We are seeing intermittent (not all jobs fail) Backend errors when attempting to create a table from a JSON file in GS from the Browser UI.
There is only a red pop up with the "Backend Error" message, no other information and it goes away after a few seconds.
How can we troubleshoot or gather additional information about this error?
Related
All of a sudden we are seeing this random error / exception in our web application.
Failed to load resource: the server responded with a status of 502 (Bad Gateway).
In the Log Stream, we are seeing the following details, with specific error code as 502.3 - Bad Gateway: Forwarder Connection Error (ARR).
Also, sometimes in the browser itself we see "The CGI application did not return a valid set of HTTP errors." getting displayed.
Most of the searches for these error codes refer to "IIS / Proxy Server" configuration. But, we haven't changed any such settings.
The error happens very randomly and not specific to any user action/function. Same functionality works first and on second execution immediately after first one throws this error.
How to figure out what is causing this and how to fix?
I google this question, because the program was normal at the beginning, and the subsequent 502.3 error. After I checked the information on the Internet, I feel that it can only give us an inspiration, and it cannot solve your problem immediately.
So my suggestion is that first you browse post1 and post2 I provided.
Next, proceed to Troubleshoot according to the steps of the official documentation. Specific errors require specific analysis.
I'm trying to make my first load adf test. I have chosen software from Oracle "Oracle Application testing suite". But after probably first good test everything that I'm trying to test receives an error response even empty scenario ( of course from open script).
I tried lots of combinations in settings in OpenScript and all of the requests in this app are passed. We can find errors only in Oracle Load tests app. I don't have any idea what should I do. Can you give me any advice?
Every request responds showing this error:
Request error : authentication failed empJAgent.exe
I'm writing a Python CGI script and trying to test the behaviour of the system when I need to return Status: 500 Internal Server Error.
My script is something like that:
#!/usr/bin/env python3
print("Content-type: text/html")
print("Status: 500 Internal Server Error")
print()
When I run this script there is a report in apache access log with code 500, but it's not reported in the error log. I also don't get a "500 page" in the browser.
If an internal error is caused by some other means (e.g., a script that is not executable, or contains bad HTTP header) I do get the "normal" behaviour of internal server error.
It seems like apache is ignoring, somehow, the status returned from (my) CGI scripts. I've searched for an answer but couldn't find anything.
Just for clarity, CGI is working fine on this server in any other aspect.
Any thoughts? Am I missing something?
Thansk,
Amit
Answering to myself: it seems that I was barking up the wrong tree. Based on some clues and more empirical results, It seems that when passing a request to an external script (e.g. a cgi script, php etc) the apache server expects the external script to handle any error, and it's the responsibility of the external script's to return a document that includes the error code and an error message. The external script is also responsible to log the error (it's usually enough to print it to the standard error, and it'll be picked by apache and be written to its error log).
So, for example, if my cgi script needs to report an "Internal Server Error" it is not enough to return just the header (see in my question), but it should create and return the whole error message, in HTML format. In addition, it should print an error message to the standard error.
I haven't found an official source for that, but perhaps I somehow overlooked it.
I currently registered my Quecklink GV500 to cumulocity and I'm able to receive some events and measurements.
But when I try to send command to my Quecklink GV500 registered in Cumulocity but I always hava a FAILED response. For example, I tried to send this command (which is fully supported by the GV500) from SHELL tab: AT+GTTMA=gv500,+,1,0,0,,,,,,FFFF$
And as result I got:
Failure reason: Command currently not supported
I also tried to get the agent logs by using "Log file request" in the "Log" tab of my Device and as result I got:
Failure reason: Cannot build command. Search parameters only allow the
following characters [a-zA-Z0-9_]
Is it normal?
When I look the general information in "Info" tab I have:
Send connection: online
Push connection: inactive
Is it normal that Push connection is marked as inactive?
The tracker-agent in it current state does not use a push connection for receiving operations but does a polling of the operations. Therefore the push connection is always shown as inactive.
If you receive "Failure reason: Command currently not supported" it is an error from the agent not the device. The agent seems not to support shell operations for Queclink.
As for the error on the log file request it seems that there was an unsupported character in the search parameter. Maybe you can share what you entered for the parameters in the UI
Thanks for your answer. For the log file request I let blank value in the search input field. If I try to enter "gl200", I get the following error: Command currently not supported.
So to resume can you confirm that Quecklink devices can't be managed from Cumulocity for the moment? It's supported for which device?
I am trying to do multiple file upload simultaneously to google big-query using command line tool. I got following error :
BigQuery error in load operation: Could not connect with BigQuery server.
Http response status: 503
Http response content:
Service Unavailable
Any way to workaround this problem ?
How do I upload multiple files simultaneously to google big-query using command line tool.
Multiple file upload should work (and we use it every day). If you're getting a 503, that indicates something is wrong with the service. One thing you might want to make sure of is that if you're using a * in your command line that you have it quoted so that the shell doesn't expand it automatically before it gets passed to bq.
If you're getting a 503 error, can you retry the command the flag --apilog=- (this needs to be one of the first params) which will dump the interaction with the server to stdout. The problem may be obvious from that log, but if it isn't can you update your question with the relevant portions of the log? If you're not comfortable posting that information on a public forum, can you e-mail it to me at tigani at google dot com?