Various and different installations issues on a VM

FreePBX Installation / Upgrade
1

Hello,

The forum has pointed out that I spend a lot of time here and suggested I post . So here I go.

I’m quite passionate about open-source software, and while I’m not a computer scientist, I manage to get by—enough to run Nextcloud on a shared host, for example.

But I must admit, I’ve rarely struggled this much with an installation.

On VMs with maxed-out RAM, 4 CPUs, and 60GB of storage, I’ve never managed to complete the installation successfully. When I do reach, sometimes, the UI login, I enter a username and password, only to get a “credentials error” message.

The script on Debian 12, after five or six attempts, has at best brought me to this stage. Most of the time, it seems to be issues with users in the database, from what I can gather.

SangomaOS crashed once during the npm installation, and now it can’t find the Debian installation mirrors (though pings work fine).

What really baffles me is that the errors are different every time.

I’ve read that this should take about 30 minutes to set up. I feel like I’ve lost two days. Apologies if I come across as the typical newcomer complaining before even introducing myself.

If anyone has a step-by-step tutorial with precise VM configurations, I’d be very grateful.

Thank you,

(P.S. I’ve done my best to read the documentation.
I’m using Ubuntu 24.04 and VirtualBox.)

2

I will keep you updated on any developments. The installation of “in1click,” which was initiated one hour ago, has been displaying “Installing npm…” for the past 18 minutes.

Is this normal? After how much time should I start to be concerned?

THX

3

Hi, welcome to the forum!

The official installation method is described at https://www.freepbx.org/downloads/

“in1click” is a third-party tool (that I do not know anything about other than having heard of it).

I can vouch for the official script; it has worked many times over for me, on hosted VMs with as little as 1 CPU and 2GB RAM.

4

Hello and welcome to the community!

​Since the installation is failing, the best place to find the root cause is the installation log.

​Please look in the folder /var/log/pbx/. You will find a file starting with freepbx17-install-…

Could you check that file to see if there are any error messages recorded?

5

Hello,

@billsimon By mistake, I used the term “in1click”, but I was actually referring to the latest ISO.

@Mauro I was on my way to click post with the following question : “Do you have any suggestions on how I can retrieve the installation log? “. You did reply to it. however last time I got stuck in a booting loop with no access to the terminal.

At the moment , My CPUs are processing something, but I am stuck on “installing npm…” for 45 minutes.

:astonished_face: Interestingly, as I was typing this message, I pressed “Escape” on the terminal, and it suddenly jumped forward, confirming that npm had installed successfully and proceeded to ipset.

If this can help anyone: if your terminal gets stuck during installation, try pressing “Escape”.

I keep you informed,

Best,

1 Like
7

Hey,

I hope my comments will be helpful.
After reboot, I now have an “installation failed” message.

I am being asked to run a dpkg --configure -a or a systemctl disable installation command (exact command to be confirmed) to exit the boot loop, but… I don’t have time to log in before the auto-login occurs, and then it just bootloops again.

And if I try to enter rescue mode, it asks for the superadministrator password, which is not the same as the Sangoma password…

So, no access to the console, and therefore no access to the logs…

Frustrating!

I will therefore attempt the installation for the 10th time, and this time I will change the distribution version.

Thank you for your attention to this matter :roll_eyes:

8

Have you tried the install script on top of a plain Debian 12 OS, rather than the ISO?

9

Yes I did.

By memory, a lot of extensions were missing. Logs related to some issues with pbxuser in DB.
I got once to the admin page but after entering user + password no way to log in and credentials error message …

Thank you for your time.

PS : Another OS installation attempt has just thrown an installation error mentionning a kernel installation issue now. Madness. That must be something to do with virtualbox but that’s never the same issue…

10

Sorry, your story makes no sense, as you said you never got an installation to work in the first place. Extensions missing? As in you restored a backup and the data was corrupted?

Yes, your virtual environment is broken. Please try it on an environment that is known to work; DigitalOcean and Vultr are good choices.

11

Thank you once again.

I am indeed working on a personal server. I went a step further by disabling the host computer’s power-saving screen—not entirely sure if this is related.

At one point, I had to press Escape twice to exit what seemed like a freeze (with a single step taking over 5 minutes to complete).

After about an hour of installation, everything appeared stable. I was able to log into the system using the username sangoma and the personal password I created during installation (of the PUB Spice at the end of the day). However, instead of the admin page, I was greeted by the Apache2 Debian Default Page.

Upon checking, the /var/www/html directory contained only the default index.html file.

The last two lines of the installation log were as follows, and then nothing further:

Preparing to unpack …/378-node-ssri_9.0.1-2_all.deb …
Unpacking node-ssri (9.0.1-2) …

I attempted to run dpkg --configure -a, but it was locked by an apt-get process. After terminating the process, I managed to execute dpkg --configure -a, which triggered a reboot—but I was back in the boot loop.

However, via SSH, I had just enough time to connect—thanks to the short password I had set—and perform the well-known command: sudo systemctl disable sng-install-freepbx. It rebooted once more, but at least it waited for further instructions this time.

Another dpkg --configure -a initiated numerous configurations, including node-ssri.

On instinct, I then ran: sudo systemctl enable sng-install-freepbx followed by sudo shutdown -r now.

The system restarted, and additional installation processes (such as ipset, fail2ban, and htop) began, lasting at least 12 minutes. I had allocated 4 CPUs and 16GB of RAM to the VM.

