Issue 7723: slapd crashes on multi core machines if a search request is "immediately" followed by an unbind Status: VERIFIED FIXED Reported: 2013-10-10 13:27 UTC by michael.vischchers@7p-group.com Modified: 2014-10-23 07:31 UTC Product: OpenLDAP Component: slapd.2.4.36 Hardware: All Key Information: Description of the Bug: slapd crashes on multi-core machines when a search request is immediately followed by an unbind operation. Reproduction Steps: - Make sure that threads can run freely. - Start slapd with the rwm overlay. - Run a client loop that binds as a valid user, sends a valid search request that is "immediately" followed by an unbind request. Discussion Points: - The issue seems to be related to the rwm overlay and how it handles session variable substitution in certain conditions. - Several contributors provided patches and insights to address race conditions and reference count issues. - The problem was eventually fixed, as indicated by the status "VERIFIED FIXED." Attachments: : A test scenario to help reproduce the issue. Comments Summary: Comment 2: The bug persists in 2.4.36. Comment 4: Detailed steps to reproduce the issue. Comment 6: Confirmation that the issue is not a misuse of the API but a day-to-day network problem. Comment 9: Insights from Valgrind and ElectricFence tests, along with proposed fixes. Comment 13: Request for review of a proposed patch. Comment 16: Confirmation of a crash with GDB backtrace. Comment 17: Discussion on the build environment and potential issues.