• We value your experience with Plesk during 2024
    Plesk strives to perform even better in 2025. To help us improve further, please answer a few questions about your experience with Plesk Obsidian 2024.
    Please take this short survey:

    https://pt-research.typeform.com/to/AmZvSXkx
  • The Horde webmail has been deprecated. Its complete removal is scheduled for April 2025. For details and recommended actions, see the Feature and Deprecation Plan.
  • We’re working on enhancing the Monitoring feature in Plesk, and we could really use your expertise! If you’re open to sharing your experiences with server and website monitoring or providing feedback, we’d love to have a one-hour online meeting with you.

Resolved Apple OS and iOS - NSPOSIXErrorDomain:100

DerDanilo

Basic Pleskian
Since the migration from Debian 9 to Ubuntu 20.04 LTS all Apple customers report that they are having trouble accessing any of the websites on the server.
It was very difficult to diagnose this issue. The problem and it's solution is written down on the page below and seem not new. Why wasn't this addressed from Plesk earlier? Caused us a lot of headache.


We use default settings in Plesk and did not (yet) modify any templates or system files. But this issue should be addressed on a larger scale or there might be more trouble when people start migrating to recent systems.

Maybe there is also another issue but this seems to be the only solution for now.
How to properly impelment this serverwide?

---

Simplest soltuion for the moment seems to be:


Bash:
root@serverx:/etc/nginx/conf.d# ls
aa500_psa_tweaks.conf  fix_apple_conn_issue.conf  ssl.conf  zz010_psa_nginx.conf
root@serverx:/etc/nginx/conf.d# cat fix_apple_conn_issue.conf
proxy_hide_header Upgrade;
root@serverx:/etc/nginx/conf.d# systemctl restart nginx
root@serverx:/etc/nginx/conf.d# systemctl status nginx
● nginx.service - Startup script for nginx service
     Loaded: loaded (/lib/systemd/system/nginx.service; enabled; vendor preset: enabled)
    Drop-In: /usr/lib/systemd/system/nginx.service.d
             └─limit_nofile.conf, respawn.conf
     Active: active (running) since Wed 2020-09-02 20:20:29 CEST; 7s ago
    Process: 1519045 ExecStartPre=/bin/sh -c [ '${NGINX_ENABLED}' = 'yes' ] || { echo Not starting nginx as it is disabled in Plesk. Run \'plesk sbin nginxmng -e\' to enable nginx in Plesk.; exit 1; } (c>    Process: 1519050 ExecStartPre=/usr/sbin/nginx -t (code=exited, status=0/SUCCESS)
    Process: 1519057 ExecStart=/usr/sbin/nginx (code=exited, status=0/SUCCESS)
   Main PID: 1519058 (nginx)
      Tasks: 13 (limit: 23922)
     Memory: 23.9M
     CGroup: /system.slice/nginx.service
             ├─1519058 nginx: master process /usr/sbin/nginx
...
root@serverx:/etc/nginx/conf.d#
 
Last edited:
  • Like
Reactions: mow
.....The problem and it's solution is written down on the page below and seem not new. Why wasn't this addressed from Plesk earlier?
@DerDanilo Is this a Plesk problem or, a problem that's been generated from a local config / setup somewhere?

The reason we're asking the question is because we've used Ubuntu 18.04 LTS** OS since 18.04.1 was released c/w Plesk Onyx originally, now Obsidian and we've never, ever had any isssues with Apple (including the h2 related ones that you've posted above) on any of the sites / domains that we host. To be 100% sure, we've just run all the checks in the articles you've posted links to, using Apple devices running the latest releases (IOS 13.7 and MacOS 10.15.6 - so that's the latest Safari release in both cases) and, there's still no Apple problems anywhere. There's no Apple problems when using Firefox or Chrome on those Apple devices either FWIW.

As you are no doubt already aware, you can run lots of 'external to Plesk' tests on say apache / nginx / headers / h2 etc (in this case) to check things in detail
We use nginx as proxy but we don't need a /etc/nginx/conf.d/fix_apple_conn_issue.conf file as the issue is not present (in our case) so there is nothing to fix.
Attached is an extract from the curl check showing the h2 section. The image was taken using the latest curl release on MacOS 10.15.6

**We haven't upgraded to Ubuntu 20.04 LTS - yet - as there's still a slim chance that Plesk will actually provide a Dist-Upgrade option for Obsidian Users to upgrade their OS from 18.04 LTS > 20.04 LTS. So we'll wait and see if that happens. If not, we'll change to 20.04 LTS via a fresh OS install / Plesk migrate for 2021.

curl.png
 
Thanks for your detailed response. I appreciate that you took the time to give detailed feedback!

