Zammad 6.0 Mobile client issues

Technical assistance
1

After upgrade [Centos 7, Zammad 5.x (latest current), nginx] to Zammad 6, i start testing mobile app on Android phone (OnePlus 8T, OxygenOS 13.1, MS Edge browser).
Zammad mobile version shows red banner: “The connection to the server was lost.” almost constantly.
I can login, browse tickets, change my avatar photo, but I am unable to add reply to open ticket:
got an spinning circle, if i refresh page, browser returns to ticket body, and again red banner - connection to server was lost. I see no errors in nginx zammad access/error logs. no errors in production.log
Solution: Your reverse proxy MUST support websockets.
About production.log: I found, what it is now 133MB, and my uploaded avatar photo is logged in it.
See attached photo.
I discovered, what with single photo upload via mobile client, i did it like 6 times. When edited my avatar via desktop web, there was 1 old photo and 6 copies of new photo…
And with desktop client i never got server connection lost error.
After few hours of uptime, there is few errors in production.log:

WARN – : Scoped order is ignored, it’s forced to be batch order.
ERROR – : No route matches [GET] /wp-login.php (ActionController::RoutingError)
ERROR – : No route matches [GET] /wordpress/wp-login.php (ActionController::RoutingError)
ERROR – : No route matches [GET] /blog/wp-login.php (ActionController::RoutingError)
ERROR – : No route matches [GET] /wp/wp-login.php (ActionController::RoutingError)
Where is wordpress taken from?
Answer: This was external script scanning Zammad. Not an error, but attempt to scan for wordpress.

P.S. Log level can be changed in /opt/zammad/config/environments/production.rb
(was) config.log_level = :info
(now) config.log_level = :warn

Zammad_avatar
Zammad_avatar1920×508 311 KB

1 Like
2

Same issue here, even on fresh install on Ubuntu 22.04

3

Make sure the Redis service is active

systemctl start redis
systemctl enable redis

And that these lines are inside your NGINX

1 Like
4

redis is running. without it you will get “bad gateway” error, and no zammad at all.
/etc/nginx/conf.d/zammad.conf contains first block at the beginning, i did not put it there.
and second block i did copy-paste right after upgrade, as per release notes, right under location /ws {} block

5

Have you looked at the Nginx log?

6

Take a look here

7

I did. And yes, it was my case. After i sucsessfully run postinstall script, I got Zammad working in desktop mode. Unfortunately, mobile mode looks like different beast.

8

Blockquote /var/log/nginx/error.log
[error] 36652#36652: OCSP_basic_verify() failed (SSL: error:27069076:OCSP routines:OCSP_basic_verify:signer certificate not found) while requesting certificate status, responder: ocsp.sectigo.com, peer: 104.18.15.101:80, certificate: “/etc/pki/tls/private/helpdesk.crt”

Blockquote /var/log/nginx/zammad.error.log
*37470 open() “/opt/zammad/public/robots.txt” failed (2: No such file or directory), client: 10.10.x.x, server: support.xxxx.lv, request: “GET /robots.txt HTTP/1.1”, host: “support.xxxx.lv”

Blockquote /var/log/nginx/zammad.access.log
10.10.x.x - - [11/Jun/2023:10:16:13 +0300] “POST /api/v1/message_send HTTP/1.1” 401 102 “https://support.xxxx.lv/” “Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:109.0) Gecko/20100101 Firefox/113.0”
10.10.x.x - - [11/Jun/2023:10:16:14 +0300] “POST /api/v1/message_send HTTP/1.1” 401 102 “https://support.xxxx.lv/” “Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/113.0.0.0 Safari/537.36”

9

I updated to zammad 6.0 and see the red banner too on the mobile page.
grafik

from: /var/log/zammad/production.log

I, [2023-06-11T11:02:40.535150#2897-113000]  INFO -- : Started GET "/cable/"[non-WebSocket] for 91.21.250.245 at 2023-06-11 11:02:40 +0200
E, [2023-06-11T11:02:40.535211#2897-113000] ERROR -- : Failed to upgrade to WebSocket (REQUEST_METHOD: GET, HTTP_CONNECTION: Upgrade, HTTP_UPGRADE: )
I, [2023-06-11T11:02:40.535250#2897-113000]  INFO -- : Finished "/cable/"[non-WebSocket] for 91.21.250.245 at 2023-06-11 11:02:40 +0200

I have ran the postinstall script again but no success, no errors during postinstall.

10

Same problem here - same errors in the logs as icecoffee.
Zammad 6 (update from 5.4.1)
Debian 11 with apache webserver,
Desktop version runs without flaws

11

Hi @voljka, hi @donbertl, hi @icecoffee :slight_smile:

Are there any other errors reported in your log files?

Did you update your nginx config file for Zammad?

12

Yes,I did, as per post @YosefAdPro

13

Ensure HTTP_TYPE and FQDN are set correctly to what you’re using in your Browser.
In my tests this was crucial when everything else (like nginx) was configured correctly.

14

Please be more specific. I do not understood, where and what to check .

15

See Base — Zammad documentation .

16 
rails r "pp Setting.get('http_type')"
rails r "pp Setting.get('fqdn')"
17

Hello, there are no more errors in any log files. Yes I updated the nginx file correctly.

root@***:~# zammad run rails r "pp Setting.get('http_type')"
"https"
root@***:~# zammad run rails r "pp Setting.get('fqdn')"
"<yes.subdomain.correct>"
1 Like
18

Redis is up & running?

19 
root@***:~# redis-cli ping
PONG
20

@MrGeneration Yes, both are correct.
zammad run rails r “pp Setting.get(‘http_type’)” => https
zammad run rails r “pp Setting.get(‘fqdn’)” => support.our_domain.lv

I just realised, what we use Microsoft Web Application proxy, but it set simple as
(internet) https://support.our_domain.lv => (dmz) https://support.our_domain.lv
it has wildcard certificate: *.our_domain.lv

Zammad host has nginx installed, configured as per instructions, with https and same wildcard certificate for zammad site.

With this config desktop clients works fine. What I did not check yet, is how mobile client will work inside our LAN, without going through MS WAP.