For this rather unusual setup, I did follow the following steps:

  1. Set a short password instead of the default one to ensure you have enough time to execute: sudo systemctl disable sng-install-freepbx

  2. Wait for the next restart.

  3. Run dpkg --configure -a (after killing any locking processes) as many times as necessary.

  4. Execute sudo systemctl enable sng-install-freepbx.

  5. Finally, run sudo reboot -n.

ie : The system remained stuck for 15 minutes on “installing asterisk-sounds-*” before I intervened. I had to press Ctrl-D and repeat the above steps.

Unfortunately, the installing FreePBX 17 did not succeed. Despite these efforts, the furthest I could get was a restart that froze at “making sure installation is sane”. Even a manual sudo apt install freepbx17 stalled at 20%.
sudo apt-get install --reinstall apt relaunched the installation freebpx17 with no success neither.

I let my computer working on this but there is no ressource in use about that. Installation logs for what I see are about this apt installation not succedding.

Any help on this last attempt of an OS installation on a VM started 4 hours ago would be more than welcome. Any clue to help or set the VM

(I add VM in the title of this thread)

12

I recommend that you start over, taking care to install the base OS correctly.

Follow the guide at https://sangomakb.atlassian.net/wiki/spaces/FP/pages/230326391

especially the instructions at https://sangomakb.atlassian.net/wiki/spaces/FP/pages/295403538

If no luck, post first thing that goes wrong.

13

@Stewart1 :
This is what I did on first intention.

By the way,on this last attempt the Debian Installation process went good.

However, after 20min, the script got stuck on “Installing npm” for 23 min. Thanks to another ssh session, I check the logs .They were moving about dependencies.
Another story again about “Installing Freepbx 17”. It took 33 min with absolutely nothing happening in the logs. I believed the few activities displayed by the VM were relative to the SSH sessions but at the end of the day: “no”. It reached another screen.

The console tells me now that I have 4 notifications in the UI. I did not see any.

I have been able to connect to the UI, without hassles. A F5 had been necessary after activation.
But I have started to “play” with the settings.

My guess is that on my first attempts were I gave the full night to script to apply, the screen saver of the host messed up. So I would advise to disable it and to be patient. It took me 3 hours between the Debian OS and the Script.

Sorry to have bothered you.

Thank you @Stewart1 @billsimon and @Mauro

1 Like
14

:face_with_peeking_eye: The system told me “Firewall activated” and I have no more way to access to the UX. But this is another story.

At the end of the day, I keep this diary of a beginner open :slight_smile: :

I got mad because the UI was not accessible anymore. My browser added an S to the htpp. I have to find a way to create a SSL certificate on a local adress.

At the moment I am stuck on the firewall settings with : “Ajax request broken or aborted for an unexpected reason. Please check console logs for more details.” . I took me a while to understand that it was speaking about the browser console where I did not see anything revelant before being quick out of the UI (failed connexion …)

15

If you really are completely stuck with FreePBX 17, try downloading the FreePBX 16 ISO. It is solid, and it works 100% on VM. The install is super quick and easy. Using that ISO, you don’t have to install any OS first at all.

16

@userjf Hello.
At the end of the day, I have been able to install Feepbx17 with the script .
I still face some issues, (auto-login on the server start, setting of the firewall that tend to block any legitimate access) but I have to focus on other matters at the moment (I did not plan to spend so much time on it this week…).
I may post a kind of tutorial here later on.
Best,

17

:index_pointing_up:IN1CLICK

Option 1: On a fresh Debian 12 installation, run this curl command as root:

curl freepbx.in1.click | sh

Option 2: On a fresh Debian 12 installation, run this wget command as root:

wget https://raw.githubusercontent.com/20telecom/IN1CLICK/main/freepbx17-on-debian12 -O /tmp/IN1CLICK && chmod +x /tmp/IN1CLICK && /tmp/IN1CLICK

You may still run into underlying issues, but it might tell you what’s wrong along the way :rocket:

Option 3: Follow the official guidance. As @billsimon says, this is indeed a third-party tool.

1 Like
18

I use this a lot. Thank you.

1 Like
19

My advice after lots of experience with VMs is that I would bet you are trying to run the installers at the various virtual consoles that the hypervisors supply. This is always going to kill you. Never do this.

As a general rule with virtual UNIX or UNIX-alike systems, the hypervisor console exists for ONE reason - to get the operating system installed. When installing UNIX-alike or actual UNIX systems under hypervisors, always use either real-live Xwindows xterms, or install xdrp and use Windows Remote Desktop Client, if you absolutely must have a GUI into the OS. Otherwise, you should use SSH exclusively. Only use the hypervisor console to install the OS and select OpenSSH during the installation. After that, the hypervisor console becomes the ultimate stunsail. Star Trek writers used that term to invent the insult dunsel, which you may be more familiar with. (or from the Dictionary of American Hayseed sayings “useful as teats on a boar”)

1 Like
20

I am going to explore this

21

Well. My previous installation was so instable that I decided to change my VM supervisor.
I went to Qemu / KVM. After setting the connection bridge, it took the server only 40 minutes to finish the installation of the Sangoma freepbx17 OS
Logs shows no errors in the installation.
I had to manually login in my Sangoma user in the terminal.
However, once I set login/password on the admin page, I have a credential error to connect.
I did fwconsole unlock id session. Blank screen on the UI…

I decided to restart the VM, I have now an apache2 error on starting but once logged systemctl shows the service as “active”.
Blank screen on the browser… :weary_face:
And no way to access the console via ssh (The VM is connected on my local network, and I had no issue with this last point on Virtual Box). To finish this week to nowhere, I am going to try the IN1Click solution of @kierknoby on QEMU and let you know.

Regards,