We installed a plain Ubuntu 20.04 LTS (netinstall ISO), added our SSH keys, our external DNS server, default ssh hardening and installed Plesk with the installer script. No further changes to the system whatsoever.
Usually we stay with Debian for stability but Plesk took so much time before officially supporting Debian Buster, that we decided to switch to Ubuntu on our Plesk servers. I don't like this choice but we cannot migrate customers every few years because of dropped support for older systems. At least Ubuntu LTS will get ~10 years of updates and most probably ~5 years support from Plesk.

We never had any issues with specific devices (exceplt extremely old ones) as well. This is the first time that something like this happened. I am wondering why this happens anyways. I don't think that Plesk did this on purpose but I am wondering why this didn't happen to any other customer yet. Maybe there are not yet enough Ubuntu 20.04 LTS users out there to receive enough feedback.

But to be honest, there were also some issues with Ubuntu 18.04 LTS and MariaDB back when it was released and we were told that we caused the issue, even though we installed a plain, naked Ubuntu 18.04 LTS and any further software with the Plesk installer. We had to build our own solution. This kind of problems never happened with Debian since we use Plesk 6 Years+, that's why we kept distance to Ubuntu until a few weeks ago. Hosting needs to be rock solid and not problematic now and then. No intension (neither time) to discuss this here. :)

The following packages are installed by default:
Plesk 18.0.29 Ubuntu 20.04 1800200825.21, Operating System: Ubuntu 20.04.1 LTS, Kernel: Linux 5.4.0-42-generic, nginx version: nginx/1.18.0, Server version: Apache/2.4.41 (Ubuntu), Server built: 2020-08-12T19:46:17, mariadb Ver 15.1 Distrib 10.3.22-MariaDB, Postfix 3.4.13, Dovecot 2.3.10.1
 
Are we really the only ones with this kind of issue? We are not running any magic on our systems. We don't even modify them.
Clean Ubuntu and then Plesk installer. Using Plesk Migrator to migrate sites, not erros reported there.
 
Yes i can confirm that problem exists and plesk support tries to make it look like this is problem with wordpress websites hosted with plesk. Have played all weekend with this problem and it is driving me crazy. Problem can't be solved, can't find any information with it.

Debian 10 + Plesk latest version.

If i use plesk bin http2_pref disable - websites opens at Safari browser mac/ios. STILL WORKS WITH ANDROID, WINDOWS.
If i use plesk bin http2_pref enable - websites on Safari MAC,IOS doesn't load and gives error: WORKS WITH ANDROID, WINDOWS
 
Yes i can confirm that problem exists and plesk support tries to make it look like this is problem with wordpress websites hosted with plesk. Have played all weekend with this problem and it is driving me crazy. Problem can't be solved, can't find any information with it.

Debian 10 + Plesk latest version.

If i use plesk bin http2_pref disable - websites opens at Safari browser mac/ios. STILL WORKS WITH ANDROID, WINDOWS.
If i use plesk bin http2_pref enable - websites on Safari MAC,IOS doesn't load and gives error: WORKS WITH ANDROID, WINDOWS
I litterally posted a workaround. Maybe you didn't see that. No need to drive yourself crazy. Still Plesk should fix this.

Here is a one liner that should to the trick:
Bash:
echo 'proxy_hide_header Upgrade;' > /etc/nginx/conf.d/fix_apple_conn_issue.conf; systemctl restart nginx; systemctl status nginx
 
  • Like
Reactions: mow
Thank you @DerDanilo ! Worked perfectly!!!!!

I have tried this one before:
echo "proxy_hide_header Upgrade;" >> /etc/nginx/nginx-includes.conf
This one worked perfectly:
echo 'proxy_hide_header Upgrade;' > /etc/nginx/conf.d/fix_apple_conn_issue.conf; systemctl restart nginx; systemctl status nginx
Yes i think plesk must fix this... but they just said it is wordpress problem...p.s. not all websites runs on plesk... :)!
 
Come on Plesk - nearly a year later and still this is an issue with a standard Plesk Apache + Nginx setup. How is that the case?
 
  • Like
Reactions: mow
Back
Top