Hide Forgot
Created attachment 1209620 [details] Task content host destroy Description of problem: Bulk deleting content hosts can trigger a deadlock in postgres. See attached task export that it deadlocks on the import_data to update the katello_pools. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. Register 100 Content Hosts 2. Bulk Delete 100 Content Hosts 3. Actual results: Deadlock Expected results: Success Additional info: In the latest upstream the https://github.com/Katello/katello/pull/6187 removes already the finalize() code block that does the import_data() call.
Peter, thank you for filing the bugzilla. Justin, will the PR that Peter has referenced resolve the deadlock observed?
Indeed it does. We likely could backport *just* the piece that would fix it if we want to pull this into 6.2.z
http://projects.theforeman.org/issues/12556
This is not just bulk deleting also with re-registering hosts the deadlock is triggered. E.g. re-register (susbcription-manager register --force) on 10 hosts at the same time.
Please add verifications steps for this bug to help QE verify
*** This bug has been marked as a duplicate of bug 1419985 ***
closed as a duplicate as https://bugzilla.redhat.com/show_bug.cgi?id=1419985 already has the single fix we need and not the GIANT upstream change that happened to include it.