How to automatically restart XiboClient when it freezes

How to automatically restart XiboClient when it freezes

Hello.

I think that the Xibo Client may freeze due to insufficient capacity of the PC or long operation of the PC.

(In my case, it was the latter. There were subplaylists and datasettickers on the screen, but they were all frozen. The operating system was normal. I looked at the log for that display on the CMS and there were no errors.)

As a countermeasure for this case, I am currently looking for the method in the title.

My current plan is to use PowerShell in windows.
If the CPU time of the XiboClient does not change for a certain period of time, it is assumed to be frozen and the XiboClient is restarted.

However, I wish I could do this with the standard WatchDog without using such a complicated method.
https://xibo.org.uk/docs/setup/windows-player-watchdog
(However, it seems that watchdog checks for the existence of the XiboClient process itself, so I think this is difficult.)

Or, if there are any points in the above method, please let me know.

Environment
CMS 2.2.1
Client v2 -201 (windows)

I look forward to your comments.

The Watchdog already does some quite complex checks on the Player (memory usage, whether certain key threads are running inside the Player), so the question is why it didn’t detect the Player as hung in your case.

When it’s in that state, first thing to do is open the Player status screen (press i) and that may give some indication of what the Player state is.

I would also suggest upgrading your Player to 2R253 (released yesterday) before you do anything else.

Hi, Alex.
Thank you for your reply.
Since then I’ve looked at all the settings for watchdog (including those not described in the documentation).
As you said, Alex, I certainly confirmed that players have already taken quite advanced measures.
(The Process.Responding method seemed to be used to detect hangs.)
However, the hang in this case could not be detected by this detection method.

I was able to check the player’s status log from the CMS.
Only the “No active Process” log remains after manually restarting the player application after the hang.
No information about the hang itself was available.
So, as Alex says, “Why did the player hang but not detect it this time?” is the issue.

I have no plans to take advantage of the new version 2R 253, but I would like to test it if I can make some progress on this issue.

Based on the above, my questions are as follows.

・This happened during a long running test of the player (About a month), is it mandatory to reboot regularly?

・The player application after 2R 202 seems to be able to log the state of threads. Are there any plans to consider the function to perform some actions based on this?

I am thinking of various ways to detect hangs that don’t appear in the status log like this, such as CPU time check, but I am in the middle of trial and error due to my lack of technical ability.

I would appreciate it if you could comment.

Hello.

I put this issue on hold, but it’s always in the back of my mind. I hope the topic will not be closed.
For verification, I would like to have the client freeze intentionally, what kind of method is there?

As mentioned above, the operating system was normal.
I want to freeze only the application.
By the way, I also have experience with C #, PHP, etc. used for xibo, so there is no problem with source level work.

I want to contribute to the stable operation of Xibo, but I don’t know how to verify it.

The Xibo Community site uses cookies. What are cookies?