The missing uptime explained
It seems that my box loses its uptime record after I have migrated it to the Yizhuang IDC. The two things are related, but technically speaking, they are not. Actually I have discovered gokeeper’s hosts losed their uptimes as well, but I did not know why before. Sometimes I thought that was because the Linksys switch we have used, but there are some evidence that it is not; also I have at a point doubted the Cisco devices were misconfigured or have some firewalling features, but that’s not the answell, either.
The problem surrounding it is that some changes in the FreeBSD kernel, say, in the 6-CURRENT age, has made uptime provided by TCP fingerprint inaccurate, as system timer was changed to 1000Hz.
So, it’s not Linksys, nor Cisco to cause the uptime hidden under the way.