[Zope] Heap Corruption Problem

Apparently, we have a serious heap corruption problem. The problem manifests itself in different ways: The Python process * receives signal 9 (SIGKILL) by Solaris 2.7 * receives signal 11 (SIGSEGV) * enters an infinite loop (and blocks the complete process) In all cases, the analysis detected that the fatal event occured in a memory management function, not always the same, however. The problem occurs about 1 to 2 times per day on a lightly loaded site (still under construction and test). Environment: Zope 2.1.6 binary release for Sparc-Solaris 2.6 Solaris 2.7 ZOracleDA 2.1.0 with Oracle 8.1.6 no other C level extensions I will keep you informed. Dieter