Last Modified: Oct 16, 2023
Affected Product(s):
BIG-IP LTM
Known Affected Versions:
11.1.0, 11.2.0, 11.2.1, 11.3.0, 11.4.0, 11.4.1
Fixed In:
11.5.0, 11.4.1 HF2, 11.4.0 HF6, 11.3.0 HF9, 11.2.1 HF11
Opened: May 24, 2013 Severity: 2-Critical Related Article:
K15715
In certain cases, mcpd can crash during configuration synchronization. Objects were being prematurely deleted and then referenced. MCPD closes this connection as shown by the message: May 21 23:22:53 USLAS-FLTM3900LB-COR-02 notice mcpd[6022]: 0107072a:5: Disconnecting user svc.scmger
The end result is that a BIG-IP system that was active will go to standby when MCPD crashes. This is because the object of MCPConnection class is being deleted. Such deletion causes a chain of events - request_group is deleted. but file_sync_unit is associated with the request_group and it also registers itself to the ISC event system. when the request_group and the file_sync_unit are deleted (the core file displays these as 0xdeadbeef) the ISC event system is not aware of that. the next ISC event will trigger a memory access that is invalid.
This often occurs after an upgrade.
Sometimes initiating the sync from a local (e.g. root or admin) user can work around this issue, but this has not been shown to be true in every case.
MCPD will no longer crash during configuration synchronization. Steps were taken to ensure that objects were not prematurely deleted.