nickva commented on pull request #3553: URL: https://github.com/apache/couchdb/pull/3553#issuecomment-836996618
@bessbd the error comes from trying to acquire a map context. Would this usually be considered a user error (bad javascript, values are too large, etc) or server-side error, as in we something is broken with the backend (operator misconfigured the language server backend). If the error is from a misconfigured query language server, wonder if failing the job after 3 tries would make sense. If the backend is fixed, would the user still get a failed error, if so it might be better to perhaps keep retrying longer. How retries happens is a bit subtle too it seems... Notice when we retry we simply abandon the job. We exit the but still keep ownership of the job https://github.com/apache/couchdb/pull/3553/files#diff-228d2a0ae65ae417bf4b3fb19f0de9f886c2ae168ab9fbec8e5f86b93934e6f7R148. It would eventually timeout based on https://github.com/apache/couchdb/blob/c75a9b66965939e470ed5f3f2749626d3f35d311/src/couch_views/src/couch_views_jobs.erl#L33 and rely on the job activity monitor re-enqueuing the jobs. That's not the worst and it would provide a backoff of sorts but it certainly subtle. -- 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. For queries about this service, please contact Infrastructure at: [email protected]
