Detailed analysis of ten common faults in Unix system

SCO OpenServer 5.0.5, as a highly efficient and stable multi-user operating system with high security performance, has been widely used in finance, insurance and telecommunication departments. In the daily maintenance work of the system, some system faults are encountered sometimes. The author of the common 10 questions summed up a bit, I hope to be helpful to everyone.

First, after powering up the computer, the main console screen appears as follows: Boot not found cannot open stage 1 boot failure:error loading HD ()/boot, and then panic.

Analysis: This indicates that the boot file in the system root directory is missing or not found. The boot file is an interactive program for loading and executing UNIX programs, primarily for loading and executing the core of the UNIX operating system, during which boot reads/etc/default/boot, displays the boot process information, loads and executes/unix, So when the boot file is missing or not found, the system boot fails.

Workaround:

If the boot file is missing:

1. Boot disk (emergency floppy disk or installation CD).

2. Enter HD UNIX at system boot prompt, which is used to load the UNIX core from the hard drive. Then enter the root user's password and enter Single-user mode, at which point the root user prompt # appears.

3. Uninstall boot file system, execute command: Umount/stand.

4. Load boot file system, execute command: Mount/stand. (Note: SCO openserver 5.0.5 The/stand file system is set to read-only and cannot change properties when the system starts, so that the/stand directory is writable. )

5. Use the command to install the boot floppy disk into the system, execute the command: MOUNT/DEV/FD0135DS18/MNT, and then use the command: Cp/mnt/boot/stand to copy the boot file to the/stand directory.

6. Use the command Umount/dev/fd0135ds18 (or umount/mnt) to uninstall the boot disk.

7. Remove the boot disk, typing haltsys or reboot command to reboot the system.

8. After the system normal boot into the root user, the/boot file of the owners and groups to bin can be.

If the boot file cannot be found:

This situation is generally due to the CMOS configuration of the hard disk parameters, at this time just let the system to automatically detect the hard disk, the parameters can be corrected.

Second, after the computer power, the main console screen appears the following information: Unix not found, and then panic.

Analysis: This indicates that the UNIX system core is missing. Because the/unix contains the core, a corrupted UNIX file will cause the system to fail to mount the parallel kernel, causing the panic to occur.

Workaround:

1. Boot Disk Guide.

2. Typing FD UNIX under system boot prompts. Z ROOT=HD (41) This command is used to load the UNIX core from the boot disk and install the root file system on the hard disk. Then enter the root user's password and enter Single-user mode, at which point the root user prompt # appears.

3. Uninstall/stand file system, execute command: Umount/stand.

4. Load/stand file system, execute command: Mount/stand.

5. Install the boot floppy into the system and execute the command: mount/dev/fd0135ds18/mnt.

6. Implement Cp/mnt/unix. Z/stand restores the Unix files in the root directory of the hard disk. This operation enables the compression of the Unix kernel files on the boot disk to be copied to the hard drive without the user having to extract the kernel files.

7. Uninstall the boot disk and reboot the system. After the system boots into the root user, the owner and group of UNIX files are changed to Bin and mem.

Third, the machine after the power, at the system boot prompt type carriage return, when the first screen is refreshed after the display of the second screen, there are many error messages, such as: Init:command EXEC/ETC/BRC 1>/dev/console 2>&&1 failed To execute, errno=9 (exec's shell failed), and so on.

Analysis: This indicates that the system/bin/sh file is damaged or missing. Because SH is the interpreter for Shell standards, job control, and restriction commands, the damage or loss of the file will cause system startup to fail.

Workaround:

1. Boot Disk Guide.

2. Press ENTER at the system boot prompt, and when prompted to insert the root disk, remove the boot disk and insert the root disk so that the system is all booted by the emergency boot disk until a prompt appears: #.

3. Execute commands mount/dev/hd0root/mnt and cp/bin/sh/mnt/bin/sh restore/bin/sh files under the root directory of the hard disk.

4. Uninstall the boot disk, and reboot the system, the system after normal boot into the root user, the/boot file of the owner and group to bin can be.

Four, the system can guide, also can normally handle daily affairs, but every time shutdown will be on the main console screen error message:/etc/initscript:/etc/uadmin:notfound, and then the system crashes, resulting in the shutdown can not be normal. So every time you start to clean the file system, waste a lot of time, too many times can also seriously damage the file system, resulting in system paralysis.

Analysis: System file/etc/uadmin is lost or damaged.

