RedHawk SDR waveform won't run in Domain Manager - redhawksdr

In the RedHawk SDR, I am trying to launch a waveform under REDHAWK_DEV as described in chapter 22 of the Redhawk Manual Version 1.10. The waveform (junk2) launches fine into the sandbox, but fails when I try to launch into the Domain. The error is
Failed to create application: junk2_097_121512019 'load' failed for
component: 'HardLimit' with component id:
'HardLimit_1:junk2_097_121512019_1' with implementation id: 'cpp'; on
device id: 'DCE:9bca8a73-55e6-443f-9925-081db1fcad5b' in waveform
'junk2_097_121512019_1' error occurred near line:2386 in
file:ApplicationFactory_impl.cpp;
IDL:CF/ApplicationFactory/CreateApplicationError:1.0
Looking at the Console, the following error was reported
2015-04-07 14:09:30 INFO DeviceManager_impl:1626 - Registering device
GPP_crancentos1 on Device Manager DevMgr_crancentos1
2015-04-07 14:09:30 INFO DeviceManager_impl:1665 - Initializing device GPP_crancentos1 on Device Manager DevMgr_crancentos1
2015-04-07 14:09:30 INFO DeviceManager_impl:1840 - Registering device GPP_crancentos1 on Domain Manager
2015-04-07 14:10:21,782 ERROR DCE:9bca8a73-55e6-443f-9925-081db1fcad5b{1}:625 - [Errno 13] Permission denied: 'components'
Traceback (most recent call last):
File "/usr/local/redhawk/core/lib/python/ossie/device.py", line 594, in load
os.mkdir(loadPoint)
OSError: [Errno 13] Permission denied: 'components'
Basic things I have checked:
Confirmed Domain Manager (REDHAWK_DEV) and GPP_crancentos ("crancentos" is the machine name) are running.
Changed the device from "AUTO" to "GPP_crancentos1" on the "Assign Components to Devices" tab
Tried a different waveform. The first component in each waveform throws the same error, even if the component is different. The components are all from the standard group provided in the basic install.
Since the Console implied there was a file access problem, I opened permissions to universal read/write/"do what you want" for /Var/redhawk/sdr/dom/components, HardLimit and cpp subdirectories, and files for the two subdirectories (including the cpp, java, and python diretories).
The RedHawk version is 10.2R2015013000031 and the operating system is CentOS 2.6.32-504.e16X86-64.
It confuses and confounds me that it runs in the sandbox (even before I opened the permissions), which implies the problem is not in the waveform itself, but not under the Domain Manager. I am sure that I am one setting or line of code away from success... but don't know what it is.
Help would be much appreciated!
Brad Meyer
AECOM

When a component is launched by the Domain Manager it is "loaded" onto the executable device, the GPP, and then executed. The loading consists of copying the component to the running the GPP into a local cache which is located at $SDRROOT/dev in a hidden directory for each device manager. In this case it appears the user running the Domain Manager or the Device Manager running the GPP don't have permission to write to that directory. One possible explanation might be that the user is not in the redhawk group. When the component is run in the sandbox it does not execute it on the GPP device so this step (and this error) will not occur.

Related

Linux: How to setup the environment for React Native correctly?

Sorry for bothering you with my problems, but it has become a little bit weird ...
Meanwhile, it looks as if my problems were not really caused by this "Android virtual device (unavailable)" I decribed in my previous post ...
In the meantime, I'm relatively sure that all my problems were caused by the user rights ...
By mounting my partition in /etc/fstab by Device ID with the option values
gid=users,fmask=133,dmask=022,umask=007,locale=en_US_UTF8
the message "Permission denied" from yesterday disappeared (see below) after executing
npx react-native start
But when I tried to start my app using
npx --debug react-native run-android
this occurs
* What went wrong:
Execution failed for task ':react-native-gradle-plugin:processResources'.
> Could not copy META-INF to '/windows/D/02 Web/02 Javascript/AwesomeProject/node_modules/react-native-gradle-plugin/build/resources/main/META-INF'.
> Could not set file mode 755 on '/windows/D/02 Web/02 Javascript/AwesomeProject/node_modules/react-native-gradle-plugin/build/resources/main/META-INF'.
It says that the file mode cannot be set ... there must be another problem with my option values. Any idea?
All these problems appeared after buying a new computer ... on the new computer, this development partition is located on a primary partition of a SSD, on my old pc, it was an extended partition on a HDD ... but that cannot be the cause? My development partition is an NTFS partition, could that be the cause?
I created a React native project in my user home directory, there were no such access problems?
Any idea what's wrong with the user rights I indicated? Or maybe the best way is to format my development partition to BTFS?
Thanks a lot in advance for your efforts, and kind regard, Frank
=====
I'm trying to learn React Native on my new computer using a linux system (Open Suse Leap 15.4).
I think the problems mentioned below are caused by the user rights.
I created an NTFS partition and mounted it in /etc/fstab by UUID with the option values
users,uid=my_user_name,gid=users,fmask=133,dmask=022,umask=007
Those rights are different from those below, and I got a different error message
...#localhost:/windows/D/02 Web/02 Javascript/AwesomeProject> npx react-native start
sh: /windows/D/02 Web/02 Javascript/AwesomeProject/node_modules/.bin/react-native: Permission denied
The rights for my development partition must be wrong ... but what can I do to make this work? Any ideas?
Best wishes, Frank
=======
I'm trying to set up my React Native development environment as described
[here at https://reactnative.dev/][1]
on a Linux OS (Suse Leap 15.4).

