Hi Nino,
Your medm screen shot shows that you are running ADEiger version 3.0.0. There have been 2 releases of ADEiger since then, and both have entries in the release notes documenting that they address the problem you are seeing:
Release Notes
=============
R3-2 (January 22, 2022)
----
* Improved the R3-1 fix for the race condition with the Stream interface.
In R3-1 the ZMQ socket was always created, whether or not DataSource was set to Stream.
This introduced a problem for users who want to use ADEiger to control the detector,
but use another ZMQ client to receive the data.
It caused the other client to only receive every other ZMQ message.
Now the ZMQ socket is only created when DataSource is set to Stream, and it is
deleted when DataSource is set to anything except Stream. This allows use of other ZMQ clients.
R3-1 (November 18, 2021)
----
* Fixed a race condition with the Stream interface.
- The detector sends the first message over the ZMQ socket as soon as the "arm" command is sent when starting acquisition.
- However, the driver was not creating the ZMQ socket to receive those messages until after the "arm" command was sent.
This was a race condition. It was failing on an Eiger2 at APS sector 34.
- This release creates the ZMQ socket before acquisition begins.
The problem was documented in this Github issue:
https://github.com/areaDetector/ADEiger/issues/47
It was fixed in these commits:
https://github.com/areaDetector/ADEiger/commit/690db9a13b4e58fbba231d4a6b7d004001223b25
https://github.com/areaDetector/ADEiger/commit/c6a69fabc1e4db5281999876faf4ea16c1d8d4b4
Mark
Has anyone see these errors with ADEiger like this?
eigerDetector:streamTask: failed to get header packet
eigerDetector:streamTask: got Strat packet, ignoring
See IOC and MEDM screenshots here:
It did not happen immediately after a fresh reboot of IOC, while it occurred after we took data for a while when the IOC gave some error message about the Stream Api. You can also
see in the last photo that the Steam state shows "disabled" even though it was set to enable. This happened no matter I set the header detail to "None" or "Basic". As this issue happened, the number on image counter didn't update while the default FileWriter
still wrote frames into hdf5 files.
(Ignore the autosave error).