Enterprise Data Collection

Customer Information:

Use the procedures most relevant to your environment:

If you are asked for a Mode31 dump for deep trace analysis or performance issues please review these instructions: Mode 31 Dump FAQ

Service Personnel Information:

The RAID "Auto Dump" Facility is required for most problems that relate to Hitachi RAID subsystems.

This function is a procedure that merges the separate DUMP and FDCOPY functions into one operation, and provides a single compressed file available for transfer from the SVP. The default location of this file is:

• C:\DKC200\TMP\hdcp.tgz (for USP-V, USP-VM, VSP, VSP G1000, HUS-VM)

For instructions on how to perform an Auto Dump on a RAID subsystem, refer to the SVP section --> 2. Function of the SVP of the Maintenance Manual.

Note: A file called dump.tgz is NOT the correct file. Always collect file hdcp.tgz which is usually in the C:\DKC200\tmp folder.  Do not zip or archive the dump file for upload it is already compressed.  You can rename to a meaningful and descriptive title but must have "hdcp" in the name for our automated processing to work correctly.

Note: For Storage Navigator issues, click here

Types of Auto Dump:

1. LIMITED
   Used specifically for initial investigation into 5000 series arrays. This should be requested for hardware specific errors as it contains all hardware incidents and information and is very quick to collect and process. It is suggested to collect a normal or detailed dump as soon as the limited dump has been completed.
2. NORMAL 
   Useful for many hardware faults but should only be collected if requested by service personnel or if time is of the essence and the array is not a 5000 series. Performs and collects dumps (all microprocessors, ABEND, WCHK1), in addition to files gathered by RAPID. Monitor data is not collected. This type should be collected by default when no instruction has been given and the problem types referred to in "DETAIL" below do not apply.
3. DETAIL 
   This is the default dump type as it gathers all possible information needed. It Gathers SVP monitor data on the MPs and CACHE, in addition to information gathered by RAPID and NORMAL dumps. This type should be collected when there is some suspicion that the problem is related to performance, subsystem workload, sidefile or HUR journal overflow. It is also important to collect this type of Dump for CHIPs and when attempting to determine a good time frame for online code upgrades.
4. MODE 31 
   Helps with Performance Cases. Only useful if executed during the point in time a problem is occurring as this type of dump collects around 1-second of IO from each port. Make sure to turn SOM31=ON five minutes before starting a DETAIL dump. Turn SOM31=OFF after you have collected your dump (unless instructed otherwise).  Note: SOM31 is a public System Option Mode and can only be set via SVP or MPC maintenance panels by the Service Engineer using the "MODE" mode.  If using the local Hitachi Remote Operations Console the MODE31 dump can be requested and sent directly to Hitachi in one operation.
5. RAPID 
   Gathers minimal information and should only be collected if requested by service personnel.  Only includes logs, SVP operation history, configuration information. NOT collected: Dump data (Processor, ABEND, WCHK1) and monitor data. This type is typically used for Backend (Hard Disk) related errors and for system health checks.
6. Archive PIN 
   For cases reporting PIN TRACK, you must ARCHIVE PIN (See Maintenance Manual SVP section for details) via the Maintenance panel BEFORE collecting the dump or else the PIN DATA will not show up.

Before Performing the SVP Dump Procedure:

1. Perform LAN test from Diagnosis panel 
   This will check to make sure all MPs are communicating properly and the results of the test are updated in the dump data.
2. Check for Pinned Data from the Maintenance Panel 
   If "Pin..." button is not grayed out, click it and archive the pinned LDEV.

   Note: For VSP onwards, using Recommended GA code or higher, pinned data is automatically archived during Dump collection.

3. Open Version Information from the Maintenance panel 
   This refreshes the microcode data on the SVP and makes sure the latest information gets into the dump.
4. In the case of Mainframe subsystem 
   Display Mainframe Paths

If any of these steps are missed, the latest information will not be collected and can delay, and increase the difficulty of troubleshooting.