problem:
After a certain operation of upgrading the system kernel, after the server restarted, the alarm monitoring platform did not receive the device restart alarm and recovery alarm at that time, and the alarm was displayed in the alarm after a delay of eight hours
System Suse 11 SP1
Troubleshoot:
Check /var/log/messages and find that the time jumps during the restart period
Check the local time configuration:
HWCLOCK="-u" configured by /etc/sysconfig/clock
The time zone configured by /etc/localtime is CST-8
Avoid
HWCLOCK="--localtime" configured by /etc/sysconfig/clock
Thorough solution: regenerate the init system driver file of the current kernel version,
backup the initrd file of the current system, soft link
backup the original init file
cd /boot;cp initrd-2.6.32.12-0.7-default initrd-2.6.32.12-0.7-default_bak
After confirming that it is correct, regenerate the init file
mkinitrd
Whether the restart observation time is normal
The reason for this problem is that in addition to the above two time zone and time configuration files, the suse11 system also has these two files under initrd under /boot. When the system is enabled, the file under initrd is first used to set the system time, /etc/sysconfig/clock and Startup after /etc/localtime is only to fine-tune the time, and will not reset the time. When the file configuration in the two places is inconsistent, it will cause time confusion, which has been resolved in suse11 sp3 (should)