【 Weak current College 】 Server Administration most likely to make the top ten errors

1.UPS (uninterruptible power supply) for the use of

A shop is squeezing its network systems last life. Despite the very old server almost difficult to meet the needs of store operations, but the operator even replace the battery in the UPS are reluctant, in fact the battery

Every two years should be replaced. Finally one day suddenly, and the UPS is too old, it has no control over the server shuts down. Most had already run out of battery power, so it will soon be dead. As a result, control anomalies controller was unable to properly shut down the server role. Store daily functioning and trading issues on the server. Fix server took three days, and compensate for the resulting losses may take a longer time. In fact, a $ 75 value only batteries can effectively avoid such things from happening.

2. no tidy up the need

A telephone call into corporate headquarters said a branch server problems, network administrator ran out of the Office, drove an hour to reach the scene. As a result, he found that the operating system files are damaged, so, all he can do only reinstall the operating system. The branch network administrators unable to find the Setup disk, no matter, the headquarters of the administrator, but in his Office-from here, an hour, round-trip fare for two hours. Picked up a backpack and carry it. All the things you might need, including your user's operating system backup, you should put in a backpack. This is very easy to do, not what it takes, but update on issues such as the operating system can play a key role.

3. do not install the patch

A toy manufacturer's server connection problems, allowing staff and their files are lost, production was forced to stop for a couple of days. Loss serious! a newly hired by the network administrator is responsible for resolving the crisis. He quickly discovered that the previous network administrators in three years ago when the installation system does not have to installthe patch.

A patch is available free of charge, maybe they're not open, but they are usually very easy to configure, no system administrator are able to do. Before the Administrator said, because the system is running very well, so no need to patch. This argument was apparently very irresponsible.

4. backup work improperly

A hospital Office storage a lot of medical files and records server has failed. Because the server has no information can be accessed, the Office of the functioning of basic stop. Network administrator immediately brought out a backup tape of the document, try to repair. When she discovered that the tape is blank, the heart sank. She checks the additional backup tapes, actually all is empty! she has checked the log and found that the Office colleagues every two years to replace backup band, just strap placed into never install the backup software on the server, no one knows the backup with a problem, because no one checked, but only two years to replace blank backup every day. This makes the hospital suffered heavy losses. In fact, as long as a database administrator to do a simple backup checks you can avoid this major problem. In fact, they are doing that every day is.

5. poor quality cable line engineering

A bank's network is often a problem, for which they dedicated to wiring box do the check, check, many RJ-11 and RJ-12 plug into the socket RJ-45. And in each socket can insert a toothpick, Chief of the cable line no wonder will be a problem. Many network problems are attributed to the wrong cable connections, therefore, shrewd Manager the best allows a business license, secured and trustworthy cable line contractor to frame add cable lines.

6. equipment changed hands many times too many, the internal Constitution of the problem

A professional sports organizations to buy a brand-name server, but it just started working on the problem. The network administrator to operating system vendors and hardware vendors to help requests. Later discovered, providing the "brand name" Server resale businessmen to this server configuration is a brand of memory, non-brand-name of the disk controller, as well as non-branded external disk drives, just because these components are cheaper. The hardware manufacturer and operating system vendors refuse to provide support, because the server is non-brand-name of the widget by confusion, its structure is not easy to be identified.

7. no contract signed authorization

An Office last year, spend a lot of cash to purchase a server. The system has a RAID5 redundant protection, power, and 24 × 7 support, and a year later, the drive is broken. Ensure 24 x 7 support staff came, he telephoned the hardware manufacturers, manufacturers asked what his contract number, and the Office has not signed the authorization prior to the contract. "It doesn't matter," manufacturers, "they leave authorization expires two years time, I will be in five or six days to give you a replacement drive. ”

However, manufacturers of tolerance is not enough, it is best to ensure you have the full set of 24 × 7 support, go to an office supplies store label on each label contract number written on the authorization and technical support telephone number, and then put them on each machine.

8. in the absence of a test environment

A few years ago, a software development company installed a new workstation, it is the fastest RAM, the fastest hard drives and the fastest processors, it will serve as the Chairman of the new station. Shortly after the installation is complete, the Chairman of the Board received from his biggest partner in the company of a request, using it to test a new artificial universe station communication platform, the result is a blue screen. When he surprised the restart of the computer, what internal saved information at all. InAfter rebuilding the system, he also spent four days from that heap operations guide get over it.

Another company's Board Chairman smarter, built a test network. In the system before you upgrade, they will do a test network, do the test again and find the error, set the network continuously test over and over again, until it is completely correct. Only in the test results are fully and correctly, they will really expand the system upgrade.

The first company President never leave their server when meter.

9. storage capacity planning

An art company five years ago to buy a server, the server can support six 8G-byteRAIDArray5 drive. In order to save money, the company adheres to only buy four 4G-byte drive. Network Administrator said that soon they will need more space, the company eventually compromise, bought two drives, so settled the matter. In three years after they had serious problems of space, they even had to delete only the small files 50Kbyte. They need to expand capacity, while the drive has 8G-byte buy, bigger servers also support it. To resolve the issue of a new additional subsystem will need more than the original server. In this way, they have more than two years ahead of the replacement server. Therefore, do a good storage capacity plan that will make you save money, or even greatly extend the life of your system.

10. error action is the greatest risk

Afternoon Office power suddenly broke, tight Office Manager that would impair their two servers, therefore he took quick action – he went off the server. Back home, he also for his fast action. Can the next morning, when he returned to Office opens two servers found nothing inside content. In fact, yesterday when he presses the server switch, the server is a critical file complex update work, he discontinued server work affects another key database file. The results of repair network work took two days.

All protection network to work, this is the most unpredictable.