Localhost - "Welcome to CentOS" displayed - docs followed


  • Used Zammad version: 3.6? (whatever is pulled using install doc)
  • Used Zammad installation source: (source, package, …)
  • Operating system: CentOS 7 (2009) fresh install - with or without GUI
  • Browser + version: any

Expected behavior:

Zammad Welcome when browsing to localhost

Actual behavior:

browsing to “localhost” does not show Zammad Welcome
Instead the “welcome to CentOS” page is shown.

Steps to reproduce the behavior:

follow (exactly) docs here

Note: The “elasticsearch” instructions here

Are not complete either
Installing ingest generates a plugin permissions error
–batch must be used to bypass

TLDR - these docs are messy, apparently outdated and incomplete and following them does not result in a working installation. I have seen multiple threads with the same issue and no clear reason or resolution other than to “check the config” - ehh thats not a help to me.

Attempting to use Centos 8 results in similar errors but the nginx welcome is shown instead - and elastic install command results in file not found for one of the components.

For those with the same issue on CentOS

If you follow the (actually rather poor) install docs you will end up with a default nginx configuration that includes and also overrides the zammad.conf file.

include /etc/nginx/conf.d/*.conf;

server {
listen       80 default_server;
listen       [::]:80 default_server;
server_name  _;
root         /usr/share/nginx/html;

You need to remove or comment out the bolded line:
root /usr/share/nginx/html;

It would take 10 secs for the docs to be updated - but alas, I assume this will be ignored.

1 Like

Pull requests are welcome if you’re unhappy with the docs.


You pretty much just made my point for me, twofold.

A) You could have fixed the docs in the same time it took you to respond.
B) Chances are almost 100% that you or a team member will not go fix what was reported… even after this dust-up.

–Predictable open source project management 101

My “happiness” with the docs has nothing to do with my point. The docs (at least for centos) have two glaring omissions that result in a broken install for all but centos gurus. There are numerous threads here and several in the tracker, most ignored or closed with no resolution.

1 - the conf.d issue mentioned here. zammd will not work per intstall doc.
2 - elasticsearch can NOT be installed per the docs unless --batch is added to the ingest install directive. Otherwise ingest plugin fails and zammad will not run.

So right out of the gate - anyone trying to get up and running with this project is stopped dead in their tracks with two insurmountable hurdles. I would say that is a BIG deal…

No offense, but the same applies for typing the part with the remarks to that topic.

Stunning, thanks for your trust in us and our project.
In fact we do accept pull requests and change suggestions if they make sense and do cover the broad user base.

Okay two things.
Discourse closes stale topics automatically without a chance to adjust that as of now.

This community board is community driven. The Zammad team tries to answer and help on issues if in application scope and the time allows it. This is not always possible I’m afraid. Nobody ever guruanteed solutions for everything on here.

Still, I got you covered:

There’s a rework of the installation docs going on. It takes time though.

Now to your points.

Yes and no.
This is a configuration specific issue. The documentation, except for the last sentence “go to localhost blah” aims for a public reachable Zammad instance. This means that your setup just might not fit the described use case that the documentation aims for. This is the part the “broad userbase” comes back in. We can’t fit any little detail. We also can’t cover everyone.

I’m very sorry for that, but you can’t always expect the “I get everything package” for giving nothing back. Please don’t get me wrong here. It’s just impossible to cover everyones needs and requirements.

I can’t reproduce that.
The documentation is tested against default installations of each system we do list and cover.

Your two above points do not happen in my enviroments. I do regulary test and re-test installation steps and ensure that they are working. That is part of my many duties. While I do agree that the installation documentation is not ideal and an ugly desaster in terms of “grown things”, this is something that takes time. If you’re working alone on such big documentationary that we do provide you’ll have to have grey spots etc.

I’m sorry that those do cover you - but that’s something clear feedback or maybe even a pull request can cover. The posts you mention are not always covering the same directions, this is why saying “we have plenty of those and they’re all unanswered” are not a benchmark.

I’m currently behind up to 16 days - and will take my next week off to cool down on this topic.

Have a nice weekend.

1 Like

So - That kind of brings us full circle. Thank you (honestly) for the response.

I should have just posted what i found and left my frustration out of it.

FWIW - It is not lack of faith/trust in THIS project or its team - it is the frustration of attempting to consume/use the majority of projects like this. I think often the devs and regular contributors often lose sight of the fact that most of us are not devs too!

The open source community is pretty much divided into the gurus and the code consumers. Most code consumers do not know how to make a pull request - even if they did stumble upon the correct answer somewhere. Searching for answers often results in more frustration when numerous similar issues are also unanswered.

regarding conf.d
I am not sure that I understand. Unless I am missing something, Will the result will not always be the nginx default server and not zammad, regardless of what is put in the zammad.conf - as it is negated by the root directive that comes AFTER the include. I know very little about nginx - just what I have read while researching my issue.

Regarding ElasticSearch
I have tried on 2 fresh installs of CentOS 7 - following the zammd docs 100%
Elastic Generates a plugin error
Typing “y” at the permissions prompt is ignored and the install is rolled back

per the link above using --batch fixed the issue and allowed the install to work

Thanks again for your thoughtful response.

I can only guess.
But: If you use a named server within Zammads conf file ( e.g. server_name my.host.name or server_name nginx will exactly know which host configuration it should react to.

The default.conf file is working with a wildcard that technically reacts to any request that does not fit other requests. So if your config for Zammad uses wildcards too, the easiest way would be to remove default.conf if you don’t want to use specific named configurations (which I would suggest for Zammad).

In nginx and, as far as I’m aware apache as well, the first fitting configuration file will trigger.
This means if default.conf would be renamed to zzz_default.conf, you should see the Zammad page after reloading nginx. Whatever you do, don’t rename zammad.conf in your nginx directory. The package installation will other wise copy a new file version into the directory which will cause your nginx to have a faulty configuration and no longer start.

As I already said, I’m currently working on a rework documentation on those parts which will introduce a webserver part ( https://github.com/zammad/zammad-documentation/blob/mh/rework/installation-documentation/getting-started/configure-webserver.rst sneak peak).

However: It will be and is impossible to catch every one. Our documentation (especially docs.zammad.org) require basic linux and administration skills. I mean yes, we could cover the whole part from “how to install my machine” to “finish Zammad setup”, but that’s an absolute nightmare in terms of maintaning given the fact that I’m basically working alone on those documentation parts beside my regular todos. (I could fill my whole day with documentation only)

Can you please point out which ISO of CentOS 7 you’re using?
During the rework of said documentation I’ve been touching CentOS several times this month and never experienced this kind of issue with latest updates.

If it helps, I’m using CentOS7 x64 minimal for those situations.
No xServer because it’s a server.

When working on those documentation part’s I’m always starting from a clean install which is why specific environments might react different to what I see. That’s the part where it will be hard to cover all single situations even though I do understand your frustration at that end.

This topic was automatically closed 120 days after the last reply. New replies are no longer allowed.