As of 6.2 OSTree sync only pulls the latest branch, it does not sync the full history. This affects customers wanting to deploy to a specific version of the host. Steps: 1) Sat with ostree enabled 2) manifest with ostree + enable atomic repo or create product + repo with one of the upstream repos. 3) Sync Actual Notice only the latest version (7.2.6) as of now shows up. Expected All versions traversable via history,
What do we want the default depth to be? Do we need to included the additional commits (and metadata) in the summary information pulp stores in the (pulp) repository scratchpad?
I would say lets keeps the depth infinite and get full history by default. The user using this is likely going to want to preserve history and use the ability to roll back. We can always override specific repos if history is not desired. We are not using scratch pad from summary in Katello. We assume pulp does a pull all and populates its mongo db accordingly. We use :type_ids => ["ostree"] to get the branch information from there.
The Pulp upstream bug status is at ASSIGNED. Updating the external tracker on this bug.
The Pulp upstream bug priority is at High. Updating the external tracker on this bug.
Note: When testing with F24 trees at: https://kojipkgs.fedoraproject.org/atomic/24/ using a depth=-1, the sync time slightly more than doubled. Went from 7 min (depth=0) and 16 min (depth=-1). Only pulled (1) branch.
The Pulp upstream bug status is at POST. Updating the external tracker on this bug.
This commit will need to be cherry picked downstream.
The Pulp upstream bug status is at MODIFIED. Updating the external tracker on this bug.
Here is the commit: 7735b5c8ba2e5b9bfeb4d6228b5956cd3402e6c9
Katello will need to configure both the importer and distributor with depth=-1 for this to be completely resolved.
The Pulp upstream bug status is at ON_QA. Updating the external tracker on this bug.
Created redmine issue http://projects.theforeman.org/issues/17427 from this bug
Upstream bug component is Repositories
The Pulp upstream bug status is at CLOSED - CURRENTRELEASE. Updating the external tracker on this bug.
Moving this bug to POST for triage into Satellite 6 since the upstream issue http://projects.theforeman.org/issues/18076 has been resolved.
verification blocked by https://bugzilla.redhat.com/show_bug.cgi?id=1477751
Failed QA with Satellite 6.3 SNAP 13: satellite-6.3.0-16.0.beta.el7sat.noarch , tfm-rubygem-katello-3.4.4-1.el7sat.noarch - imported manifest - enabled : Red Hat Enterprise Linux Atomic Host Trees - synced : Red Hat Enterprise Linux Atomic Host Trees - after syncing, there is only 1 branch visible - attaching screenshot of the result
Created attachment 1320039 [details] Failed: Satellite 6.3: Ostree branches only shows 1
Requesting needsinfo from upstream developer jortel because the 'FailedQA' flag is set.
Could you post details of the API calls which Katello does in this scenario? To be sure that depth=-1 is configured for both importer and distributor.
Tanya is asking the appropriate question.
When initially testing this, it was not obvious from the description how depth actually works within Satellite. After some digging around, I now see the following as a configurable attribute on the repository: attribute: "Upstream Sync Policy" possible values: "Latest", "All History" and "Custom Depth". In the initial test, it was set to the default behavior, "Latest", which appears to generate the correct behavior. Only 1 branch is synced. Setting it to "Custom Depth" of 3 and sync, the result was 2 branches that appear to be identical from 'Content => OSTree Branches'. Is that expected behavior? Setting it to "All History" and sync, the result was 3 branches that appear to be identical from 'Content => OSTree Branches'. Is that expected behavior?
Hey Brad, For detailed description and verification steps look at https://github.com/Katello/katello/pull/6637#issuecomment-288629620 I suspect you ran into the """ Note that the UI metrics of a sync operation are going to look identical whether you are syncing using All History/Latest Only/Or Custom even though these mean different traversal depth. This is because pulp does not capture "all" the results of the sync in its mongo db. It only captures the "branch heads" and makes units out of them. So Katello has no way of knowing whether all versions got synced. This is being addressed in https://pulp.plan.io/issues/2594 """ We may choose to pull in the fixes for https://pulp.plan.io/issues/2594 from 2.14 to 2.13 (or create a bz for that)
In order to verify this bug, really need to have access to the pulp fix in bug 1425605. Without it, the behavior described in comment 27 will be observed.
Upstream bug assigned to paji
The Pulp upstream bug status is at NEW. Updating the external tracker on this bug.
The Pulp upstream bug priority is at Normal. Updating the external tracker on this bug.
*** Bug 1514009 has been marked as a duplicate of this bug. ***
Requesting needsinfo from upstream developer ttereshc because the 'FailedQA' flag is set.
Importer was not updated properly in the certain cases. Fixed, see Pulp issue #3210. An importer/distributor update is performed via asynchronous task, thus Katello should check if the task was finished successfully or not. As per comment 38, it seems like such check is not performed and Katello assumes that importer/distributor was updated successfully.
Hey Kedar, Looks like the pulp change didn't make it to the 6.3 GA, but should be available 6.3.1 and above.
We can now perform OSTree Sync with "all History". VERIFIED With Sat6.4.0-snap9
Created attachment 1456938 [details] OSTree sync showing all branches of RHEL7
Created attachment 1456939 [details] OSTree branches sync Coomplete
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2018:2927