[asterisk-bugs] [LibPRI 0014947]: [patch] Changes in LibPRI 1.4.10 cause E1 to drop during Siemens Realitis CPU switchover

[Asterisk Bug Tracker]

A NOTE has been added to this issue. 
====================================================================== 
https://issues.asterisk.org/view.php?id=14947 
====================================================================== 
Reported By:                shawkris
Assigned To:                mattf
====================================================================== 
Project:                    LibPRI
Issue ID:                   14947
Category:                   General
Reproducibility:            sometimes
Severity:                   minor
Priority:                   normal
Status:                     assigned
Asterisk Version:           1.4.24 
JIRA:                        
SVN Branch (only for SVN checkouts, not tarball releases): N/A 
SVN Revision (number only!):  
Disclaimer on File?:        N/A 
Request Review:              
====================================================================== 
Date Submitted:             2009-04-21 18:21 CDT
Last Modified:              2009-10-16 10:38 CDT
====================================================================== 
Summary:                    [patch] Changes in LibPRI 1.4.10 cause E1 to drop
during Siemens Realitis CPU switchover
Description: 
Asterisk is connected to a Siemens Realitis PBX via a Sangoma A102. The
Realitis contains two CPUs running active/passive but swaps the roles over
each night at midnight. The system can maintain calls whilst the switchover
is in progress.

Up until LibPRI 1.4.9, this CPU switchover went unnoticed by Asterisk.
However, since 1.4.10 Asterisk drops the E1 during the Siemens CPU
switchover.
====================================================================== 

---------------------------------------------------------------------- 
 (0112370) frawd (reporter) - 2009-10-16 10:38
 https://issues.asterisk.org/view.php?id=14947#c112370 
---------------------------------------------------------------------- 
I completely confirm this issue, happening for me when there is a small
link error and the operator sends a SABME, expecting all counters to be
reset.

Reading Q921, I think the reset should be done regardless of state (even
simpler), but AFTER having sent the Unnumbered Acknowledgement (UA), like
so:

/* Send Unnumbered Acknowledgement */
q921_send_ua(pri, h->u.p_f);
q921_reset(pri); //SHOULD BE HERE WHATEVER HAPPENS
return q921_dchannel_up(pri); 

Issue History 
Date Modified    Username       Field                    Change               
====================================================================== 
2009-10-16 10:38 frawd          Note Added: 0112370                          
======================================================================