Project

General

Profile

Bug #20329

Issue du jour #1 at protoDUNE, 17-Jul-2018: including the dispatcher prevents data from flowing

Added by Kurt Biery about 1 year ago. Updated about 1 year ago.

Status:
Resolved
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
Start date:
07/17/2018
Due date:
% Done:

0%

Estimated time:
Experiment:
-
Co-Assignees:
Duration:

Description

In a system with 2 data loggers and a dispatcher in partition 1, the presence of the Dispatcher prevents data from flowing.

History

#1 Updated by Kurt Biery about 1 year ago

Turning off the RootNetOutput module in one of the DataLoggers allows data to flow through it, but the other DL does not allow data through.

WIth a 0.7 Hz trigger rate, I'm seeing half that at in the DIM DL plot. Then it went to zero. One minute at 0.7, one minute at 0.35, then zero.

#2 Updated by Kurt Biery about 1 year ago

That was run 2624.

#3 Updated by Kurt Biery about 1 year ago

  • Status changed from New to Resolved

this was traced to the fact that the TCPSocketTransfer inside the DataReceiverManager in the Dispatcher was only waiting for 10 seconds before giving up. Because the start of dataflow in Partition 1 was contingent on the operator pressing a special Run Control button, the delay between when the Dispatcher was told to begin the run and when the data started flowing was longer than 10 seconds.
Part of the reason for this is that the art modules inside the DL art processes don't make the connection to the Dispatcher until events start to arrive.
Following the lead of a change that Eric made on the develop branch, I commented out the line in TCPSocketTransfer that was giving up aftet 10 seconds, so now we effectively wait forever.



Also available in: Atom PDF