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

It took two days to build oracle for the first time.

Has been through too much hardship.

So make some summaries.

Summary, experience:

Because the installation was unsuccessful many times yesterday, before deployment today, we first summarized the problems that occurred yesterday, and took some corresponding measures ahead of time to avoid some unnecessary troubles.

Let's first describe the general process of yesterday:

During the installation yesterday, I first found the documentation on the Internet, and then began to deploy oracle's running environment SUSe11 on VMware. At that time, the configuration given on VMware was as follows: (processor 1p/1c memory 1G hard disk 10G and other systems installed after the completion of the configuration of the network, so that the virtual machine and physical machine interconnection, and then through the image mapped on the physical machine to configure the software source to install lrzsz tools Then transfer the installation package on the physical machine to the virtual machine for decompression through the xshell tool, by referring to the documentation, solve some dependencies (software packages) needed before oracle installation, create oracle users and corresponding groups, create the directories required for oracle installation and change the corresponding permissions, and then configure kernel parameters, configure users, and finally switch to oracle users. Start the graphics installer and start reporting errors, errors.)

I won't talk about the errors reported after the oracle installer starts, but there are a lot of problems during the above description of the process.

1) first of all, when solving the dependency (the software package was suddenly interrupted during installation-> through troubleshooting because the software source I configured was gone. At that time, the configuration was configured through the image mapped by the physical machine, but somehow the image suddenly disappeared, so the path in XXX.repo failed. Now I don't know why it is suddenly gone. But it's been here all day, and I don't dare to use it again. )

2) since the mirrors are gone, I can only send the images on the physical machine to re-build the software source. I have no choice but to wait. I finally see that the progress bar is almost over, but it is unexpectedly broken again.-> through error troubleshooting, it turns out that / mnt space is insufficient.

3) instantly speechless. You can only add a new disk, partition, remove the package under / mnt, mount it, restore the package back, and repeat the previous actions to configure the software source. Wait

4) then, when configuring the kernel parameters, the value does not know how to give the corresponding value, look up the data, and solve the problem.

5) then by browsing the document, I saw that the memory when oracle was running should not be too small. At that time, my memory was 1G, so I was afraid of something wrong. Add memory

6) forget that the environment variable is temporary, the system is rebooted when the device is added, and the failure of the variable causes the installer to start with garbled code.

7) because there was a mistake when installing oracle before. So when I finally installed it, I couldn't get up, so I had to install SUSE... from scratch. I didn't even think of taking a snapshot at that time.

8) start to repeat the previous steps again. It was an unknown error during the installation. Check the information. Wait

To put it bluntly, I was doing these meaningless things all day yesterday, dealing with mistakes and all kinds of strange problems (unreasonable plans, confused thoughts).

So this morning, before deployment, I reviewed yesterday's problems and took the following measures:

1) first of all, the configuration is sufficient at one time: processor 1p/1c 2G memory hard disk 40

2) after configuring the network, configure the software source with the mapping image to install the lrzsz software, then transfer the oracle package and image to the virtual machine, and immediately mount the image replacement software source and write it to the fstab file.

3) there is no problem to restart the test. Take a snapshot of the system to avoid errors and reinstall the system.

4) the following steps are the same: after each step is completed, a snapshot is made after the test has no problem (for example, dependency resolution, kernel parameter resolution. Etc.)

Since there are no worries, I can do whatever I want, and there is plenty of time for me to find information and analyze the meaning of each parameter in the configured file.

It's time for installation again. After a long wait, what I didn't expect was that I successfully checked the process and port at once, and I was happy to start it at that time (I also wanted to do it for nothing).

But it's a little too ostentatious. I thought there would be no problem when I got up. I didn't expect that there would be a problem in the modification of the database process limit. (generally, after installing oracle, the default processes of oracle is 150. sometimes it is not enough, and it is necessary to change the process to 400.) so, don't be happy too soon before the last step. It's a bad feeling to fall from a height.

After a period of information inquiry, it is because the problem of the initorcl.ora document has finally been solved, and today is a successful conclusion. Wait this Saturday, through the history file, and then sort out the specific steps and the meaning of the relevant parameters!

Note: the above is just a description of the experience I encountered in building oracle in the past two days, because the first contact does not have any deep insight and is not detailed. I will study it carefully as soon as I have time in the later stage, so please forgive me for the mistakes.

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.