Overview
It is possible that you have to perform preventative maintenance on a server that contains a virtual machine (VM). For this activity, you are required to restart the operating system. In addition, this maintenance may be performed on a VM that is the slave of a slave/master configuration while the master is on another server.
You wish to know more details for the execution of this activity in the server/node.
Solution
- In order to shut down the system safely, perform the following actions:
(log in as root)
su - textpass
tp_stop
exit
shutdown - Then boot up the node again. Once the node boots up, all processes should start up automatically. you can verify by running command
tp_status
(as user textpass). All processes should be in 'operating' state, just like below: - If you find that any of the above processes display the state as Not Active, e.g.
You can start it up manually using thetp_start
command. In the example above, the command to use would betp_start --tp_fclient
, as the fclient is the process that is Not active. - Check that the MGR function has not started up (if this is the slave node on stand-by). Check with command below:
ps -ef | grep MGR
- Check the NTP sync status using the
ntpq -p
command, as there is a possibility that the system clock is lost after reboot and server starts with hardware clock.:
The line beginning with a star indicates the NTP server which is actively being used. Check the OFFSET column: the number should be near 0 (zero). If the number is high (in thousands), then please force sync with the NTP server with the following command run as root:systemctl stop ntpd; ntpdate ntpserver1; systemctl start ntpd
- Note that the Master Node should not be affected at all by rebooting the Slave Node, as these are independent servers. You can confirm that the Master Node is functioning normally using the same commands as above:
tp_status
:- The MGR function here must be Active. Run the
ps -ef | grep MGR command
: