You marked this 'regression' but do not mention which version was known to work.

Please provide the information requested here: https://wiki.jenkins-ci.org/display/JENKINS/How+to+report+an+issue

Is this really about console-tail plugin, or did you just not read the note that says issues with Jenkins itself should use the 'core' component?

Sorry for unclear information, updated components and added previous version where I have seen this working.

Does the spinner (or "indeterminate progress indicator") at the bottom of the log disappear as well? Does the browser stop sending ajax update requests?

IIRC X-More-Data only applies when there's more data, so if logging is quiet, it'll not have more data for a while. The client should still continue polling for changes.

The spinner disappears and the browser stops polling for changes.

There is a javascript function fetchNext(...) in the console page which does polling. The code is pretty straightforward, it doesn't re-instantiate setTimeout when X-More-Data header is missing in the response and it stops polling (I think) forever:

    if (rsp.getResponseHeader("X-More-Data") == "true")
        setTimeout(function() {
            fetchNext(e, href);
        }, 1000);
    else
        $("spinner").style.display = "none";

Stapler sets X-More-Data if the file being read hasn't been completed yet though, so I was simply wrong about the behavior of this header:
https://github.com/stapler/stapler/blob/master/core/src/main/java/org/kohsuke/stapler/framework/io/LargeText.java#L266

The Jenkins parts also look reasonable:
https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/Run.java#L1439
https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/Run.java#L501
https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/Run.java#L234

Does this occur for all jobs? Are you doing weird things in Groovy scripts?

It happens for all job types, I have pure Java Maven jobs (e.g. "mvn clean compile test") and some Grails jobs ("grails prod war"). No weird things I hope, I'll try to attach screencast.

Screencast demonstrating bug, currently running job was Maven test goal.

The screencast is very nice, thank you. And the behavior demonstrated is very weird.

Does this happen for all jobs, all builds?

Is Jenkins accessed through a reverse proxy?

I'm having this issue as well. Has anyone made any progress toward finding a fix for this issue?

I have the same problem but it does not seem to be server related... in IE update is working as expected but not in FF. So it seems to be an issue on the client side.

In my case it doesn't work in all browsers (tested on Chrome 38, FF 33, IE 11).

Polling stops when the job doesn't write to console for about a half seconds or so. When the job is very verbose, it seems to work.

@Daniel:
Yes, it happens for all jobs that are not so verbose.
Yes, Jenkins is accessed through a reverse proxy, which is the Apache server in front of Tomcat on the same machine.

Is the X-More-Data header set for all responses?

When polling, typically for couple of seconds from refresh, there are X-More-Data and X-Text-Size headers present in all responses to 'progressiveHtml' requests. Responses look like this (only Content-Length, Date, X-ConsoleAnnotator and X-Text-Size headers differs between requests):

Cache-Control:max-age=0, no-cache
Connection:Keep-Alive
Content-Encoding:gzip
Content-Length:151
Content-Type:text/html;charset=UTF-8
Date:Thu, 30 Oct 2014 12:03:39 GMT
Keep-Alive:timeout=5, max=91
Server:Apache
Vary:Accept-Encoding
X-ConsoleAnnotator:7MYs6ttMMoModS4EVo7TUCG79S3ndnWRJpG1rPfajEK/KFxnFhMBC/D3zeM0xiuEYJDHZ4w/GiQAwBpuiCiRbTshPhZm+Jjau4J0XfzamZptCfN++vXbE0qLVLZbcpyl
X-Frame-Options:SAMEORIGIN
X-Mod-Pagespeed:1.7.30.4-
X-More-Data:true
X-Text-Size:2625

The last response, after which the page stops polling, looks like this:

Cache-Control:max-age=0, no-cache
Connection:Keep-Alive
Content-Length:0
Content-Type:text/html;charset=UTF-8
Date:Thu, 30 Oct 2014 12:03:41 GMT
Keep-Alive:timeout=5, max=89
Server:Apache
X-Frame-Options:SAMEORIGIN

I must admin that this is frustrating for me and colleagues. Every time refreshing when some process is computing for while can get for example me to madness.

The last response, after which the page stops polling, looks like this:

What's the HTTP response code for this response?

A colleague of mine installed a new server with Jenkins in the same version (1.582) from scratch and the console is polling OK now. Sorry, I cannot reproduce it, apparently it happens under certain conditions only.

Martin: Clearly, otherwise I'd be able to reproduce the problem

Would be interesting to know whether all affected instances run behind Apache.

HTTP requests and corresponding responses

Same here with fresh Bitnami stacks (1.575 and 1.585) and updated (currently 1.580.2), running behing Apache (2.4.10) + Tomcat (7.0.55) on Ubuntu 14.04.1 64bit.

See attachement for the 3 last requests and corresponding responses.
Let me know if you need more info.

Nothing in the logs.
Here are the logs that are enabled by the way:

The interesting line is the one without logger name, just log level.

Thanks for your quick answer.

I guess I won't be really helpful because I did not have a default log level (it seems it is removed on startup), and setting it to ALL does not produce any more info in the logs.

Is snyone experiencing this issue...

• Without Apache in front of Jenkins?
• Without Tomcat running Jenkins?
• On a box that's not Bitnami?

Cannot reproduce the problem on the VMware VM bitnami-jenkins-1.596-0-ubuntu-14.04 I just downloaded. Tested with freestyle project executing the following build script:

for I in $( seq 0 1000 ) ; do
  sleep 1
  echo JENKINS-25026 $I
done

The console output keeps updating.

Neither Chrome 39 nor Firefox 34.0.5 on OS X show the reported issue when accessing the mentioned instance.

Any console annotators enabled, like Timestamper, or Ant plugin or Gradle plugin adding a list of executed targets/tasks?

When using a freestyle project executing the script you gave (without any console annotators), it seems to works. But putting the sleep value at 2 fails.
I have a bunch of console annotators installed though, I will try to check that if you still can't reproduce this issue with a greater sleep value.

Reproduced with sleep 3. Investigating...

And that's why you don't use some random vendor's preconfigured shit. It's broken.

Disable mod_pagespeed (whatever the hell that is) in Apache and things start working.

What a waste of time.

Thank you very much for your help. It's always good to know that this configuration is broken.