Related questions

Schedulers in Linux vs CFS

2018-07-11 16:17 Qwerto imported from Stackoverflow

In linux, which are the differences between sched_rr,sched_fifo,sched_other and the Completely Fair Scheduler ?

See also questions close to this topic

  • Connection to secure lampp in linux remote server

    I am using lampp in remote linux server.I have secured access to /phpmyadmin and accessing it through node mysql with this code.

    var mysql = require('mysql');
var con = mysql.createConnection({
host: 'IP',
port: 90,
user: 'root',
password: 'PASSWORD',
database: 'DB',
native: true,
pool: { maxConnections: 50, maxIdleTime: 30 }
});


con.connect(function(err) {
  if (err) {
    console.error('error connecting: ' + err);
    return;
  } else {
    console.log('connected as id ' + connection.threadId);
  }
});

    It is giving me error as 

       error connecting: Error: Connection lost: The server closed the connection.

    while i can access it from browser "IP:8080/phpmyadmin".which gave window to enter username and password where username "root" and password "PASSWORD" gives access to database but not from code.

    Note:

    I could access it from same code when i didn't secure phpmyadmin with password and also changed port from default 80 to 8080.It might be issue with securing or the port changing.

    Need help plz.i can provide additional information if needed.

  • How to open Java swing UI in remotely connected linux machine through ssh

    I am connected to a remote linux machine terminal through xshell from a windows machine. I wanted to run java program which opens a swing UI in the linux machine but shows the UI in the windows machine from where I have connected to the linux machine.

    1. X (Windows machine ) -> Xshell -> Remote linux machine terminal
    2. Open UI here --> Xshell --> If I run the java program in this machine terminal.
  • Error while installing openstack using packstack in RHEL 7.5

    Openstack install on RHEL 7 ends with error

    192.168.*.*_controller.pp:                    [ ERROR ]
Applying Puppet manifests                         [ ERROR ]

ERROR : Error appeared during Puppet run: 192.168.*.*_controller.pp
Error: Execution of '/usr/bin/yum -d 0 -e 0 -y install openstack-selinux' returned 1: Error: Package: openstack-selinux-0.8.13-1.el7.noarch (openstack-queens)
You will find full trace in log /var/tmp/packstack/20180919-115228-waFUVu/manifests/192.168.*.*_controller.pp.log
  • How does the android split screen feature work on an OS level?

    I want to understand the underlying the concept of multitasking used in this split-screen functionality.Are they two different processes with constant context switches between them?

    I recently read about multitasking in Operating Systems and was intrigued by this functionality in Android. Being new to how the internals of multi-threading works, please help me understand these concepts. :-)

  • What is the relation between reentrant kernel and preemptive kernel?

    What is the relation between reentrant kernel and preemptive kernel?

    If a kernel is preemptive, must it be reentrant? (I guess yes)

    If a kernel is reentrant, must it be preemptive? (I am not sure)

    I have read https://stackoverflow.com/a/1163946, but not sure about if there is relation between the two concepts.

    I guess my questions are about operating system concepts in general. But if it matters, I am interested mostly in Linux kernel, and encounter the two concepts when reading Understaning the Linux Kernel.

    Thanks.

  • *NIX read()/write(): Are they re-entrant?

    What would happen if an interrupt/thread preempts in middle of execution of read(fd,buff,size) or write(fd,buff,size) functions and then it returns to where it was in read()/write()? Can we assume it will continue reading or writing to the file descriptor without loss of data and order?

    Assume there is no shared variables to worry about synchronization.

  • Scheduled task with Django

    I'm working on a web app which manages events. I store data about them such as their starting time. I would like to execute a task (in this case, sending a mail to the event administrator). There shouldn't be there more than 10 events a day and the sending would just involve one mail. I've been reading about different approaches such as a writing custom command amd executing it with cron, but as long as the sending time depends on each event, I think it wouldn't be a nice solution. I have read as well about Celery, but in this case I think that would be a too powerful and heavy tool for my needs. Is there any other alternative?

  • Schedule a Task to run X times every Y seconds

    Currently I am trying to develop my own Timer in Java. An issue I found however, is utilizing the ScheduledExecutorService to run a task every few TimeUnit's and execute a Runnable X times. Currently I have an idea, however when the milliseconds are shortened, precision is lost.

    import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.*;

public class Timer {

    private static final ScheduledExecutorService SCHEDULER = Executors.newSingleThreadScheduledExecutor();
    private static final ExecutorService SERVICE = Executors.newSingleThreadExecutor();

    private final Runnable action;
    private long initDelay;
    private long delay;

    public Timer(Runnable action, long delay) {
        this.action = action;
        this.delay = delay;
    }

    public Timer(Runnable action, long initDelay, long delay) {
        this.action = action;
        this.initDelay = initDelay;
        this.delay = delay;
    }

    public Future<?>[] runSynchronously(long count) throws InterruptedException {
        List<Future<?>> futures = new ArrayList<>();

        synchronized (SERVICE) {
            SERVICE.wait(initDelay);

            while (count > 0) {
                futures.add(SERVICE.submit(action));
                SERVICE.wait(delay);
                count--;
            }
        }

        return (Future<?>[]) futures.toArray();
    }

    public ScheduledFuture<?> runAsynchronously(long count) {
        ScheduledFuture future = SCHEDULER.scheduleWithFixedDelay(action, initDelay, delay, TimeUnit.MILLISECONDS);
        this.cancel(future, count);
        return future;
    }

    private void cancel(Future future, long count) {
        SCHEDULER.schedule(() -> future.cancel(true), initDelay + (delay * count), TimeUnit.MILLISECONDS);
    }

    private void cancel(ScheduledFuture future, long count) {
        SCHEDULER.schedule(() -> future.cancel(true), initDelay + (delay * count), TimeUnit.MILLISECONDS);
    }
}

    Currently when I use this...

    final AtomicInteger atom = new AtomicInteger(0);

Timer timer = new Timer(() -> {
    atom.incrementAndGet();
    System.out.println(atom.get());
}, 1000, 1);

timer.runAsynchronously(10);

    The counter will reach 4. When I previously changed the delay to 333, it could produce all the numbers. Before this, I tried creating a for-loop to schedule it repeatedly, however I disliked the accumulation of multiple ScheduledFuture<?> instances.

    Is there a way to define how many times to run, or to create a reliable system that cancels the event without the loss of precision?

  • Is job swapping possible for blocked process?

    I am reading materials for operating system. And I learned that job swapping is a one of scheduling method. When the processes in ready queue are having too much memory, job swapping happens and some processes are swapped out and swapped in later.

    But, is is possible to swap out processes from blocked processes? In state transition diagram of process, there are many states; ready, running and blocked. When running process faces I/O, the process will be blocked and will wait for I/O job ends. I wonder if memory is not sufficient and some jobs needed to be swapped out, process from blocked processes can be swapped out or not.

    I briefly heard someone says it is not possible and it is related to 'direct memory access'. But I don't remember where did I heard/read it. I think it is just fine to swap blocked process except it will take more time than swapping out from ready queue. Is there any reasonable explanation?