Problems with new install of 3.4.x

Hi all,

New to Zammad here. Please bear with me if this was already answered but I could not find the solution…
I am trying to introduce Zammad to my company but even at version 3.4.x it does not seem to be ready for prime time/production…

  • Used Zammad version: 3.4.x
  • Used Zammad installation source: official RPM from Zammad repo
  • Operating system: CentOS 8
  • Browser + version: Firefox, Chrome, M$ Edge (all latest)

Expected behavior:

  • Normal use of Zammad

Actual behavior:

  • Zammad was working flawlessly after a complete restore on another server where I installed Zammad from scratch and then restored the PostgreSQL dump. I thought “wow, this is really robust!”… I logged off and called it a day.

This morning I still could access Zammad but it went straight to the Manage menu.
My Dashboard, Stats, and the Knowledge Base are all marked as “Unauthorized”.
Furthermore, clicking on tickets I created takes me to the same Manage menu…

Steps to reproduce the behavior:

  • Install CentOS8 + elasticsearch 7.9.0 + Zammad 3.4.x
  • import PostgreSQL database from oñder/dead system (dump is all good, all old tickets were successfully imported and showed up on columd on the left hand-side)
  • Leave it “at rest” for a few hours; i.e.: finish your working hours and disconnect
  • Login the next day and find a non-working Zammad system

Thank you very much in advance!

Best regards,
Martin

Can you please define what exactly you mean by “non working Zammad” ?
What was the source version you migrated from?

Any errors during import or in your production log in general?

“non-working Zammad” = exactly that. Zammad is not working as expected (or as it was yesterday right after the migration). Whenever I click on my tickets, I get redirected to the Manage page. I cannot access my Dashboard or the Stats… Lately I found out that I can, for now, access the KB…

Old version is 3.2.

No, no errors during the import process.

Users are imported from LDAP and I can see them under the Users menu.
A particular LDAP group grants Admin rights to some users, mine is among them.

Contacts are being read from Exchange as well.

And if I restart PostgreSQL, Zammad and nginx, I get a CSRF error and I must revert to the latest snapshot on VMware… Still it does not work as expected, but at least I can login :slight_smile:

Do you guys need logs? if so, which ones?

Thanks in advance!

Could (or could not) be a issue somehow related with the problem I’ve had shortly, when trying to import a zammad 3.2 db dump on a fresh new zammad 3.4 installation - here are all the details: Upgrading Zammad from version 3.2 to version 3.4

TLDR: after importing the db dump on the zammad 3.4 installation try running “rails zammad:db:init” - might be worth a try…

I get this both as root and as zammad:

$ /opt/zammad/bin/rails zammad:db:init
Traceback (most recent call last):
4: from /opt/zammad/bin/rails:8:in <main>' 3: from /opt/zammad/bin/rails:8:in require_relative’
2: from /opt/zammad/config/boot.rb:3:in <top (required)>' 1: from /usr/share/rubygems/rubygems/core_ext/kernel_require.rb:59:in require’
/usr/share/rubygems/rubygems/core_ext/kernel_require.rb:59:in `require’: cannot load such file – bundler/setup (LoadError)

yum install -y zammad installed all prerequisites… am I missing something?

TIA,
Martin

did you have a look in the admin documentation: https://docs.zammad.org/en/latest/admin/console.html

As I run a “dockerized” zammad (and the rails command was just working right from the start inside the zammad container) I really dont know whats exactly needed to get it working in your setup…sorry. I only could guess you need maybe to set some env variables…

for some reason, Zammad’s RPM does not install Ruby…

After installing it, I get the same error message…

Time to run “terraform destroy” again and start over again…

Getting a bit disappointed…

Zammad does come with a pre built and thus shipped ruby and rails version.
This allows Zammad to be independent from the environment you may have installed.

Zammad is strict on the required ruby version.
Also, as it seems like your system runs perfectly fine after restoration for some minutes / hours, that seems a bit odd to me.

If you get CSRF errors, this is due to your webservers host configuration, this is covered on point 2.1 here: Zammad 3.2 is available

Above run commands (directly calling bin etc) are not correct.
This is partly covered here: https://docs.zammad.org/en/latest/admin/console.html

a db:init shouldn’t be required in your case, at least that’s what I currently expect as it does seem to be a login issue on your end.

If it still doesn’t work, please provide your production logfile (/var/log/zammad) and we’ll continue from there.

I ended up ditching Zammad. Too unstable and not production-ready for my taste.
Sorry guys. I will keep an eye on it because I liked it a lot but it’s a pity that it is so…sensible.

Best,
Martin

Sorry for that impression.
We have dozens stable production enviroments.

I can totally understand that the first impression wasn’t too good for you.