Deploying and scaling swift requires some manual steps. This is outlined in this upstream patch. https://review.openstack.org/#/c/293311/ We need to document this process for OSP8. Let me know what other details you need.
Updated release flags and priority.
@dmatthew, just for clarification, what's the difference between the ring that the director creates vs the one that you manually create?
@dmatthew, also does this only affect OSP8, or does it affect previous OSP7 Overclouds?
Don - Yes, this is a workaround for a bug in director. (In reply to Dan Macpherson from comment #6) > @dmatthew, just for clarification, what's the difference between the ring > that the director creates vs the one that you manually create? I don't have enough swift knowledge to answer this in as much detail as you might need. However, the primary issue is that new Swift nodes when scaling up will not be added to the existing Swift Ring. It seems that new Rings may be created and the state of Swift then becomes unclear. I think there may also be a problem with the defaults of the ring that director creates. If that doesn't help, I would recommend needinfo'ing cschwede. He has a good understanding of this issue. (In reply to Dan Macpherson from comment #7) > @dmatthew, also does this only affect OSP8, or does it affect previous OSP7 > Overclouds? I believe this will also affect OSP7. This is a limitation that has always existed in upstream TripleO since initial Swift Ring support was added and still does.
This bug was accidentally moved from POST to MODIFIED via an error in automation, please see mmccune with any questions
*** Bug 1322614 has been marked as a duplicate of this bug. ***
Assigning Radek as the QA contact. Radek - can you take a look at the changes for this bug?
Looks great.
This content is live on the Customer Portal. Closing.