Guide for Building a Disaster Data Recovery Plan Server System

Here is tutorial in building or setting up Disaster Recovery Planning. This tutorial is an implementation of the authors in performing daily tasks as IT System Administrator who is obliged to maintain the stability Server System so that we can get ZERO DOWNTIME Server System.

Disaster Recovery Plan is the protection ServerSystem stages that must be prepared for in case of disaster, the data recovery process can be done quickly. DRP is very closely related with ZERO DOWNTIME Server System where the primary purpose of the Server System or DATA CENTER provides non-stop service to the Client and the Customer.

If the Server System or DATA CENTER is not designed properly then the server will experience DOWN or even damaged and takes many minutes even hours. And can we imagine, for example tens up to hundreds of bank customers would queue up in front of an ATM machine or even customers become not believe in banks and withdraw his savings.

Stages of Disaster Data Recovery Plan are as follows:

1. Redundant or Dual Input POWER SOURCE
Prepare the power source of adequate and ready to use, and can also apply on DUAL Input power to the UPS. If no Genset, or may utilize other input resources such as Solar energy, etc..
a. Input Power of Electric Power.
b. Input Power of generator sets.
c. Input Power Source Power from another.

2. DUAL UPS, or Redundant UPS, to PSU
a. UPS A
b. UPS B
We use 2 UPS with Input Power Source is different to supply a server with dual Power Supply Unit. This certainly applies to servers that have 2 pieces Power Supply Unit (PSU). The goal is if there is a problem in one of the Power Source the Server is also still able to live from another Power Supply, or other Power Source.

3. DUAL POWER SUPPLY UNIT (per server)
a. PSUA with input power from the UPS A
b. PSUB with input power from the UPS B
Not all servers have dual power supply facilities, so if our servers have dual power supply then we should use as optimal as possible.

4. LOCAL STORAGE RAID System for OS
System RAID (Redundant Array of Inexpensive Disks) is a group of hard disk that serves mutual replace / redundant to maintain a functional hard drive.

The goal is if one or more hard disk from a disk group is destroyed, then the disk group is a function of not having problems so we are not to experience a loss of data. In a RAID system is recommended using hotplug hard drive, or disk hotswap, so with this hard drive we do not need to turn off the server to process the replacement of the damaged hard drive.

RAID system that we can use are:
a. RAID 1 +0 / Mirror (minimum), better use RAID5 or RAID6.
b. RAID5 => (N = N-1), 1 piece of hard disk allocated for Fault Tolerance.
c. RAID6 / RAID ADG (Advanced Data Guard) => (N = N-2), 2 pieces of hard disk allocated for Fault Tolerance

5. DUAL / Redundant Connection in every server
Using two or more LAN Card will certainly guarantee Availability servers in the network if there is damage to the LAN Card Server. So that if one LAN connection drop out, then another LAN connection can take over the connection or the automatic Take Over.

Redundant Connections can be:
a. NIC / LAN Card for Redundant Connection & Load Balancing
b. FO for Redundant Connection (Server to the SAN / NAS & FO between theSwitch)

6. EXTERNAL STORAGE Redundant Protection for OS, database and fileserver
External Storage a SAN (Storage Area Network) or NAS (Network Attach Storage) nowadays has become a staple in the Server System. Redundant External Connection from Server to Storage is very important because it is very helpful in improving protection of storage according to the functional Data Storage to function properly.

Connection from Server to the External Storage of Fiber Optic (FOR) or Ethernet Connection (iSCSI) storage and protection to us, in the form:
a. System RAID (RAID 1 +0, RAID5 or RAID6)
b. ASM & OMF (for Oracle Database)

7. TAPE BACKUP, Tape Library, or Virtual Tape Library (VTL)
Tape Backup, Data Protection is to further both the External Catriedge and Virtual Tape Library Tape Catriedge hereinafter be saved into a special place for the event of disaster can be used for data recovery quickly.

a. Tape Backup with Catriedge adequate conventional
b. FO for Redundant Connection (from Server to the Library or VTL / Virtual Tape library)

8. Replication SERVER TECHNOLOGY
Server Technology implemented in our very important role, for example on a Single Server in case of minor problems such as Restart Server, Patch Updates, etc need time to Downtime 5 minutes until 15 minutes to process Running Up Server. Moreover, fatal problems then it took about 1 hour more, to Re-Building the same server as before. So with the Replication Server Technology, Server Downtime can be reduced up to minimize even ZERO Downtime.

There are 2 kinds of techniques in the Replication Server, namely:
a. Mirrored SERVER
b. Clustered SERVER

In the mirrored server is needed IT Administrator interventions to perform switching or takeover Servers including running the script for Passive server can take over the current Active Server Down. While on Clustered Servers are no longer required intervention because of Clustered Server IT administrators can perform automatic takeover. On the server writer, Clustered Server takeover process from NODE1 to NODE2 in Windows Server 2003 only runs in a matter of about 5 seconds.

Clustered System is a technique combines the ability or power of several servers into a Powerful Server System. The phisical, Clustered Server consists of 2 pieces Server or more even up to hundreds of servers. But the system is recognized as a fruit Server System. So Clustered Server is a manifestation or a miniature rather than mainframe servers that cost is very expensive, so with a Clustered Server Mainframe servers purchase cost can be replaced by building a Clustered Server.

9. SERVER CO-LOCATION
Server Colocation is a production server we use day-to-day operations at the replication on our servers which are beyond our Site Server. For example, outside the city, outside the island and even abroad. This implementation is highly dependent on the speed bandwidth connections that we have or we lease from the ISP.

There are 3 kinds of techniques in the Server Co-Location, that is:
a. Mirrored Server Co-Location
b. Clustered Server Co-Location
c. BACKUP Storage to Co-Location Storage (NAS with iSCSI generally use)

There are a few more things to note in the Additional Protection in building a data center. That is:
A. Cooling System is suitable and adequate (AC Precision & redundant).
B. Fire Protection & Fire Alarm (Fire Suppresson, Gas FM-200, etc.).
C. Security Access procedures (finger print, access cards, CCTV, operators, security guard, and others)

Article "Building a Disaster Data Recovery Tips Server System Plan " These tips are suitable for the Protection Server System for Medium & Enterprise Company. Because the costs incurred are also big as needed and appropriate enterprise budget provided. The higher the "High Availability Server System? the desired protection, the more points that should be done also means greater costs.