No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

The asynchronism between the virtual machine time and DDC time causes a user on DDC occupying multiple virtual machines

Publication Date:  2012-10-23 Views:  53 Downloads:  0
Issue Description
Version: R2C00SPC200
The problem description
In the pool mode, user login in and be responded by the feedback that there is none available virtual machine, but actually there are more machines than the number of the positions, and occasionally a few users are not have the available virtual machines.
Alarm Information
None.
Handling Process
1. Examine the using state of the virtual machine on the DDC, we find some users are occupying two virtual machines, and then some other users may encounter the state that there is not any available virtual machine. As the displayed as the following picture:


2. Analyze the logs of the virtual machine, we find there is time hop in the virtual machine which is occupied by mistake. This virtual machine on the site has installed the CSP system, it will change the virtual machine’s time when the user login in the CSP system, the time in the CSP system is different from the time in the AD domain, this cause the time hop in the virtual machine, and the virtual machine may lose connection. This problem has been described in the case SC0000682121.
3. The time of the virtual machine is modified by the CSP is asynchronous with the DDC time, and the heartbeat between the virtual machine and DDC loses its function, this lead to the DDC can’t identify the false virtual machine, when a user login again he/she will be distributed to another virtual machine, therefore this user has occupied multiple virtual machines.
4. Meanwhile, we configure to let user connect the same virtual machine once he/she lose the connection within two hours, as a result, the un-connected virtual machines can’t be released for the other users’ use. The configuration is alike as case SC0000697683.
5. The function described in the Step 4 is only effective at the state that the user breaks the connection at the breaking node, and it’s useless to deregister, close or restart. Therefore, a quick method to solve this problem is to restart the false virtual machine, if there are many false virtual machines and each one’s restart speed is too slow, you can close all DDC and restart them again.
6. Adjust the time in the CSP system to be synchronous with the one in the AD domain, there isn’t any time hop of the virtual machine, and it’s normal for a testing time. Thus, the problem has been solved.
Root Cause
1) The modified time of the virtual machine is not synchronous with the DDC time, and the heartbeat between the virtual machine and DDC loses its function, this lead to the DDC can't identify the false virtual machine, therefore a user can login in multiple virtual machines.
2) In order to convenience the position server to work and don’t need to restart the office environment during the interval, we configure in the early pool strategy to let user connect the same virtual machine once he/she lose the connection within two hours, as a result, the un-connected virtual machines broken unconventionally can’t be released for the other users'  use.
Suggestions
None.

END