Can't access to configuration files at session start on Gnome

I have troubles accessing to my gnome session.
I work on a Linux machine with red hat & gnome, and when I want to open my session, I get the error :
"An error occurred while loading or saving configuration information
for gnome-session. Some of your configuration settings may not work
properly.
Failed to contact configuration server; some possible causes are that
you need to enable TCP/IP networking for ORBit, or you have stale NFS
locks due to a system crash. See http://www.gnome.org/projects/gconf/
for information. IOR file '/tmp/gconfd-lepolmartin/lock/ior' not
opened successfully, no gconfd located: Permission denied"
I have checked, there is no folder gconfd-lepolmartin in /tmp. For some reason the files that Gnome are searching for do not exist in my session, but they do for other users of the same network that connect on the same machine.
Any idea ?
SLP

Using JavaFX in a headless Raspbian server

Based on this tutorial, I would like to implement a media player website running in a Raspberry Pi connected to an old (but excellent) amplifier/home theather. This means that a servlet (a thread created by a servlet) will play the music and I will control it via Internet.
What I already did and is working? I removed all graphic references (button, panel, etc) from the example and put the resulting code in a thread. When I try to run this on a Tomcat8 in Ubuntu Desktop, everything works fine! But when I try to run it in the Raspberry without GUI, I got the following error:
Exception in thread "Thread-0" java.lang.UnsupportedOperationException: Unable to open DISPLAY
I know where is the error, but I don't know how to fix it. The problem is in the line below:
com.sun.javafx.application.PlatformImpl.startup(()->{});
Without the line above, I got an error stating that to Toolkit was not initialized.
So,
(1) Is it possible to run JavaFX in a servlet?
(2) How to solve this problem?
Thanks!
Alex

