ruby - CURSOR_NOT_FOUND - my cron jobs started dying in the middle -
a cron job running years started dying after 80% completion. not sure if because collection results steadily growing , reached critical size (does not seem big me) or other reason. not sure how debug this, found user @ whom job died , tried run job user, got cursor_notfound message after 2 hours. yesterday died after 3 hours of running users. still using old mongoid (2.0.0.beta) because of multiple dependences , lack of time change it, mongo date (i know bug in versions before 1.1.2).
i found 2 similar questions neither of them applicable. in this case, used mopped not production ready. , here problem in pagination.
i getting error message
mongodb cursor.refresh() cursor xxxxxxxxx rake aborted! query response returned cursor_not_found. either invalid cursor specified, or cursor may have timed out on server. any suggestions?
a "cursor not found" error mongodb typically indication cursor timed out (after 10 minutes of inactivity) potentially indicate client code has become confused , using stale or closed cursor or has corrupted cursor somehow. if 3 hour runtime included lot of busy time on client in between calls mongodb, might give server time timeout cursor.
you can specify no-timeout option on cursor see if server timeout of cursor causing problem.
Comments
Post a Comment