, then the old behavior will be used. Likely worth testing.
There was a problem hiding this comment.
Pushed a removal of omero.pixeldata.memoizer.dir.local
|
Relaunched travis with gh-109 |
This reverts commit 1dcfc48.
| yum: | ||
| name: mlocate | ||
| state: present | ||
| when: idr_debug | default(false) | bool |
There was a problem hiding this comment.
Optional: If it's going to make things easier in future set a default in https://github.com/IDR/deployment/blob/IDR-0.4.7-1/ansible/group_vars/all.yml
There was a problem hiding this comment.
I thought about that, but since I haven't rolled it out elsewhere, I think I'll leave this as "internal API" for the moment. But agreed, having a section in all.yml with "expert" variables will be useful.
There was a problem hiding this comment.
When we move to Ansible 2.5 we can use the new never tag instead http://docs.ansible.com/ansible/latest/user_guide/playbooks_tags.html
| openstack_volume_device: /dev/vdb | ||
| openstack_volume_source: "{{ idr_volume_dockermanager_data_src }}" | ||
|
|
||
|
|
There was a problem hiding this comment.
This is fine for now, but we'll need to clean up all references to these servers in idr-proxy at some point.
There was a problem hiding this comment.
Agreed. I couldn't tell where I'd conflict with your work, so tried to go for the simplest thing possible.
|
From my point-of-view as soon as all 0.4.8 is merged, this can be merged. |
|
👍 What are the remaining steps to getting this merged and deployed on |
|
Happy to run on prod50 today from a merged stated if there are no objections. |
see: ome/openmicroscopy#5684