Upgrade the server Have you considered these ten things?

Follow the method when upgrading the server to avoid problems and get peak efficiency. Erik Eckel lists ten things to note in this article.

Server deployment is always targeted, at least initially. Regardless of whether the server is deployed in a small company or a large enterprise, its role always changes over time. Single-purpose servers often need to perform other services and assume responsibility for economic growth, budget cuts, rack limitations, or other reasons.

This is why it is important to check the system regularly. Checking system resource load helps the organization optimize server performance and prevent downtime and other failures. However, system administrators can't just be satisfied with solving small problems. Add a bit of memory here and upgrade the disk there. You need to plan carefully before the server is upgraded, and keep the following 10 things in mind when upgrading the server to ensure that the system can run at peak efficiency.

1. First, you need to verify the data backup

Don't make any changes to the server, even if it is a minor upgrade, before ensuring that a verified data backup exists. Whenever the server is turned off, there is no guarantee that it will restart and work properly. Although the probability of this happening is small, I have seen it just because the Windows performance and security patches are installed and the server cannot be restarted after it is shut down.

2. Consider creating a mirrored backup

Some manufacturers offer IT professional disk cloning technology, which simplifies the recovery process in the event of a server failure. Some companies, including Acronis and StorageCraft Technologies, offer a common recovery option that can cause a failed server to be restored to a different bare metal chassis, resulting in extremely reduced downtime. If the upgrade fails, disk mirroring can not only recover data, but also easily recover those complex configurations.

3. Don't do multiple upgrades at the same time

Most IT professionals know the importance of reducing the number of system reboots, so novices always complete multiple upgrades with one shutdown. However, tasks such as adding disks, replacing memory, and installing add-on cards should be done separately. Why do you need to do this? This is because, after a day or two of anomaly, you want to isolate the upgrade that caused the exception. If you have performed multiple upgrades at the same time, the process becomes very complicated. If only one upgrade is done, it is much easier to track the underlying cause of the anomaly.

4. After monitoring, closely monitor the log file

After the system is upgraded, don't show the error because the server is restarted into the operating system, and take it for granted that everything is normal. You should monitor log files, error reports, backup operations, and other critical events more closely than before. Use Windows internal performance reports or third-party software (such as HoundDog from GRI Software or PacketTrap from Quest Software) to ensure that any upgrades or upgrades are completed as planned.

5. What OS is explicitly used?

When upgrading, it is easy to ignore the operating system running on the server. This is especially true when the server room is not standardized and multiple operating systems are included in multiple chassis. Even experienced administrators try to install 8GB on a 32-bit Windows Server 2003 server in the face of confusing chaos (which is common in corporate information services). Memory. Only systems that are initially updated (including 32-bit or 64-bit verification) will be updated to ensure compatibility with the operating system and to install additional RAM or other resources.

6. Make sure the chassis supports upgrade

The various hardware of the server is notoriously inconsistent. Manufacturers often change model and product configurations. No matter when you add a hard disk controller, hard drive or other components, you need to know the manufacturer's technical specifications online before ordering the parts needed for the upgrade. However, only when you open the chassis can you 100% determine if the server actually deployed within your organization is suitable for this upgrade.

7. Don't think that plug and play is a matter of course.

When installing new hardware, don't think that the device can work "plug and play" under the server's operating system. Ok (even if the manufacturer states that the component is compatible). Before scheduling an upgrade, you need to use Google Search to learn about the experience of other technicians installing the same components under the same operating system. Since the upgrade is done on the server, you also need to make sure that the component is included in the hardware compatibility list by the operating system vendor. You may wish to enter the server manufacturer's forum to learn about technical issues you may encounter when installing the same device on the same server.

8. Performance Optimization

Be sure to track all upgrades that require related software adjustments. For example, adding memory to Windows does not automatically optimize the performance of Windows using RAM. System administrators must also upgrade the server's virtual memory settings to optimize Windows operation to match memory upgrades. In addition, when a new disk is introduced, the page file needs to be moved to a new disk for performance advantages.

9. Cheap and good goods

Of course, cheap hard drives, memory, power supplies and other accessories are always easier to obtain. But when it comes to servers, cutting corners does not bring much benefit, only high-quality, highly available components can be applied to the server. While the cost of these accessories is slightly higher than other (poor quality) options, the performance and uptime advantages are sufficient to offset the additional costs that need to be paid.

10. Document Upgrade

To be sure, you will maintain a document for each server. For servers that have just been upgraded, update their documentation to indicate which components, manufacturers, vendors, and, if possible, the order number and serial number, or include warranty and support information. The more documents you have on hand, the easier it is to isolate and resolve problems that may arise in the future.

(Original: 10 things to remember when upgrading servers)