HP HSG80 manual Command Scripter Communication Loss Due to Scsi Error, Association Sets

Page 11

Common Platform Issues

Intermittent Double Normalization After a Full Failback or Failback to New Hardware Procedure

After you have performed a full failback or failback to new hardware, a second normalization (from the initiator to target) can occur. This double normalization happens after the point in the failback procedure where you add the target and normalize your storage sets (from the target to initiator).

The only effect of this extra normalization is to add the additional controller overhead of doing a full copy—it causes no data corruption. When normalization is complete, you can proceed with the failback and add the load at the appropriate point in the procedure.

Command Scripter Communication Loss Due to SCSI Error

If Command Scripter v1.0A has a loss of communication while a script is running, it will cause the script to abort. You must then put the controllers in a known state by issuing the appropriate CLI commands from the controllers. You can examine the .log files in the $CLONE_HOME/log directory to determine which commands were executed before communication was lost.

If the communication failure is due to a SCSI error, then the loss of connection is only momentary. You can verify that the connection is re-established by issuing the following command from the host:

cmdscript -f <device Name> "show this"

However, if this command continuously gives a communication failure error, a hardware failure or configuration error is indicated. Troubleshoot the loss of connection for causes such as broken links, failed devices, controller configuration, and so on.

Waiting for Write History Logging Disks

After clearing invalid cache and lost data, remote copy set processing will halt. When you issue the CLI command SHOW REMOTE_COPY_SETS FULL, you may see the following error message:

Waiting for write history logging disks to become ready.

Restart the controller to clear this condition.

Association Sets

Association sets can contain up to 12 remote copy sets. However, because all remote copy sets within an association set are moved between controllers as a group, all remote copy sets within an association set must be accessed by the same server.

For instance, 6 remote copy sets (one association set) could be accessed by one server and 6 remote copy sets (another association set) by the other.

Data Replication Manager HSG80 ACS Version 8.7P Release Notes

11

Image 11
Contents Release Notes Copyright 2000-2004 Hewlett-Packard Development Company, L.P Release Notes Contents Related documentation Controller Firmware Upgrade PathWhat’s New Intended AudienceOperating System Support Matrix Operating System and Fibre Channel Switch SupportOperating OS Version Fibre Channel Switch DRM Supported Switch Firmware Swcc and OpenView Management Appliance Limitations Common Platform IssuesPerformance Considerations-Full Copy Operation Performance Considerations-Write History Log MergeRAID 5 Remote Copy Set Target Drop Pause During Fabric ReconfigurationController Saturation Startup of Only One Fabric at a Time Required Delay Time Before FailbackRemoving Targets from the Proper Controller Association Sets Command Scripter Communication Loss Due to Scsi ErrorWaiting for Write History Logging Disks HP Tru64 Unix Platform-Specific IssuesHP OpenVMS Auto Failback Booting Windows 2000 Servers Over the FabricMicrosoft Windows Increased LUN SupportChanging Host Connection Unit Offsets HSG80 Controller Soft Shutdown/RestartLUNs Lettered Incorrectly Association Sets Windows Using Large LUNs While in SCSI-2 ModeMicrosoft Windows NT NetWare 6 Storage Planning Considerations NetWare Cluster Services NwcsNovell NetWare Planned Failover/Failback with NwcsPartitioned LUNs CPU Hog AbendsMemory Requirement Sun SolarisMaximum Number of Host Bus Adapters