[Home]

Summary:ASTERISK-05042: inserting of install commands in modprobe.conf to automatically run ztcfg considered harmful
Reporter:Steve Davies . (stevedavies)Labels:
Date Opened:2005-09-10 05:08:44Date Closed:2005-10-04 19:35:21
Priority:MinorRegression?No
Status:Closed/CompleteComponents:Core/General
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:
Description:
The junghanns quadBRI board's driver breaks if ztcfg is used more than once after the driver is loaded.
You don't see any errors, the board just simply plays dead from Asterisk.  I for one don't use Klaus' boards regularly, so it took quite some time to eventually find the cause, which was that the zaptel Makefile adds lines to modprobe.conf to automatically and silently run ztcfg without telling you.

This is half-baked and positively harmful as I describe.

Why half baked:

 1) many systems generate modprobe.conf from a modprobe.d, so the commands are likely to be lost anyway
 2) if you are loading multiple modules (wctdm and wct4xxp for instance), ztcfg will be called multiple times and will fail until all are present due to the zaptel.conf mismatch.

So Klaus should fix his problem, but I really would like to see this "feature" of the zaptel install removed or made optional.
Comments:By: Roy Sigurd Karlsbakk (rkarlsba) 2005-09-11 02:01:41

IMHO, Klaus should fix this issue. This is probably just a small driver issue...
By: Kevin P. Fleming (kpfleming) 2005-09-13 22:08:20

If someone can come up with a better solution that will work on all distros (if it's an initscript, it's got to work everywhere) I'm all for it, but we need to run this command as soon as we have loaded the drivers for all boards in the system.
By: Matt O'Gorman (mogorman) 2005-10-04 10:45:44

Seeing as the bug is most likely in the driver for the BRI card and those drivers are not maintained here I opt to close this bug.

Matt
By: Kevin P. Fleming (kpfleming) 2005-10-04 19:35:01

ztcfg has been modified to not touch channels whose configuration has not changed since the last time it was run, so that will likely be a reasonable workaround for this problem.