On Fri, 2019-03-29 at 12:06 -0700, Mark Cafaro wrote:
> It then eventually times out when my socket timeout is reached:
>
>
This clearly suggests that the server fails to send a response back to
the client, so this looks like a server side issue to me.
Oleg
>
> 2019-03-29 19:03:42,884 DEBUG
I've come up with two theories. In addition to this long running request, I
have another thread making short requests frequently (every second).
1) Perhaps Travis CI is identifying my program as a spam bot given that it's
making so many requests, and therefore blocking the response of my long
It then eventually times out when my socket timeout is reached:
2019-03-29 19:03:42,884 DEBUG [org.apache.http.wire] http-outgoing-0 << "[read]
I/O error: Read timed out"
2019-03-29 19:03:42,884 DEBUG
[org.apache.http.impl.conn.DefaultManagedHttpClientConnection] http-outgoing-0:
Close
The wire log didn't reveal much. It doesn't output anything after getting stuck
on httpclient.execute(). See below (I replaced sensitive names, IPs, and JSON
with generic info):
2019-03-29 18:31:41,815 DEBUG [com.company.ci.Client] Executing post...
2019-03-29 18:31:42,254 DEBUG
No exception. It just stays stuck on the httpclient.execute() line. I'll gather
the wire logs.
On Fri, 29 Mar 2019 08:44:02 -0700 Oleg Kalnichevski
wrote
On Fri, 2019-03-29 at 08:37 -0700, Mark Cafaro wrote:
> I mean to say when I'm using the Apache client, I see the server
On Fri, 2019-03-29 at 08:37 -0700, Mark Cafaro wrote:
> I mean to say when I'm using the Apache client, I see the server
> sends back a proper response but the client never receives it (it
> just stays stuck on httpclient.execute()). So, on the server side,
> irrespective of the execution path it
Hi Mark,
Your best bet is to post your HttpClient wire log (and probably server log).
Gary
On Fri, Mar 29, 2019 at 11:38 AM Mark Cafaro wrote:
> I mean to say when I'm using the Apache client, I see the server sends
> back a proper response but the client never receives it (it just stays
>
I mean to say when I'm using the Apache client, I see the server sends back a
proper response but the client never receives it (it just stays stuck on
httpclient.execute()). So, on the server side, irrespective of the execution
path it is taking, I ultimately know it is sending back a
On Fri, 2019-03-29 at 08:09 -0700, Mark Cafaro wrote:
> What makes me think it's a client side issue is that I can take the
> same request and run it with cURL and properly receive a response. I
> can also view the logs on the server side and see that a response was
> sent back.
>
>
So what?
What makes me think it's a client side issue is that I can take the same
request and run it with cURL and properly receive a response. I can also view
the logs on the server side and see that a response was sent back.
I'll try turning on wire logging and see if that gives me any clues.
On Fri, 2019-03-29 at 06:34 -0700, Mark Cafaro wrote:
> The call to httpclient.execute() is just stuck until the eventual
> socket timeout is reached.
>
>
What makes you think this is a client side issue? Have you tried
running your application with HttpClient wire logging turned out and
The call to httpclient.execute() is just stuck until the eventual socket
timeout is reached.
-Mark
P.S. Sorry if this created a new thread. I did not subscribe in time to get the
first reply.
> What exactly do you mean by "never receives a response"? What happens
> instead?
>
> Oleg
On Thu, 2019-03-28 at 17:28 -0700, Mark Cafaro wrote:
> Hi,
>
>
>
> I am running long HTTP post requests on Travis CI (Ubuntu 14.04)
> using the Apache HTTP Client. Strangely if a request exceeds about 8
> minutes, the client never receives a response. I can see on the
> server side, however,
13 matches
Mail list logo
