Supervisor: Tail -f option in the web interface doesnt work.

[ghost]

On the web page to control the processes, the Tail -f link doesn't work (the server doesn't respond anything, so the browser doesn't show anything). The href attribute of the link is
http://localhost:8080/logtail/recommender_server%3Arecommender_server_01
where recommender_server is the process name.

Nevertheless the Clear Log link works well. In this case, the href attribute of the link is:
http://localhost:8080/index.html?processname=recommender_server%3Arecommender_server_01&action=clearlog

My operating systems is: OS X Mountain Lion (Mac Book Pro)

; MY CONFIGURATION FILE FOR SUPERVISORD

[unix_http_server]
file=/Users/rmerinogarcia/Proyects/temp/supervisor.sock
username=user
password=123

[inet_http_server]
port=127.0.0.1:8080
username=user
password=123

[supervisord]
logfile=/Users/rmerinogarcia/Proyects/temp/supervisord.log
logfile_maxbytes=50MB
logfile_backups=10
loglevel=info
pidfile=/Users/rmerinogarcia/Proyects/temp/supervisord.pid

[rpcinterface:supervisor]
supervisor.rpcinterface_factory = supervisor.rpcinterface:make_main_rpcinterface

[program:recommender_server]
process_name=%(program_name)s_%(process_num)02d
numprocs=1
numprocs_start=1
command=/Users/rmerinogarcia/.Envs_Python_2.7.3/USG_ENGINE/bin/python /Users/rmerinogarcia/.Envs_Python_2.7.3/USG_ENGINE/bin/start_recommender.py -p 91%(process_num)02d
stdout_logfile=/Users/rmerinogarcia/Proyects/temp/recommender.out
redirect_stderr=true
autorestart=true

[dalanmiller]

Same thing happening to me here on Mac. @rmerinogarcia did you ever find a fix?

[netandreus]

+1 Same at Fedora 3.7.6-201.fc18.x86_64
in console works fine, but in web interface respond does not returned (timeout in browser).

[netandreus]

Here is the fix: https://github.com/Supervisor/supervisor/pull/195/files

[tobeportable]

After applying this fix : there is still an error when the log file exists but it is empty.

[mnaberez]

403 notes that it used to work on Chrome but is broken after Chrome/Chromium 33.

[corywalker]

I can confirm that it no longer works for me under Chrome 33.

[sw00]

Doesn't work on Chrome Version 34.0.1847.116 either.

[duxiaoyao]

Not work on Chrome Version 35.0.1916.114 either

[tlatorre-uchicago]

Also doesn't work on Firefox 24.6.0 (RHEL 6.0). It seems to be a problem with how these browsers deal with the chunked transfer protocol (see http://code.google.com/p/chromium/issues/detail?id=156023).

@mnaberez, it seems like relying on the chunked transfer protocol is not very stable. Do you think it's worth adding some javascript to the page to send AJAX requests to update the log files?

[jensenbox]

I would be happy to put something together for the the JS version of the tail. Is this something the project is open to?

[bfleming-ciena]

I'm having this same issue. I will mention I am running supervisord in a docker container and I've tried chrome and firefox. Stop/start works but not the logtail tail -f

[ngocson2vn]

I just proposed my fix #471

[dawncold]

It works fine on Chrome/Chromium Version 36.0.1985.125 without any modification

[armab]

The issue is still there.
Doesn't works with Firefox 25

[Shellbye]

I have the same issue but fixed by add childlogdir = %(here)s to [supervisord]. And all my [program] have the directory=%(here)s

[ue-it]

It works well with http://127.0.0.1:9001/logtail/test, but not work after nginx with http://my.url/logtail/test
All the others work well, so I think the problem is not nginx. Can anyone help me to find the reason?

[alzmcr]

Same problem here as @ue-it - work well locally but not from remote. Anyone aware of the reason?

[sokokaleb]

Same problem. The page says "Error response. Error code 404. Message: Not Found."

[armyja]

It doesn't not work after nginx with http://my.url/logtail/test.
The page says "504 Gateway Time-out nginx/1.6.2".
The fix: https://github.com/Supervisor/supervisor/pull/195/files can not solve my problem.

[simonegymondo]

I have the same issue but I get

Error response
Error code 410.
Message: Gone.

[heman4t]

Me too facing the same issue like @simonegymondo

[arnuschky]

Same here. Nginx 1.10.1 Chome 51.0.2704.79

2016/06/17 12:34:08 [error] 17986#17986: *10 upstream timed out (110: Connection timed out) 
while reading response header from upstream, client: XX.XX.XX.XX, server: supervisor.example.com,
request: "GET /logtail/crashmail HTTP/1.1", upstream: "http://127.0.0.1:9001/logtail/crashmail", 
host: "supervisor.example.com", referrer: "https://supervisor.example.com/"

[evgenity]

This patch helped our team: https://github.com/Supervisor/supervisor/pull/195

[julienfabre]

Had the same issue (404 not found) with nginx 1.10.2.
This nginx config (particularely the proxy_buffering off;, proxy_http_version 1.1; and proxy_set_header Connection "";) fixed the problem for me: https://serverfault.com/a/836855

[mnaberez]

All of the known issues with the tail -f output on the web interface are believed to be fixed as of Supervisor 4.2.0 (released 2020-04-30). That release includes the most recent fix, for the nginx proxy buffering issue immediately above, that was committed in 7d39a0fea0aa69061c3d2d1474d01559cd18ddfd.