Everything is Linux

Wednesday, December 17, 2008


The Intelligent Platform Management Interface (IPMI) is an industry standard that defines interfaces to the TT platform management subsystem. It include remotely accessing the platform through standard messaging channels such as IPMI Over Serial, LAN, or Serial Over LAN (SOL) connections,as well as generating and sending a Platform Event Trap (PET) when a monitored event condition occurs.IPMI configuration can be carried out by different software programs and utilities both graphical user interface (GUI) and command-line interface (CLI) options available. Configuration can be divided into the following categories:

• Configuration of a LAN connection for IPMI messaging and alerting: Includes enabling IPMI communication over a LAN connection, selecting the source for the remote-access IP address (statically assigned or obtained using the Dynamic Host Configuration Protocol), assigning a static IP address, choosing the highest privilege level allowed on the connection, and configuring virtual LAN (VLAN) settings and the baseboard management controller encryption key.

• Configuration of a serial connection for IPMI messaging and alerting: Includes setting the connection mode, baud rate, flow control, terminal mode, and highest privilege level allowed on the connection

• Configuration of a SOL connection: Includes enabling SOL and configuring the baud rate and settings for how characters should be accumulated and sent

• Configuration of IPMI remote-access users: Includes enabling users; assigning usernames, passwords, and privileges on each connection channel (LAN and serial); and enabling SOL payloads on the LAN connection

• Configuration of Platform Event Filters (PEFs): Includes enabling PEF alerting on the LAN connection, selecting actions to take, and configuring alert destinations

• Restoration of IPMI configuration parameters: Returns these parameters to the reconfigured default values


Monday, December 1, 2008

Introduction to Dump

Dump: Smart Archiver for Filesystems, not FilesThere are other popular tools to manage backup archives such as cpio, tar and afio.
These utilities handle files as the target to be archived and they are capable of excluding specific files and/or directories from the target.
They can even create a single archive that contains files from multiple filesystems.In contrast, dump handles a physical filesystem as an archiving target and the restore command usually uses the archive to restore the filesystem as it was dump’ed.
Each file is managed by the i-node number and basically, dump cannot exclude specific files from the archive (actually, you may do so in a different manner.

Dump is indeed a simplistic and primitive tool, but it does come with a brilliant feature for incremental archiving. It identifies newly created or modified files after the previous backup and efficiently stores them to an archive very fast.

For example, suppose a file ‘foo’ was backed up during the last archiving and removed afterwards. On the next incremental archiving, dump puts the record in the archive as ‘there used to be a file foo, but it was removed’.
If you use tar for your regular incremental backup tasks and attempt a full restoration one day, you may run out of the disk space by trying to restore a large number of files that had already been removed.
With dump, you will never face such a problem.In summary, it would be fair to say² Cpio, tar or afio is suitable for archiving specific files or directories.² Dump is suitable for archiving whole filesystems.Just pick a right tool for you job.


Dump and Restore

The dump(8) and restore(8) commands have traditionally been used on the BSD systems to backup and restore filesystems. Dump backups a filesystem as a whole into an “archive”, and restore retrieves files from it. Although the archive may be created as a regular file on a regular filesystem, it is usually stored on an external backup device such as a magnetic tape. Some features are implemented in dump to support such devices.