　Workaround One:

1. After the system boots and cleans up the file system, registers with the root user.

2. Under the super User, use the command mount/dev/fd0135ds18/mnt to install the root disk into the system.

3. Copy the/etc/uadmin file from the floppy disk to the hard drive: Cp/mnt/etc/uadmin/etc/uadmin.

4. Uninstall the root disk and change the owner and group of the/etc/uadmin file to bin.

Workaround Two:

Solution one is for stand-alone, only the situation of emergency disk. In the case of multiple machines interconnection, there are actually more concise ways:

1. Use the network FTP or RCP command to obtain uadmin files from the normal machine.

2. Use a floppy disk using the TAR command to obtain the Uadmin file from the normal machine and tar into the corresponding path of the faulty machine.

After the machine is power on, type the return key at the system boot prompt, and when the first screen is refreshed, the following line appears at the end of the second screen: kernel:i/o bufs=600k, and then the system hangs.

Analysis: This indicates that the system's/etc/init files are missing. The INIT program begins execution in the final phase of core initialization with a process number (PID) of 1. The process starts the processes listed in the/etc/inittab file in a certain rule, booting the system into the specified run level. It first reads the Initdefault entry in/etc/inittab, and when all the processes are created successfully, Init enters the loop wait. If the/etc/init file is compromised, the new process cannot be created, causing the system to hang.

Workaround:

1. Boot Disk Guide.

2. Press ENTER at the system boot prompt, and when prompted to insert the root disk, remove the boot disk and insert the root disk so that the system is all booted by the emergency boot disk until a prompt appears: #.

3. Execute command mount/dev/hd0root/mnt and cp/etc/init/mnt/etc/init to restore the init files under the root directory of the hard disk.

4. Uninstall the boot disk and reboot the system. After the system is properly booted, enter root and change the owner and group of the/etc/init file to bin.

Six, the system starts prompt: NO OS.

Analysis: If the system can work correctly, the boot block of the system hard drive is destroyed.

Workaround:

1. Boot Disk Guide.

2. Enter HD UNIX at system boot prompt, which is used to load the UNIX core from the hard drive. Then enter the root user's password and enter Single-user mode, at which point the root user prompt # appears.

3. Execute command INSTBB hd/dev/hd0a to write the boot block partition to the hard disk, and Dparam _w to write the master boot block code to the hard disk.

4. Reboot the machine.

Vii. system startup cannot enter multiuser mode.

Analysis: If the system after the start of the display: Init:can not open/etc/inittab error 2, in ctrl-d after the system can not enter multi-user mode, that means/etc/inittab file loss,/etc/ The inittab contains instructions for init. When the Inittab is lost, Init cannot execute the system boot instruction and the system cannot enter multi-user mode, and the system still maintains single-user mode when the ctrl-d is pressed. Because the/etc/inittab contains instructions for other parts of the system, a special/etc/inittab is written to the root file system floppy disk when the "emergency disk" is created. Therefore, if the problem of/etc/inittab loss to the system can not simply Instead of copying the etc/inittab from the emergency disk to the hard drive, the/etc/conf/cf.d/init.base is copied to the/etc/inittab and then relink to the core.

　Workaround:

1. Start the machine and enter the system maintenance status.

2. Execute command cp/etc/conf/cf.d/init.base/etc/inittab and/etc/conf/cf.d/link_unix-y to re-establish a new/etc/inittab file. and reboot the system.

Eight,/ETC/BCHECKRC file is missing.

Analysis: If the system in the boot display the following information:/etc/initscript:/etc/bcheckrc:not found, indicating that/ETC/BCHECKRC file is missing. This file is a symbolic connection file and can be recovered from the source file if the source file is intact.

Workaround:

1. Start the system and enter the system maintenance mode.

2. Run fsck to clean the root file system.

3. Execute command CP/OPT/K/SCO/UNIX/5.0.5EB/ETC/BCHECKRC/ETC/BCHECKRC, or ln-s/opt/k/sco/unix/5.0.5eb/etc/bcheckrc/etc/ BCHECKRC. Reboot the system.

Nine, the system boot to display the following information: Init:single USER mode/etc/initscript:/bin/sulogin:not found.

Analysis: Description/Bin/sulogin file is missing. This file is a symbolic connection file and can be recovered from the source file if the source file is intact.

Workaround:

1. Boot the system into multi-user mode (sulogin file loss cannot enter Single-user mode).

2. Run fsck to clean the root file system.

3. Execute command cp/opt/k/sco/unix/5.0.5eb/bin/sulogin/bin/sulogin or ln-s/opt/k/sco/unix/5.0.5eb/bin/sulogin/bin/ Sulogin. Reboot the system.

Login:error-failed to initialize policy manager is displayed when the system is started. (Ifor_pm_fatal).

Analysis: The reason for the above error is usually a system time error, which is earlier than the installation time of the operating system. This is generally due to the loss of the machine's BIOS parameters.

　Workaround:

Boot into the machine BIOS parameter settings, modify the machine time for the correct time. Alternatively, start the operating system, enter Single-user mode, and modify the system with the AskTIME or date command for the correct time.