Shulou(Shulou.com)06/01 Report--

When you encounter a very terrible plot in a horror movie, or when you see something that makes you feel shocked, I believe that now we are definitely not weak shouting "so scary", "so scared" or so on. The first subconscious sentence is either "grass mud horse" or "I wipe". I am the dividing line = speaking of today's topic, it is actually very interesting and thrilling, it happened on Friday (April 12, 2013.12) off work, in fact, I absolutely did not expect this situation, nor did I expect the result of the matter. Students who often read the Nine Uncle blog will probably know that I use Citrix's XenServer more here, XenServer as a basic platform for virtualization, and the three big brothers VMWare and HyperV can be regarded as carving up the majority of the virtualization market share. Of course, VMWare is the most powerful, and the details of market share are not discussed here. To be fair, XenServer is still very good, from the use of the past few years, it has few problems, because the underlying Linux system, so the Guest system is Linux would be better. It so happens that I have a test machine in hand, the version is updated at any time, and now it is XenServer6.1, but the hardware of this machine is a little miserable, it is a double U tower server in 2007. On this test machine, mainly some of my usual test environment machines, and in order to facilitate WDS to capture the image, put Win7-32 XP 64, XP, 2003 a total of four operating systems, as well as one of my personal website. Because XenServer is stable to use, and this is a test environment, I never thought about what would happen if the machine broke down. The overall "feeling" is that it's okay to lose the data above, and it's not a big problem. I am the split line = near the end of work on Friday, I found that the speed of capturing the WDS image was very slow. I looked at the server and did not restart it for several months, so I restarted it remotely because of suspected memory leak. But this is the beginning of the tragedy. Obviously, you can get off work at 05:30 and go. After rebooting, the disk of XenServer cannot be mounted this time. XenCenter prompts that local storage mount failed and needs to be repaired. There was a big chill in my heart at that time. These virtual machines, which I thought didn't matter, broke me down at once. Although these things are not important, but if the hard drive is really down, it is still troublesome to recover, especially the WDS image, which does not have any technical content, that is, it is pre-installed with some software needed for the company's office, and it will take two days to reorganize it. In addition, there are virtual machines with personal websites on it. This makes people have a big fire. Although this is a website with poor traffic, it will take many days to re-record the information, and I don't know if I can find a backup of some things. = I am the dividing line = in fact, I finally managed to mount the hard disk, how to achieve this is a technical problem, and it has nothing to do with today's article (of course, from 05:30 to 08:30 on Friday. The suffering of the mind during this period is incomprehensible to others). Here I would like to talk about some of my views on this matter, hoping to be of some use to you. I also hope that you will encounter less annoyance caused by this kind of data loss and say less such things as grass and mud horses. Is the testing machine really just a test? There is no doubt that in the original idea and design, this testing machine is indeed the test content of running, there is no doubt that the data above is not important, there is no need to make a backup, do not need to consider any disaster preparedness. However, in practical use, we will encounter the operation of migrating something that has passed the test to the work environment. At this time, the virtual machine on the test machine is not for testing, but valuable. The question is not whether the test machine is or not, but that the test environment sometimes produces valuable data. 2. Is data confirmation optional? The data on the test machine is not important at first, because backups can be found on other online machines. But if we need to verify a feature that exists in a more complex environment, such as Lync's full best practices deployment, such as Exchange's full best practices deployment, which requires at least 5 virtual machines to participate, does it really matter if the data is lost in a test environment like this? Obviously not, even in a test environment with no reference data, it takes a lot of time to build such a platform. At this time, if we say that the environment that has just been built is inaccessible because the hard disk storage is dead, I believe there will be 10,000 grass-mud horses surging past in your hearts. 3. Is it really okay if you can't get it back? Just because of the time constraint, we can say that it will not work if the data cannot be recovered, but if the storage is really down, and this is still a physical machine billed as a "testing machine," then it is possible that apart from being self-reliant, it is really difficult to find a reason to spend a lot of money on operations such as data recovery. Everyone is happy if they can find it back, but they can't find it and cry for their father and mother. From the moment the data produces value, it has already been decided that we can't give up the data. 4. Why is your mentality changing? From the beginning, there is no doubt that we paid little attention to the test environment. However, as we have more and more things to test and we "communicate" with the tester more and more closely, we will find that we have a different emotion with the tester. When emotion becomes a habit, sudden loss can be extremely unacceptable. Don't wait until it's too late to regret it. To solve this problem, in fact, the method is very simple, that is, deploy the corresponding backup plan for the test machine, and make a good disaster preparedness plan to operate as a formal environment. However, most of the time, we take this matter lightly and don't take it seriously at the beginning, and we end up dying 100 meters before the finish line. In the IT industry, preparing in advance is a positive commendatory term. I am the dividing line = when we find the data valuable, it is often the time when the disaster has already happened.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.