[ https://issues.jboss.org/browse/ISPN-8454?page=com.atlassian.jira.plugin.... ] William Burns commented on ISPN-8454: ------------------------------------- I have been able to finally generate some logs. And it seems this happens when an memory location that mapped to a given key is removed, but for some reason it isn't removed from the memory lookup. This can cause this memory location to possibly reused and stored in a different memory address, causing it to be returned when it shouldn't. Then when the key is removed from LRU it causes a crash when it frees it. The reason that JMH test exhibits this is due to the fact that the cache is retained between runs and causes the entry to eventually be removed by LRU. This gist shows where it didn't remove the entry in the memory address. Would have expected it to update the memory address to point to the new one. https://gist.github.com/wburns/8ddb3cac6b4f9f904f8e99aa8e6238da -- This message was sent by Atlassian JIRA (v7.5.0#75005)