Solr sometimes does not index objects as it is requested to. This is probably related to system overload, since curl error 28 is logged, and since the same objects can index just fine if you retry. eZ Find does detect when this happens. See eZSolrBase::addDocs(), which returns true or false. This is propagated through eZSolr::addObject() to ezfUpdateSearchIndexSolr::runSubProcess() in bin/php/updatesearchindexsolr.php. The same occurs in cronjobs/ezfindexcontent.php.

The current version of bin/php/updatesearchindexsolr.php does notify you about object IDs for which it could not index ("Failed indexing object with ID ..."). In 2.1.0 there is no such warning. The cronjobs/ezfindexcontent.php script does nothing about it.

Both the CLI script and the cronjob should handle this better. They should automatically retry the failed objects once or twice, and objects that keep failing they should be logged, in order to call attention to the problem.

Steps to reproduce

(Need to simulate curl error 28 in order to reproduce this)