|
|
|
|
The System Processes page is displayed when clicking on the System Processes link on the Services tab of the Container dashboard. It presents a table reflecting all the running processes inside the Container. The table provides the following information:
|
Column Name |
Description |
|
|
PID |
The process ID. |
|
|
%CPU |
The percent of the CPU time the process is currently using. |
|
|
|
The percent of the RAM size the process is currently using. |
|
|
Command |
The command that is used to launch the process. |
|
|
|
The relative priority of the process assigned to it by the user. The negative values mean that the user has manually increased the priority, the positive values - that they have decreased it. |
|
|
Pri |
The absolute priority of the process assigned to it by the process scheduler. On a Linux Node, the range is from 0 (the highest priority) to 39 (the lowest priority). The usual process priority is 30. On a Windows Node, the range can be from 0 (the highest priority) to 31 (the lowest priority). The usual process priority is 8. |
|
|
RSS |
(Resident Segment Size) The size of physical memory the process really uses (in Kilobytes). |
|
|
|
The state of the process. The possible states are: R - runnable, on the run queue; S - sleeping; T - traced or stopped; D - uninterruptable sleep; Z - defunct, "zombie". If two letters are shown, the second letter means the following: W - has no resident pages; < - high-priority process; N - low-priority task; L - has pages locked in memory; s - the process is a session leader; "+" means the process is in the foreground process group of its control terminal. |
|
|
Time |
The total amount of the CPU time the process has used so far. |
|
|
User |
The user the process belongs to. |
|
To have the information in the table refreshed automatically with the current values, click the Enable Autorefresh button. It is worthy to note that only the table on the current page is refreshed, which takes much less resources in comparison with refreshing the whole Parallels Infrastructure Manager page.
On a Windows Node, you may select any number of processes by ticking the checkboxes against the corresponding processes (tick the uppermost checkbox to select all the processes at once) and end them by clicking the
End Process
button.
On a Linux Node, you may select any number of processes by ticking the checkboxes against the corresponding processes (tick the uppermost checkbox to select all the processes at once) and send them a standard signal. Choose the needed signal on the drop-down menu and click the
Send Signal
button. The following signals can be sent:
SIGHUP
- is a hang-up signal. It is often used to ask a daemon process to re-read its configuration.
SIGTERM
- sends the termination signal to the process. This is the best way to give the process a chance for an orderly shutdown and proper data saving. As the process might be able to catch this signal and stay alive, you may have to make use of the
sigkill
or
sigint
signals.
SIGCONT
- continues the process causing it to resume.
SIGSTOP
- stops (suspends) the process. The process will still be on the task list.
SIGINT
- causes the process to immediately interrupt. The signal is very close to
sigkill
, the difference being that, unlike
sigkill
, it can be caught by the process and ignored if the process gets out of hand. In this case you should send
sigkill
to shut down the process.
SIGKILL
- unconditionally kills the process. Mind that sending
sigkill
to any process removes any chance for it to do a tidy cleanup and shutdown, which might have unfortunate consequences.