ngrep | node based asynchronous grep command | Runtime Evironment library

Looks you have incorrect namespace based on wsdl

To fix yours replace

Try this. No need to delegate to localhost. lookup is running always on the controller

This is now "solved".

Turns out the informative Failed to fetch, was really trying to say Request aborted by leaving page. If it just would've said that, plenty of head-scratching could've been avoided.

So yes, the issue was that the fetch() that's at the center of this Q was being called in response to clicking a submit button in a form which had been incorrectly set up to not reload the page upon submission. Thus, when the page was reloaded the pending requests were aborted, throwing Failed to fetch.

I'm using Svelte, which I'm not entirely familiar with, and disabling the page reload on submission is kind of non-obvious. At least if you're already familiar with the classic way of doing it (i.e. return false).

In addition to this mistake, I also fooled myself by having "Preserve log" enabled in the dev console, and thinking that the message "Navigated to ..." it outputs on page reload was because of my code later in the chain (which I was "getting to later"...).

You can indeed override a previously set CURLOPT_RESOLVE entry by setting a new one for the exact same host + port, or you can clear a previously set one by passing in a line like -[host]:[port] (ie starting with a dash and without a trailing :[address]).

Just setting the value to NULL means there's no new entry to pass in, which will make the curl handle keep the data already pre-populated since before.

This is documented in the CURLOPT_RESOLVE man page.

libcurl will try to reuse previously used connections as far as possible unless you tell it otherwise. If you do a second request to the same host name that you previously used, that connection will then be reused and the new CURLOPT_RESOLVE setup might then not be used or checked.

If you want to make it really sure that your second request actually uses the updated address, you must make sure that the second transfer doesn't reuse the connection still held alive from the first transfer. You can achieve this either by setting CURLOPT_FORBID_REUSE on the first connection or by setting CURLOPT_FRESH_CONNECT on the second.

Aliases are simple prefix substitutions that take place before syntax is parsed. This gives them powers other constructs don't have (albeit powers which are rarely needed or appropriate) -- you can alias something to content that's subsequently parsed as syntax -- but it also constrains them: Because a compound command needs to be parsed before it can be executed, the ngrep command is parsed before the source command is executed, so the alias is not yet loaded at the point in time when it would need to be to take effect.

As a simple demonstration (thanks to a comment by @chepner):

There is no reason in the information you have provided to explain why starting a server on port 9090 would have any impact on another server listening on port 9001. Your description of the problem is also innacurate and misleading. The webserver appears to be unaffected and is still responding. It's php-fpm which is no longer responding.

Assuming that you've not built some hidden dependency into the system you've not told us about, then you should start by looking in your logs, and the output of netstat. Would it have been so hard to test what happened with content served directly from nginx rather than php generated content after starting the websockets server?

Problem was Tomcat and a useless semicolon in the header Content-Type. Tomcat wasn't able to read the content type with the semicolon with nothing following like charset. So when the request was read, server side, the default content type text/plain. RestAssured

So i changed tomcat by undertow and i don't use rest assured anymore but WebTestClient. See codes and pom.xml. Source of this problem wasn't expected.

after several tests, we realized that this error is only on Android platforms (we tested the APIs with Postman when Android platforms were receiving Too many follow up request.

so after some more tests, we found out, when Android platforms get 408 as a response code they show this message. we changed the status code and this problem fixed.

I put this issue on square/okhttp repository in github and they put the bug flag on it.

you can see it here

---- update ----

they said it fixed in OkHttp v3.10