Virtual TapeServer 8.2 Configuration Guide
Table Of Contents
- Virtual TapeServer for NonStop Servers Configuration Guide
- Preface
- Introduction
- Overview of Tasks
- Reconfiguring Vaults
- Enabling Licensed Features
- Configuring Ports
- Creating and Managing VTLs and VTDs
- Enabling and Performing Tape-to-tape Exports
- Enabling and Performing Stacked Exports
- Enabling and Configuring Data Replication
- Enabling and Configuring Role Swapping
- Configuring EMS Communication
- Enabling and Configuring Data Encryption
- Creating and Managing Virtual Media
- Enabling and Configuring Scan/Cleanup
- Configuring User Accounts
- Configuring Web Interface Preferences
- Managing the VTS Server
- Troubleshooting
- Maintaining GFS for VTS
- Reinstalling and Restoring VTS
- Attaching External Devices after Initial Deployment
- TCP/IP Ports and Protocols
- Index
157
Maintaining GFS for VTS
The Global File System (GFS) is an advanced feature that allows Linux servers to
simultaneously read and write files on a single shared file system on a SAN. VTS is based on
Linux, and GFS enables multiple VTS servers to access a shared set of pools and virtual
tapes. The Event Management Service (EMS) can then automatically mount virtual tapes
from the GFS pools as if they were separately mounted.
With GFS, all VTS servers connect to each other over Fibre Channel to the same storage
array, and GFS allows each VTS to access all vaults at the same time. If a server becomes
unavailable, other servers can still access the vaults.
This appendix provides common procedures for maintaining GFS using the command line.
You can also use the cluster administration user interface to manage GFS. For complete
information about this user interface, refer to
http://www.redhat.com/docs/en-US/
Red_Hat_Enterprise_Linux/5.2/html/Cluster_Administration. Refer to this URL for
information about GFS and how to install a broader range of tools: http://www.redhat.com/
docs/en-US/Red_Hat_Enterprise_Linux/5.2/html/Global_File_System/index.html.
Powering hardware and starting a cluster
If GFS is properly installed and configured, the VTS startup process automatically starts
GFS.
Note The following procedure assumes that the GFS cluster has been configured.
To power all hardware and start GFS on all nodes in a cluster
1. Power on VTS servers:
a. Power on first VTS server. Wait for this server to boot completely. Fencing may power
on subsequent VTS servers.
b. If fencing does not power on subsequent servers, power on the second VTS server.
Wait for the server to boot completely. Power on each server in this way.
c. Log in to the first VTS server’s operating system and become root:
su -
d. On the first VTS server, enter these commands:
service clvmd start
service cman start
service luci start
service ricci start
Note that luci runs on only one VTS server in the cluster.