Jira (PDB-2477) Refactor GC processes to avoid cascades

[Ryan Senior (JIRA)]

Title: Message Title

Ryan Senior created an issue   PuppetDB / PDB-2477 Refactor GC processes to avoid cascades Issue Type: Improvement Assignee: Unassigned Created: 2016/02/29 7:01 AM Labels: tcse Priority: Normal Reporter: Ryan Senior Current GC processes are somewhat simple and rely on cascading from a single delete in certname/report etc. Although the cascaded calls are pretty simple, there can be enough of them to be taxing on the database. Nick Walker has a new approach that is faster on larger datasets. Below is an example of the approach applied to reports SELECT id INTO TEMP TABLE report_ids_past_report_ttl from reports r where r.producer_timestamp < NOW() - INTERVAL '14 days'; ANALYZE report_ids_past_report_ttl; DELETE from resource_events where report_id IN ( select id from report_ids_past_report_ttl ); UPDATE certnames SET latest_report_id = NULL where latest_report_id IN ( select id from report_ids_past_report_ttl ); DELETE FROM reports where id IN ( select id from report_ids_past_report_ttl ); DROP TABLE report_ids_past_report_ttl; We should be able to recast node_purge_ttl GC in a similar way. It's worth re-evaluating facts as well, though it's probably different. Add Comment   This message was sent by Atlassian JIRA (v6.4.12#64027-sha1:e3691cc) -- You received this message because you are subscribed to the Google Groups "Puppet Bugs" group. To unsubscribe from this group and stop receiving emails from it, send an email to puppet-bugs+unsubscr...@googlegroups.com. To post to this group, send email to puppet-bugs@googlegroups.com. Visit this group at https://groups.google.com/group/puppet-bugs. For more options, visit https://groups.google.com/d/optout.