bryancall commented on issue #7192: URL: https://github.com/apache/trafficserver/issues/7192#issuecomment-4827339459
I traced this to host status being stored as persistent records, which is what gets written into records.snap. In the affected versions, when a parent or host was marked down, the state was registered as a persistent statistic, so it survived a restart hidden inside records.snap and could only be cleared by deleting that file (exactly the behavior reported here). This was reworked in https://github.com/apache/trafficserver/pull/8689, which removed host status from the persistent stats and moved it to a dedicated host_records.yaml file that is queryable and clearable with traffic_ctl host status. That change first shipped in 10.0.0. On current master, host status loads from host_records.yaml rather than records.snap, and transient parent-down state is held in memory only and recovers after the retry window, so the stuck-503 condition described here no longer applies. Closing as fixed by the rework. Please reopen if you can still reproduce stuck 503s tied to records.snap on a curren t release. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