RedHawkSDR - USRP N210 - Cannot run USRP_UHD Component

I imported this github c++ project: https://github.com/RedhawkSDR/USRP_UHD into redhawk, hoping to run it with a USRP N210.
Redhawk only allows me to run the project as a component or C++ Application, so I tried running it as a component.
Here is the exact error I get when I try to run as a component:
An internal error occurred during: "Launching USRP_UHD".
Could not initialize class gov.redhawk.ide.debug.internal.ScaDebugInstance
How can I fix this?
The USRP_UHD device is a Redhawk Device that interfaces with the N210. In Redhawk, Devices are deployed and managed by an instance of the Device Manager, which is referred to as a Node.
To run the USRP_UHD Redhawk Device in a Domain:
Install the USRP_UHD Device to the Target SDR. This can be accomplished by clicking and dragging the top level folder of the USRP_UHD project from the Project Explorer view to the Target SDR in the SCA Explorer view.
Create a new SCA Node Project using the Redhawk IDE that contains a USRP_UHD Device instance. The first wizard page will prompt you for a Node name (project name) and a Domain name. You can override the domain name later at run time if the name you choose now ends up being different from your running Domain. After clicking Next, the second and final wizard page allows you to choose from a list of Devices that are installed in your Target SDR. Select the USRP_UHD and click Finish. The Overview tab of the SCA Node Editor will appear after clicking Finish.
Configure the Node. Within the SCA Node Editor, you can edit the properties of the USRP_UHD Device using either the Devices tab or the Diagram tab. Generally, you will want to at least configure the IP address of the N210 using the USRP_ip_address property of the USRP_UHD Device so that the USRP_UHD Device will connect to the USRP hardware upon deployment.
Install the Node to the Target SDR. Again, this can be accomplished by clicking and dragging the top level folder of the Node from the Project Explorer view to the Target SDR in the SCA Explorer view.
Launch a Domain and the Node (Device Manager) that you created containing the USRP_UHD Device. This can be done by right-clicking on the Target SDR in the SCA Explorer view and selecting Launch…. In the dialog box that pops up, you can choose a Domain Name (this doesn’t have to be the same as the Domain Name specified in the Node) and a debug level for the Domain Manager. To also launch a Node, choose the Node that you created from the list of Device Managers and set the debug level appropriately for the Device Manager. Select OK to launch both.
Inspect the Domain that you launched by expanding the Domain within the SCA Explorer view. You should see the Node under the Device Managers folder, and after expanding the Node you should see the USRP_UHD Device instance (probably named USRP_UHD_1).
If this doesn't fix the issue, please provide some more information about your environment (specifically, what version of the Redhawk framework and IDE, what version of Java is reported by "java -version", what OS and version, what branch/release of USRP_UHD, what version of UHD software) and the steps you are taking to run the USRP_UHD as a component. In Redhawk version 1.9, I was able to choose Run as…->Local Component Program and it successfully launched the USRP_UHD Device in the Sandbox without the error you have experienced. You may also wish to try with both Redhawk version 1.8 and 1.9 (be sure to use the latest release of each) to see if the issue appears in both versions.
Is the version of UHD that you are using version 3.7.1? It should be reported by any of the uhd_* commands, such as uhd_find_devices.
uhd_find_devices should find your N210 device if you can ping it. I have seen the X310 not respond to uhd_find_devices, but it does respond once the IP address is specified. Try specifying the IP address of the N210 as shown below:
uhd_find_devices --args="addr=192.168.10.2"
Replace 192.168.10.2 (the default IP address for the N210) with the IP address of your N210, of course. If your N210 is still not found, try unplugging the power source to the N210 and then plugging it back in to force a reboot. Again, I’ve seen this help with the X310 when it wouldn’t respond to the uhd_find_devices command even with the IP address specified.
Then also try to probe the N210 using the following command:
uhd_usrp_probe --args="addr=192.168.10.2"
I believe that if the N210 has a firmware version that is incompatible with the version of UHD you have, the N210 will still be found and the probe command will inform you that the firmware must be updated.
If neither command can communicate with the N210, I have to think the issue is between the UHD software and the N210 instead of being a Redhawk-related issue. To load the firmware, see the link below. Also, there are instructions for setting up networking and troubleshooting communication problems at the same link. If you haven’t done so already, take a quick look and see if anything there helps. Let me know what you find.
Load the Images onto the On-board Flash (USRP-N Series only)
Did you follow the networking setup steps listed here: USRP N210 Networking Setup?
As described in the link you need to make sure your host PC has an IP address on the same subnet as the USRP. You can use ifconfig to set the static IP address for a specific interface ex. "eth0"
Java 7 update 55 (version 1.7.0_55) introduced an issue/bug with Eclipse (including the Redhawk IDE, since it’s Eclipse-based) that is described here.
The change made in Java 7 update 55 that seems to have caused various issues with Eclipse/JacORB has been reverted here and will be available in Java 8 update 22. There is a beta release available here, but being a beta release, it may have other issues and as such may not be worth trying. Instead, you can do what I did and patch the Redhawk IDE as a workaround to the bug in Java 7 update 55.
Set environment variable IDE_HOME to refer to the same directory as the eclipse executable (not the executable itself, though).
export IDE_HOME=/usr/local/redhawk/ide/R.1.9 # replace with your path
Append the following line to the $IDE_HOME/eclipse.ini file. If a line already specifies the endorsed dirs, replace it with this line.
-Djava.endorsed.dirs=$IDE_HOME/jacorb/lib
Create the JacORB lib directory at the path specified in the previous step.
mkdir -p $IDE_HOME/jacorb/lib
Find the exact name of the JacORB directory located within $IDE_HOME/plugins, which will begin with “org.jacorb.system”, and assign it to an environment variable named JACORB_DIR:
export JACORB_DIR=`find $IDE_HOME/plugins/ -maxdepth 1 -name org.jacorb.system*`
Copy the contents of the JacORB jars directory into the $IDE_HOME/jacorb/lib directory:
cp -R $JACORB_DIR/jars/* $IDE_HOME/jacorb/lib/.
This should resolve any potential issues resulting from the Eclipse/JacORB bug. Does this also fix the remaining issues you’re having with the USRP?

ClearCase: Unable to create directory when trying use 'startview' on Linux

I am really stuck in ClearCase startview on Linux. I tried the below:
ct startview <viewtag>
And met an issue like below:
cleartool: Error: Unable to create directory "": error detected by ClearCase subsystem.
I am sure the view is OK, and I can start it in other host. So perhaps my OS met some issue. Is there some good solution?
And I have restarted my Linux system.
If this is about a dynamic view, you can check technote swg21449500 "Create and startview errors: Cannot start dynamic view and Unable to create directory":
Cause
The client or view-server installation is corrupted. The symptoms would typically be seen immediately after applying a Fix Pack of iFix.
Diagnosing the problem
A cleartool -verall showing a mismatch between the MVFS and the other modules is a clear indication of corruption.
Resolving the problem
An uninstall and re-install on the machine will be necessary to resolve the issue.
On Windows, you can see a similar error when a Windows client attempts to start a dynamic view on a Linux or UNIX server running Samba.
See swg21295394.

Resources