In Fedora, the ramdisk is disabled by default, that is, it is compiled into the kernel, but as a module, which will be loaded into the kernel when actually used. So, it is necessary to compile the kernel manually to select the ramdisk as the built-in modules.

In Ubuntu, the ramdisk support is enabled by default, so the steps are easier compared to Fedora.

Using the following commands to list the ramdisk:

The output will be like:

Format the ramdisk as an ext4fs:

Then mount it:

By default, Ubuntu doesn‘t install the ssh server, so the following command will fail:

Using the following command to install the ssh-server:

Then all things will go well.

It is easy to do with real root user:

The CPU status can be checked by:

To re-enable the CPU:

The SELinux has some negative effect on the self-defined filesystem, so it is necessary to disable it, using the following command:

will list the status of the SELinux,

To change the status of SELinux, modify the configure file in:

The easiest way is as follows:

Add the following two lines to the end of the file:

This enables the Chinese characters to be recoginized by the system.

I am working on a project which has something to do with a remote server, and I have to ssh to that server, but the server is Fedora and my os is Ubuntu, and it turns out it takes a long time to resove the host/server IP, the solution to this problem is:

Afer a failed attempt to upgrade the Ubuntu13.10 to Ubuntu 14.04, the system shows a popup window everytime the system boots, the title of the window is "System Program Problem Detected". The following method can be used to solve the problem:

This command will delete the crushed software of the system so that the reporting utility won‘t work, thus eliminating the annoying message.

Recently, a project requires frequently mounting and testing a filesystem, and it is very likely that the newly-mounted filesystem will crash for all kinds of reasons, and after the accident, it is no longer possible to umount the filesystem in a normal way. The error message from the OS is as follows:

With the -l option of umount, the result can be seen as follows:

When using CloudDrive, it is likely that the software may only allow you to synchronize the files or folders under a specific folder, if you have many other directories to synchronize, you have to copy files, but there is a simple way to achieve the same goal without copying files—the symbolic link.
In Linux,the following command can be used to create a symbolic link to another directory:

\

In windows, use the following command:

1. Add extention to all the files in a directory: \

   for f in *;do mv "$f" "$f.png";done

2. Rename: \

   rename ‘s/old-name/new-name/‘ files

The first useful option of ssh is the GSSAPIAuthentication=no, which accelerates the connection startup time. Another useful option is the X11 forwarding, with the -X option, you can use the X11 application from the localhost.(Especially for emacs).

The problem I encountered is like this:
When I want to logout the xsession of ubuntu, it fails to close all the running application and return back to the login screen, so I start a terminal by ALT+CRTL+F1, from which I did a reboot. After the system booted, I typed my password to my account, but it returned back to the login session again, and the process continued.

There was some error messages in ~/.xsession-error file, and the mode bits of ~/.xauthority file has changed to root only, so the correct way to deal with the problem is use the following command:

Copy the existing source configure file to the desired machine,the file is:

I have a working copy of a cvs-controlled repository, and after I added some new directories, I failed to add the files in the directories to the version control system, because the CVS complains that the node corresponding to the new directory does not exist, the simplest method to solve this problem is the cvs add command:

Then, in emacs, just type c-x v v, all things will be handled by the smart emacs…

The repositories I use on my machine is not under cloud back-up, so I just want to backup my repositories to my cloud, so I followed the following steps:

1. Create and init the new repository but override the existing CVSROOT environment:

   cvs -d /home/wujing/KuaiPan/myrepos init

2. To make sure that the proper CVSROOT directory is correctly setup, ls the directory;
   
3. Import the project files into the new CVS repository:

   cvs -d /home/wujing/KuaiPan/myrepos import -m "start papers backup" memory_contention wujing start

   In the above command, the files in the source files in the project directory will be copied into the CVS/memory_contention directory.

4. Backup the original files and checkout the versioned files from the cvs repository:

   cvs -d /home/wujing/KuaiPan/myrepos co memory_contention

   Be sure to specify the correct cvsroot with the -d option.
   Note:Be careful about the current working directory when using import and checkout command.

I tried to copy the entire directory and checkout from the new repository, it works fine, which means that the CVSROOT repository can be migrated easily.

I am working with multiple computers, and I want to keep the source files be synchronized among all of them, and these source files are under cvs control in my localhost, so just do a checkout from the remote computer with the following command:

Make sure that the ssh connection is enabled in the remote host. When I am trying to commit the changes to the remote host, the emacs complains about the missing ssh-askpass file, the right way to solve this problem is to use password-less ssh connection, the detailed steps are as follows:
Suppose my working host1‘s IP is 172.31.8.22, the working host2‘s ip is 172.20.161.56, and the repository is located in host1, so when I ssh to host2 from host1 and want to commit the changes to host1, host2 must have access to host1 without password, do the following configuration:

Recently, I am frequently downloading/uploading data from a BT site, and I want to cause little traffic to my disk so I just download the movie to my Ubuntu /run/shm directory, the problem is that the default size of this partition is half the size of system memory, which is 4G in my case, and the following command can be used to enlarge this tmpfs:

Sometimes the filenames of *.torrent files (haha, that‘s it, you probably got it!) may be invalid coding, the following script can rename all of them:

The cat -n adds number to the output of ls, which can act as a running index to the files, and the second line of the ls | cat -n output is the original filename, so the read n f command reads exactly all the information necessary to finish the renaming, and the do… done does the renaming.