Project

General

Profile

Bug #24591

New DPM won't read on event for certain erlang front-ends on the CMTF subnet

Added by Dennis Nicklaus about 1 month ago. Updated about 1 month ago.

Status:
Resolved
Priority:
Normal
Assignee:
Category:
Data Pool Manager
Target version:
-
Start date:
07/01/2020
Due date:
% Done:

0%

Estimated time:
Duration:

Description

We are unable to read devices on-event from p2itb1 or p2itb2 (all devices, including, e.g. the guaranteed readable, U:GRWDB9)
Devices read back fine for periodic, but when we try to read on-event, it just hangs forever.
Santucci reports that it does work when using old DPM.

On the erlang front-end, with it set to the default tclk multicast clock, I do see responses from
sync:on_event(2)
calls, so the front-end is receiving the TCLK multicast (also shown by network dumps).

History

#1 Updated by Dennis Nicklaus about 1 month ago

Just to be clear: DPM=Java does work on-event for (some, at least) other erlang front-ends, e.g. CLX38e/U:GRWD7C

#2 Updated by Richard Neswold about 1 month ago

This means the Java DPMs need to listen to all multicasts and that clients should indicate which clock should be used. We mentioned this in #20115.

#3 Updated by Richard Neswold about 1 month ago

Charlie reported (via email):

The problem seems to be that the both nodes had TCLK set to false on D98. I’ve updated both to true and my test DPM works. Unfortunately I need to restart the DPMs to get the update D98 entry. Hopefully I’ll get them all restarted by the end of day tomorrow. I’m planning on removing the reliance on the TCLK entry altogether hopefully very soon.

#4 Updated by Charles King about 1 month ago

  • Status changed from New to Resolved

The problem was that both nodes had TCLK set to false on D98. DPM currently relies on the status of the TCLK setting so any node that has TCLK support from any source should set this to TRUE. I’m planning on removing the reliance on the TCLK entry altogether in the near future.



Also available in: Atom PDF