Looking for someone to test & help

A few months ago I came across an issue where if asterisk is reloaded while a voicemail is being recorded then the email address then becomes corrupt and is never sent.

I opened a ticket with asterisk and 2 months later they have came back unable to replicate

I am currently tip-toeing around this situation by making system changes then looking at active channels, if no voicemail is being left i can reload, if there is I have to wait.

My system originally consisted of FreePBX 2.10 32-bit which has since been upgraded to 2.11 and then moved to brand new hardware which I then installed a fresh 2.11 64-bit and did a restore via the backup/restore module. I then changed astmoddir => /usr/lib64/asterisk/modules in asterisk.conf and that’s it.

However I have noticed that in extensions_additional.conf there is a line that;s ASTMODDIR = /usr/lib/asterisk/modules

This should be lib64, not sure if this is an issue?

However, In the 32-bit 2.10 days, users were minimal, as the system has grown I have came across this problem, it may have been there the day I did a 32-bit FPBX restore to a 64-bit server, who knows.

With Asterisk coming back unable to replicate, I’m now stuck.

Is there anyone out there with a system that started life 32-bit and has upgraded to 64-bit that can test or, someone out there regardless of system config, can test this?

I can replicate every time guaranteed. Ring an extension and while recording a voicemail do a reload. You don’t need to change any config, just issue core reload at the CLI. Hang up when its done and see if an email gets sent.

I am currently 2.11 with all updates 64-bit Asterisk 11.13.1

I’m at my wits end with this and would appreciate some help.

If anyones interested the Asterisk ticket with debugs etc is here:

https://issues.asterisk.org/jira/browse/ASTERISK-24463

I have a 64-bit system with Asterisk 11.14.2 and FreePBX 12, and was unable to replicate your issue. I will say that the method you chose of upgrading is likely to break something since A) you went from 32 to 64-bit and B) you moved hardware and restored configs (less of a problem, but still a potential issue).

I see different IPs in the logs - Class A and C private IPs. Are the network settings correct under admin > sysadmin and settings >asterisk sip settings (and chan_sip page underneath this as well)?

I’m not sure how much, if any, custom programming is on this box, but could you export DIDs/EXTs and recreate the other settings? I’d do that and on the latest version of Asterisk 11 and FreePBX 12 as this would likely solve the issue. You could also try just updating to the latest Asterisk 11.

It’s a weird problem for sure, and it may be in your best interest to rebuild from the ground up :confused:

Thank you for responding.

It is crossing my mind that the 32 to 64 change did something however the current new hardware system was installed direct from the latest 64-bit ISO at the time. Surely I’m not the only one out there that has done it this way.

I then did a restore which (to my understanding) only restores databases and conf files and does not alter core asterisk files.

This 64-bit system has been in service for over a year and this is the only fault I currently have with it and with the majority of config all text based I’m hopeful (with some help) we can get to the bottom of it.

The different IP’s you are seeing would be the 10.x.x.x which is the primary range and the other is for VPN (rarely used at the moment)

Re-building is out of the question. I don’t have the hardware available to do this and no window for any downtime. It’s quite heavily programmed (not much custom) and would take a long time to get back to its current level.

I have managed to replicate this on a sandbox server which is using a clone of the backup (which also has the problem). This is a server sitting on its own with 2 phones completly independant from the rest of the network. If anyone has any suggesstions on what to try I can try it on this server, no worries if I break anything as I can just restore it.

Got an update on the situation.

I have found that when deleting vm_email.inc the email is sent perfect every time however restarting asterisk stops the voicemail module loading with this file missing.

What makes it even more wierd is having this file there but empty produces the fault.

The only way I can get it to work is to start Asterisk with everything as it should be then rename vm_email.inc to something else, it will then work fine. The file would then need to be renamed back for an asterisk restart

If I take the include out of voicemail.conf makes no difference. If i move emailBody to voicemail.conf makes no difference, the only fix so far is to rename the .inc

Asterisk have closed the bug pointing it to be a FreePBX issue

I have re-opened my original ticket and hoping that someone can chip in and help out.

What I also noticed and have now changed, was the ASTMODDIR in freepbx_settings in the asterisk SQL table. Changing that and restarting asterisk then issuing a core show globals is now showing the correct path.

I then applied the latest 5.211.65-21 patch taking me upto 11.14.1 along with a load of other updates but unfortunately its still the same.

It looks like I’m going to have to go down the paid support route on this one in the new year.

Thanks

http://issues.freepbx.org/browse/FREEPBX-8176

UPDATE - I have replicated this on a clean ISO from the website and have updated the ticket including a step by step on what I did.

This is the vm_email.inc file on my box with e-mail working, yours? I’m going to have to update the link to use the UCP :smiley:

; To be able to localize text in the email by using utf-8 characters you can remove the remark on the next line so that the charset will be utf-8
; charset=utf-8
; Change the email body, variables: VM_NAME, VM_DUR, VM_MSGNUM, VM_MAILBOX, VM_CALLERID, VM_DATE

emailbody=${VM_NAME},\n\nThere is a new voicemail in mailbox ${VM_MAILBOX}:\n\n\tFrom:\t${VM_CALLERID}\n\tLength:\t${VM_DUR} seconds\n\tDate:\t${VM_DATE}\n\nDial *98 to access your voicemail by phone.\nVisit http://AMPWEBADDRESS/recordings/index.php to check your voicemail with a web browser.\n

Hi Overkill

Sorry for the long delay, I didn’t get a notify that you replied.

It’s all getting very tiring now. I can replicate off a clean ISO off the website, 64 and 32 bit, FPBX 2.11 and 12 and even on the AsteriskNOW ISO. If I do a version change upto Asterisk 12 the problem goes away. I move back to 11 and it comes back every single time. I’ve posted this on FPBX tracker and asterisk and just get bounced back and fourth that it’s the other ones problem. Asterisk cannot replicate which does point to something within the FreePBX distro that’s causing it. I don’t want to move to 12 as 11 is the LTS.

If you look at the FreePBX bug report I have attached a step by step guide on what I did to replicate, on a clean install no backup/restore cross platforms etc etc

Is there anyone else out there that’s experiencing this? - I find it hard to believe it’s just me! - A lot of people might not even realise it’s happening as emails are dropped and with no non-deliverable messages being dropped by default then these could easily be missed on a large system.

My vm_email.inc is the same as yours. I think this file is a complete red herring but don’t understand how renaming it fixes the problem on a reload. I may have explained that wrong in the tickets which have caused asterisk and freepbx to misinterpret what I’m trying to say. I’m not a programmer, developer etc, just someone trying to use their program and I feel that I’m having to dig into the bare bones to get my point across without anyone actually doing an install and checking it out for themselves.

There is a problem there, I just need more people to chip in and try and replicate it with me.