We have a micro services architecture deployed on AWS ECS with Docker. The API Gateway and Eureka is from Spring. We have Java/NodeJS micro services which are working perfectly through the gateway but the .NET micro service, which uses SteelToe Eureka client gives us connection reset exception.
The following are screenshots for a better understanding.
Eureka app registration details
Feign client stacktrace
We have spent a lot of time, trying to figure it out, but at the moment with no success. Can we get some help on this to solve it please.
Thank you very much!
A lot of time has passed since the question, but my solution maybe will help someone. I faced java.net.SocketException: Connection reset during calling FeignClient. To fix it i change used OpenJDK to Oracle JDK and exception was gone! So, i really don't know root of the problem, but changing jdk works for me.
Related
I've been installing our very own ArcGIS Enterprise instance on AWS.
The instance I chose is ArcGIS Enterprise on Ubuntu.
It is important to mention that this installation was conducted without using Cloudbuilder. I know it is a tool that automates the process but I was introduced to it only after I have already started to attack my current instance problems head-on. So, please don't advise me to restart the whole process from scratch using it.
The current status of my instance is that my ArcGIS Server is working. I can access it, upload services and we have already started using it in out Staging environment.
I have authorized all of the software on the server and verified it is licensed. The Portal for ArcGIS is my main problem.
Whenever I try to access it externally(from my office computer) it seems to redirect to the internal IP for some reason, and then times out on that request.
for example typing(from my browser):
https://[dns address]:7443/arcgis/home
redirects to:
https://[internal IP]:7443/arcgis/home
and this times out. (...took too long to respond error)
The funny thing is I can access the portaladmin area.
it's only the portal itself which doesn't work.
Also, another curious thing is that if I type without using the ports, I can access a window but exceptions are thrown in the browser.
For example:
https://[dns address]/arcgis
This will lead to a window where the ArcGIS world icon can be seen but nothing else loads and there are exceptions for "resource not found" 404 on some of the components of this page.
Any ideas? What further information should I include to answer this question?
I've looked everywhere but Esri's documentation is not very forthcoming with examples and information to understand what it is I did wrong.
Also, I don't think this is a ArcGIS software issue. It looks like this might be a proxy issue. Anyone else experienced something like this?
Thanks!
I found the solution.
It was a combination of two problems:
Tomcat that was running the web adaptor service was crashing because of an entirely different and unrelated issue.
The Portal was missing a web adaptor configuration and therefore did not have the WebContext property set with the web adaptor URL.
After fixing both of these problems, I was able to access the portal correctly.
The presence insight server on bluemix has been quite unstable now. Cannot get access to the server.
Is there any way to deploy the instance on softlayer server for production?
There were definitely some problems with PI last evening, but the team worked until early this morning to get them resolved. It looks like the system is back up and functioning. Are you still seeing the issue?
Also, as a general reference, this page has service status details on it that may be of help if you notice a problem.
This issue came after #EricLaw explained me what was the problem when I could not receive a response back from my REST request.
Essentially the request was just returning blank, then I found out that if Fiddler was not running,
the application works just fine.
As Eric explains there are some security issues that I should consider when developing/debugging in Windows Server 2012.
I followed the instructions and installed the suggested utility but when I click in the AppContainer Loopback utility I get the following error: Failed to get AppContainer info: Unable to enumerate AppContainer, Is the windows Firewall Service started?
Of course, my Firewall Service is working just fine.
Additionally there is something I still don't understand, why my HTTP request works just fine whenever Fiddler is not running anyway?
Has anyone had this problem already?
PS: we need a new tag called fiddler4
I know it is a bit late to bring this up, but since I had a fix I though it might be helpful to others, too.
I used the utility described here and although it did not fix the issue for me, it had a glowing hint, the Windows Firewall.
What fixed the issue for me was starting Windows Firewall service, which I had disabled previously.
Hope this might help.
I have started working with ZK recently. I really like the framework but I am having a lot of problems in configuring it to work correctly with Websphere Application Server v8.0. Namely, I am developing an EJB application with accessing the database via JPA entities and using the ZK framework for GUI. I am developing the application using the Rational Software Architect 8.5.
When I start working on a fresh ZK project in a fresh workspace and server profile everything seems to be working correctly. However, after some time I start having strange problems. I think the problem occurs in the moment when ZK tries to initialize my ViewModel class. I do not get any exceptions or errors, but get the timeout message in my browser: "The page or component you request is no longer available. This is normally caused by timeout, opening too many Web pages, or rebooting
the server."
My ViewModel initializes by resolving the EJB bean I am using for business logic. That normally works, but as I said, after working for some time on the project, or after restarting my computer the ViewModel object cannot get initialized and the timeout error appears no matter what I do (when ZK tries to bind a variable or when I fire an event needed to be handled by my ViewModel. As before, I do not get any additional errors or exceptions.
Has anybody already had this problem? Is there a solution? Can I change the configuration of my server or ZK to avoid this problem?
Thank you.
Seems that I have solved the problem myself. I was using internal Rational Software Architect (Eclipse) browser which apparently has problems handling ZK sessions. When I switched to Firefox, the problems I explained stopped appearing.
I'm totally a newbie in Java Enterprise and I have a lot stuff yet to learn. Right now I'm working which involves JMS using JBoss and ActiveMQ. An application sends messages to queues in JBoss and my goal is to access those messages using any message broker (in this case I tried to use ActiveMQ). So I think it's better for me to embed ActiveMQ to JBoss. I used the link below as the guidance :
http://activemq.apache.org/integrating-apache-activemq-with-jboss.html
I followed every single steps in the guide except that I used JBoss 4.2.3 and ActiveMQ 5.1.1. If this problem results from the different version of the tools, I think I can't help it out because other versions seem does not work on my machine. When I run JBoss to test whether the embedding is working or not, I can't see anything running on port localhost:61616 which is the default port for ActiveMQ, although JBoss seems run well. My question is :
Is any one know how to fix this? Or has anyone ever experienced such problem?
Is there any way to access queue in Jboss?
If this question is pretty ambiguous or need more details, let me know. Thanks in advance for any help.
I'm quite sure the embedded broker is started (transport vm://localhost) which results in a working JMS broker for that VM. But this broker isn't reachable from outside (via tcp or anything else)
Do you see any error in the log when launching jboss? I managed to make it work with Jboss 4.2 and ActiveMQ 5.4, and once integrated it only gives you a small hint that is working correctly with some info message saying "broker started in port ..." .
Also you can try to use the sample producers and receivers in ActiveMQ to test if the queue is working correctly.
If you still have problems I can try to help