Removing and Inserting Enclosure Management
Modules on Attached Expansion Enclosures
The following procedures describe how to safely remove and insert an enclosure management module (EMM) from an expansion enclosure attached to the MD3000/MD3000i.
NOTICE: Failure to follow these guidelines may result in a physical disk failing during removal and/or inadvertent removal of its redundant data path.
Removing an EMM from the Expansion Enclosure
Check the Recovery Guru to confirm that there is no loss of physical disk
path redundancy.
If there is no loss of redundancy, check the channel. If it matches the EMM you are removing, you can safely remove it now.
If redundancy is lost, run the following command. Note that channel [1] is the degraded channel.
"set physicalDiskChannel channel [1]
status = optimal;"
Verify that the path to the channel is restored. The Recovery Guru may
take several minutes to update.
Inserting an EMM into an Expansion Enclosure
Make sure to always insert the EMM without SAS cables attached.
After the EMM is inserted, wait at least 30 seconds before attaching the SAS cables.
After attaching the SAS cables, wait at least three minutes for the EMM to
reach optimal state.
NOTE: You may experience a transitional failure on multiple EMMs in the MEL log while the system updates.
Removing and Inserting Physical Disks
Refer to the following guidelines to ensure that physical disks are safely removed from and inserted into the MD3000/MD3000i RAID storage array.
Wait at least 60 seconds between removing a drive and inserting a replacement.
When pulling a drive from a storage array to move it to a different slot, wait 60 seconds before inserting the drive into the new slot.
Wait at least 60 seconds between the removal of drives from a storage array.
Wait at least 60 seconds between the insertion of drives into a storage array.
In a large configuration, storage management software may take up to 10 seconds to detect hardware changes.
MD3000 Maintenance Considerations
For Linux kernels, stop and then restart the SMagent after performing one of the following maintenance tasks:
Move a controller offline or replace a controller.
Remove SAS cables from or attach SAS cables to host servers running Red Hat® Enterprise Linux® (version 4), SUSE® Linux Enterprise Server 9, or SUSE Linux Enterprise Server 10 operating systems.
NOTE: If a Resolve Topology Conflicts message appears in the Status portlet of the Summary page after restarting the SMagent, it may be necessary to restart the host server to clear this message. Do not select the Resolve selection in the Topology Conflict wizard. The MD3000 will continue to service I/O requests, but the array could become partially managed if only in-band management is used. Stop all I/O operations on any host servers that were involved in the maintenance operation and restart the system.
MD3000 Cluster Configuration Guidelines for
Standalone Host Servers
If one of the standalone host servers you are planning to configure into a cluster environment is running MD Storage Manager and has a virtual disk mapped to the array, use that host to create the Host Group and quorum virtual disk mapping.
NOTE: Failure to follow this mapping protocol can cause the host server to lose communication with the array if the server is using only in-band management and only one of the servers has a virtual disk mapped to the array. If communication is lost, restore in-band management to the other host server and complete the cluster setup using the following procedure.
Using MD Storage Manager from either the host server with restored in-
band management or from a management station, select Configure ®
Create Host-to-Virtual Disk Mappings.
Select the host server name that does not have a virtual disk mapped.
