v20

[Albano Belchior]

Saudações estimadas ,

Vim solicita a sua ajuda. Após a migração da v18 para v20, tenho enfrentado diversas situações
Quando efetuamos qualquer atualização ou redefinimos os serviços por via ssh, perdi o acesso à plataforma .

E a informação constante é esta que partilho em anexo, podem ajudar por favor a ultrapassar esta instabilidade.

 

[Guillaume Bourgeois]

Please write in English to get international assistance. If you prefer Portuguese, register on the local forum at https://www.3cx.com.br/forum/.

Was your update successful or did it fail? Normally, you should receive an email from your phone system notifying you of the success (or failure) of the update.

 

[Albano Belchior]

Greetings

I would like to ask for your help. After migrating from v18 to v20, I have faced several situations
When we perform any update or restart the services via SSH, we have frequently lost access to the platform.

And the constant information that appears is this that I share attached, can you please help to overcome this instability.

 

[Guillaume Bourgeois]

Did it work for a while after the V18 to V20 update, or did it stop working immediately during the update?

 

[Albano Belchior]

It worked for a while, which means I have more than three licences, and in between I'm constantly getting this information on my machine

 

[Guillaume Bourgeois]

execute

sudo systemctl list-units --failed

Share the command output.

 

[Albano Belchior]

I'll share the printout

 

[fxbastler]

1. Is the 3CX operated in-house or externally?
2. Is the 3CX a VM or is it installed directly on the hardware?
3. Is the underlying hardware having problems?

I could ask many more questions to clarify your problem, but my advice: create a good backup of the 3CX, store the backup outside of the 3CX, check the backup, completely reinstall with the 3CX v20 ISO and restore the backup.

 

[Guillaume Bourgeois]

try

sudo systemctl restart 3CXPhoneSystemMC01

 

[fxbastler]

From what the TO wrote, I don't think this will work for long.

 

[Guillaume Bourgeois]

It might not solve the problem, but it will allow us to increase verbosity.
Then, if the issue occurs again, we can generate a support info file after another restart.

Of course, one could perform a full backup and reinstall everything, but that isn't feasible for everyone and it doesn’t truly solve the underlying issues. Diagnosis is the first step, followed by action.

I wouldn’t want a doctor to prescribe pain relief without first understanding the underlying problem.

 

[fxbastler]

If the TO doesn't have a hardware problem, then it's a software problem (and hopefully not with the setup, that would be illogical).

Yes, you can search for a long time and maybe at some point find the cause and hopefully solve it. A clean reinstallation solves it immediately.

A 3CX is not human, you don't need to go to the doctor. This is very pragmatic, but murder and clone is much faster and cheaper

 

[Albano Belchior]

Can you share if you are also experiencing these problems and how you would solve them and avoid stress?

 

[Guillaume Bourgeois]

If the TO doesn't have a hardware problem, then it's a software problem (and hopefully not with the setup, that would be illogical).

Yes, you can search for a long time and maybe at some point find the cause and hopefully solve it. A clean reinstallation solves it immediately.

A 3CX is not human, you don't need to go to the doctor. Murder and clone is much faster and cheaper

And if it's simply because the system lacks sufficient CPU/RAM to handle the workload, a clean reinstall won’t solve the problem. Instead, increasing the resources, not reinstalling, would be the necessary solution.

I've seen a lot throughout my career. Before becoming self-employed, I started working with a partner; we had 0 PBXs... and by the time I left, we had reached Titanium level. I can assure you that reinstalling doesn’t fix everything; it addresses OS issues, but there are many other scenarios that can cause problems!

So yes, you can give advice your way, and I respect that, but don’t say that taking the time to diagnose is a waste of time. You haven’t even analyzed the logs yet.

 

[Albano Belchior]

1. Is the 3CX operated in-house or externally?
2. Is the 3CX a VM or is it installed directly on the hardware?
3. Is the underlying hardware having problems?

I could ask many more questions to clarify your problem, but my advice: create a good backup of the 3CX, store the backup outside of the 3CX, check the backup, completely reinstall with the 3CX v20 ISO and restore the backup.

1. Yes. I have it in-house.
2. It is a VM.
3. No. It doesn't.


I already did this way. I made the update by creating a new and clean 3CX version.

 

[fxbastler]

I already did this way. I made the update by creating a new and clean 3CX version.

... and the 3CX or its services die repeatedly? Without interacting? Interesting.

Are there sufficient system resources available? What kind of hypervisor is that? Has ntp been installed?

Edit:
I had a very similar behavior with a 3CX v18 on a KVM hypervisor. Even though the 3CX basically had enough RAM and CPU resources: the KVM was also running a VM with workflow software with OCR (which was only used sporadically), which reproducibly caused various 3CX services to stop after several days. They were sharing some CPU cores. Could that be a reason for your problem?

 

[Guillaume Bourgeois]

... and the 3CX or its services die repeatedly? Without interacting? Interesting.

Are there sufficient system resources available? What kind of hypervisor is that? Has ntp been installed?

Ah, now you're doing a diagnosis...

 

[fxbastler]

Ah, now you're doing a diagnosis...

Yes, now that we know it's a clean installation. That's a very good basis for it. No need to laugh

 

[Albano Belchior]

... and the 3CX or its services die repeatedly? Without interacting? Interesting.

Are there sufficient system resources available? What kind of hypervisor is that? Has ntp been installed?

Yes. Its dies but only tow of them (at liest that i noticied):

1. 3CX PhoneSystem 01 Management Console
2. 3CX PhoneSystem 01 System Server.

Those two are that stos frequently.


System Info

[General]
ProcessorCount = 4
TotalDiskSpace = 28 GB
FreeDiskSpace = 11 GB
TotalPhysicalMemory = 7.721 GB
FreePhysicalMemory = 6.357 GB
OSArchitecture = X64
OSDescription = Debian GNU/Linux 12 (bookworm)
ProcessArchitecture = X64
FrameworkDescription = .NET 8.0.7

[3CX PhoneSystem Info]
Number of Extensions = 22
Number of Queues = 5
Number of RingGroups = 2
Number of IVRs = 6
MULTICOMPANYMODE = 0
ALLOWUSEBUSYOPTFORGROUP = 0
FIREWALL_CHECKER_RESULT = 0
SYSTEMLOGGINGLEVEL = 1
VERSION = 20.0.3.659
RESTORED_FROM_VERSION =
ADJUST_FROM_VERSION = 20.0.3.659




PS: This one of the 3CX that I admininstrating. I have 4 more and the bigest one has 32 simultaneos calls and has the same issues.

 

[fxbastler]

Those two are that stos frequently.

Is this somehow reproducible? How many CPUs (not cores) does the hypervisor have? Which kind of hypervisor?

Addendum: Why use a non-production 3CX?