The current SMI - Storage Server probe performs the CIM_ProtocolControllerForUnit query that can return large sets of data. In some cases, the result set can exceed 120Mb. The probe performs this query and then parses the payload to pull out what it needs, usually a very small subset of this data set. If several of these probes are executing on the MID Server at the same time, out of memory errors may occur.
Steps to Reproduce
- This problem occurs only when a single CIM server is being used for multiple large storage arrays
- This problem can only occur when multiple storage server probes are executing in parallel against a single CIM server
- The problem occurs more frequently with Hitachi storage
- Symptoms include:
- Extremely long probe times for the SMI - Storage Server probe
- MID server out of memory while executing the storage server probe
- Partial results may be returned
This problem has been researched and fixed in Jakarta. A workaround is available for Istanbul and Helsinki but applying it requires manual intervention by ServiceNow Customer Support.
Related Problem: PRB715487