asterisk users - Aug 2023 - Segmentation fault

Hi Federico,

Segfaults are 100% not by design.  Typically if something seg faulted, 
either there is a logic bug or a component mismatch. The you should 
definitely be able to use more than one connection (we use multiple 
connections with postgres odbc with no issue).

If Asterisk segfaults when using odbc
Try this:
- use ps and get the pid of Asterisk
- run gdb, attach to the asterisk pid
- do something that would cause the seg fault
- get a backtrace (bt) and show all threads backtrace (thread apply all bt)

if Asterisk segfaults when starting up
Run Asterisk straight from gdb
Wait for segfault, get backtrace, and all threads backtrace




On 8/16/23 18:48, Federico wrote:
>
> I tested this issue with version 13 and version 18.
>
> In res_odbc.conf, if I add a second, new data source like
>
> [asterisk]
>
> enabled=yes
>
> dsn=asterisk
>
> sanitysql => select 1
>
> isolation => read_committed
>
> username=root
>
> ;password>
> pre-connect => yes
>
> forcecommit => yes
>
> connect_timeout => 10
>
> negative_connection_cache => 0
>
> max_connections =>500
>
> my odbc.ini
>
> [cdr]
>
> Description = MySQL ODBC Driver Testing
>
> Driver = maria
>
> Socket = /var/run/mysqld/mysqld.sock
>
> User = root
>
> Password >
> Database = public
>
> Option = 3
>
> I  get, immediately, segmentation fault.
>
> With only one, it works fine.
>
> Is this by design?
>
> Philip
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.digium.com/pipermail/asterisk-users/attachments/20230819/04197393/attachment.html>

I cannot follow your instructions, because asterisk segfaults on start. It never
starts

Can you give me instruction to trap this segfault on starting asterisk?

Like gdb …..asterist –gvvvvvvc 

 

 

From: asterisk-users <asterisk-users-bounces at lists.digium.com> On
Behalf Of Mark Murawski
Sent: Saturday, August 19, 2023 11:04 PM
To: asterisk-users at lists.digium.com
Subject: Re: [asterisk-users] Segmentation fault

 

Hi Federico,

Segfaults are 100% not by design.  Typically if something seg faulted, either
there is a logic bug or a component mismatch. The you should definitely be able
to use more than one connection (we use multiple connections with postgres odbc
with no issue).

If Asterisk segfaults when using odbc
Try this:
- use ps and get the pid of Asterisk
- run gdb, attach to the asterisk pid
- do something that would cause the seg fault
- get a backtrace (bt) and show all threads backtrace (thread apply all bt)

if Asterisk segfaults when starting up
Run Asterisk straight from gdb
Wait for segfault, get backtrace, and all threads backtrace




On 8/16/23 18:48, Federico wrote:

I tested this issue with version 13 and version 18.

In res_odbc.conf, if I add a second, new data source like

 

[asterisk]

enabled=yes

dsn=asterisk

sanitysql => select 1

isolation => read_committed

username=root

;password
pre-connect => yes

forcecommit => yes

connect_timeout => 10

negative_connection_cache => 0

max_connections =>500

 

my odbc.ini

[cdr]

Description = MySQL ODBC Driver Testing

Driver = maria

Socket = /var/run/mysqld/mysqld.sock

User = root

Password 
Database = public

Option = 3

 

 

I  get, immediately, segmentation fault.

With only one, it works fine.

Is this by design?

 

Philip

 

 

 





 

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.digium.com/pipermail/asterisk-users/attachments/20230820/25ec460a/attachment.html>