I am finding binary items in the search queue although we don’t allow search on fields with sys_File control. Then if covered binary has big size, I get java “out of memory” errors on startup.
Last week I had a TAR open on this and seemed to resolve it. First I checked binaries are not covered by search, then stopped the server, flushed the indexes, truncated the table and re-indexed content. Thinking was probably some items got stuck in the cache for a long time. It seemed fine in the begining. However today I caught another one in the queue and got the error.
Does Convera index the database top down or somehow crawls the nav tree? Sounds funny, but is it possible that Convera actually follows the CMS links from body fields? ( A wild guess, but there is a pattern.)
Restarted the service this morning and got the out-of-memory error again. Then I deleted offending record from the queue table, started again, this time no error.