Pat Thanks for the reply. What if we flag the documents in the core as deleted after the delta indexing?
And also I came across one more scenario: During full indexer run, if an object gets updated - DeltaJob and FlagAsDeletedJob get enqueued. Full indexer resets the 'delta' attribute of the object to false - that the DeltaJob ignores the updated object and FlagAsDeletedJob flags the corresponding document as deleted in the core index. Now the document remains unavailable in both core and delta indices till the next full indexing. Please note that we have a single queue in the DJ for processing sphinx jobs. Thanks Sundar Rajamanickam On Monday, June 29, 2015 at 5:01:13 PM UTC-7, Sundar Rajamanickam wrote: > > Hi > > I have a hourly (cron based) full indexer and delayed delta indexer setup. > > Say, when I update an object of 'Article' class - two delayed jobs get > enqueued - DeltaIndexJob and FlagAsDeletedJob. > > If only the FlagAsDeletedJob gets executed, and DeltaIndexJob does not get > executed. Now when the full indexer starts, it cleans up all the existing > jobs in the DJ queue - means the updated 'Article' object will not be > available in both the core and delta indices till the full indexing > completes. > > Our full indexer runs for about 12-15 minutes everytime because of the > size of the database. This results in a significant delay, that the object > becomes available for search. > > Ensuring that the FlagAsDeletedJob runs after the DeltaIndexJob will help > reduce the impact, right? > > Thanks > Sundar Rajamanickam > -- You received this message because you are subscribed to the Google Groups "Thinking Sphinx" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/thinking-sphinx. For more options, visit https://groups.google.com/d/optout.
