1)ERROR message
06-17-2019 22:48:08.445 -0700 ERROR CacheManagerHandler - ReverseIndex cannot add cacheId="bid|ceg_notification_dispatcher_nsprod~560~74F09ED5-8892-471C-9E6C-CD3BC6EDAC9D|" to search sid="remote_searchhead01-3.prd.cmn.mntspl.us-west-2_subsearch_tmp_1560836883.1" as this search has already completed and closed all buckets
2)Error message
06-17-2019 22:13:08.380 -0700 WARN CacheManager - Attempting to decrement refcount of cacheId="bid|creditscore_prod_pod8~66~8D530702-839D-4679-BBAA-0E408F8EDAFD|" but sid=remote_searchhead01-1.prd.cmn.mntspl.us-west-2_tpargain_tpargainsearch_search34_1560834740.1903850_B8E7D0B9-BB6E-41DB-928B-411ABBA5AD90 is not an opener
1)ERROR message
06-17-2019 22:48:08.445 -0700 ERROR CacheManagerHandler - ReverseIndex cannot add cacheId="bid|ceg_notification_dispatcher_nsprod~560~74F09ED5-8892-471C-9E6C-CD3BC6EDAC9D|" to search sid="remote_searchhead01-3.prd.cmn.mntspl.us-west-2_subsearch_tmp_1560836883.1" as this search has already completed and closed all buckets
S2 has some protection against lagging REST requests. There are times when a search process requests to open a bucket but then the search is stopped and cleaned up before the open request is handled. This used to cause us problems leaving buckets stuck with an invalid search reference which would prevent eviction (see "stale buckets" in jira). We have since noticed that some apps and sometimes search/subsearch will re-use the same SID causing these errors. Since we've recently changed the behavior for stale buckets we're in the process of relaxing this bit so we will let these requests through and avoid the error (for quake and then to be backported)
2)Error message
06-17-2019 22:13:08.380 -0700 WARN CacheManager - Attempting to decrement refcount of cacheId="bid|creditscore_prod_pod8~66~8D530702-839D-4679-BBAA-0E408F8EDAFD|" but sid=remote_searchhead01-1.prd.cmn.mntspl.us-west-2_tpargain_tpargainsearch_search34_1560834740.1903850_B8E7D0B9-BB6E-41DB-928B-411ABBA5AD90 is not an opener
The above error means a search tried to close a bucket that it didn't have opened. Any particular context you're looking for?