Scheduled Task Not on Time Error

Support
1

Hi Chris,

I knew there was a similar discussion thread on this, but I got this error message even when I was not running any experiment task.

I just attached the screen-shot. After I connected to the server and before I loaded any experiment, I could see the error message on console. I am not sure how it happens. I am using the latest nightly built version of MWorks.

2

Attachment: Screen_Shot_2015-06-18_at_5.12.43_PM.jpg (969 KB)

3

Hi Chia-Jung,

This error is telling you that one of MWorks’ background threads has fallen behind. If you only see it once, before any experiment has been loaded, then there’s no need to worry about it. On the other hand, if you start seeing it regularly and/or while running experiments, then I would want to investigate further.

Cheers,
Chris

4

Hi Chris,

I start seeing it more regularly before experiment has been loaded, as well as while running experiments.

I did not see it when I ran the experiments last week. I started seeing it more regularly since yesterday, but I did not change anything in days between. Do you have any suggestions on how I can investigate this?

Thanks,
Chia-Jung

5

As a supplement,

I also saw this before I connected to the server (as attached screenshot shown) today. It did not happen yesterday.

I also started to see that the variables are not updated in real-time in my remote end. It runs normally in the computer that runs MWServer and MWClient, and the variables are updated normally on that computer. However, if I connect to the MWServer remotely to monitor the activity on that computer, the variables and eye window are frozen after a while, and I would need to disconnect the server and reconnect to the server, and the variables are updated again.

Thanks,
Chia-Jung

6

Attachments:

7

Hi Chris,

This error message keeps popping out. However, I also noticed one thing: Both computers (the one for server and the one connecting to the server) do not have a user account called mwdev. Not sure where that user folder came from.

Thanks,
Chia-Jung

8

Hi Chia-Jung,

“mwdev” is the name of the user account that compiled the MWorks build you’re using. The file path you see in the error messages doesn’t refer to anything on your computers.

Regarding this:

I also started to see that the variables are not updated in real-time in my remote end. It runs normally in the computer that runs MWServer and MWClient, and the variables are updated normally on that computer. However, if I connect to the MWServer remotely to monitor the activity on that computer, the variables and eye window are frozen after a while, and I would need to disconnect the server and reconnect to the server, and the variables are updated again.

In the case where the variables window stops updating, is there an MWClient instance running on the MWServer machine and a remote MWClient, or is there only the remote client? How long does it run normally before things freeze? To restart updates, do you need to quit MWClient, or does disconnecting from and reconnecting to the server suffice? Does the variables window ever unfreeze on its own?

Thanks,
Chris

9

Hi Chris,

In the case where the variables window stops updating, there is an MWClient instance running on the MWServer machine and a remote MWClient. The variables window for the instance running on the MWServer is still updating, but not the remote one.

It runs normally about 10 minutes before things freeze.

To restart updates, disconnecting from and reconnecting to the server suffices. No, the variables window never unfreezes on its own so far.

Thanks a lot,
Chia-Jung

10

Hi Chia-Jung,

I’ve been trying to reproduce this problem on some of my machines, but so far I’ve had no success. Would it be possible for me to come to your lab and see the client freeze in action? If nothing else, I may be able to determine whether or not it’s a network issue.

Any time this afternoon is fine for me. Otherwise, maybe we can find a time during the week of July 6, as I’m out on vacation next week.

Thanks,
Chris

11

Hi Chris,

I will have a subject coming at around 3 pm. Is it possible that you come
at around 3:30 pm, so you might be able to see the client freeze in action?
Great thanks!!

Best,
Chia-Jung

12

Sounds good. See you then.

Chris

13

Hi Chia-Jung,

Following up on this: The “WARNING: Session Failure on ID …” messages that we saw indicate that some kind of networking error has occurred, and because of this, one of MWorks’ I/O threads is shutting down. Presumably, the terminated thread is the one responsible for sending events from the server to the client, as that would explain why client-side variable updates stop.

In order to help debug this issue, I’ve added some code to log more details about the error when it occurs. When you have a chance, please upgrade to the latest nightly build. Then, the next time you see the error, please copy the associated warning messages and send them to me. Hopefully, this will provide some insight into why this is occurring.

Thanks,
Chris

14

Hi Chris,

Thanks a lot! I will upgrade to the latest version of MWorks, and let you know if I see the same error.

Thanks,
Chia-Jung