Archive for the ‘Troubleshooting’ Category

trixbox 2.4 kernel-pae

Saturday, January 26th, 2008

If Anaconda (the CentOS installer) detects a PAE compatible processor it will install kernel and kernel-pae

kernel-PAE (x86-only): Contains a kernel with support for 64GB of memory on x86 systems through Physical Address Extensions (PAE). Only x86 CPUs that support PAE can run this kernel. This kernel is provided because x86 systems can’t address more than 4GB of memory without PAE extensions.

trixbox is package based and everything wraps around modules in a specific kernel.

If your system gets the pae kernel you can correct this by editing /etc/grub.conf

you should see something to the effect of :
# grub.conf generated by anaconda
#
# Note that you do not have to rerun grub after making changes to this file
# NOTICE: You have a /boot partition. This means that
# all kernel and initrd paths are relative to /boot/, eg.

default=0
timeout=5
splashimage=(hd0,0)/grub/splash.xpm.gz
#hiddenmenu
title trixbox (2.6.18-53.1.4.el5PAE)
root (hd0,0)
kernel /vmlinuz-2.6.18-53.1.4.el5PAE ro root=LABEL=/1 acpi=off
initrd /initrd-2.6.18-53.1.4.el5PAE.img
title trixbox-base (2.6.18-53.1.4.el5)
root (hd0,0)
kernel /vmlinuz-2.6.18-53.1.4.el5 ro root=LABEL=/1 acpi=off
initrd /initrd-2.6.18-53.1.4.el5.img

Simply use # to comment out:
#title trixbox (2.6.18-53.1.4.el5PAE)
#root (hd0,0)
#kernel /vmlinuz-2.6.18-53.1.4.el5PAE ro root=LABEL=/1 acpi=off
#initrd /initrd-2.6.18-53.1.4.el5PAE.img

Now you can ensure it will boot to the correct image.

2.6.18+ kernel users

Wednesday, December 26th, 2007

I have put together a modprobe for users of hot-plugging kernels like 2.6.18 that use modprobe.

(more…)

trixbox CE 2.2 zaptel delemma

Wednesday, December 26th, 2007

A trixbox user has done a full post on preventing issues seen with zaptel when you run yum update in 2.2 while booted to a non smp kernel. Basically add to your the trixbox yum repo file a exclude=*smp. A full explanation is at http://blogx.co.uk/Comments.asp?Entry=679 . Please note his article is for a specific issue but covers a broader scope.

yum update -y

Monday, December 10th, 2007

FYI all trixbox CE users

Just to give everyone a heads up from the Rhino’s mouth. Rhino products are proud to continually work with trixbox CE and future trixbox distributions. We also take great pride in our customers and realize that the more information that is posted the better informed we all are. That is why we take it very seriously to keep our customers informed of any ongoing issues with our hardware and distributions that we recommend.

The following issue has been popping up all over so please be aware of what you are installing at all times.

Please note that if you do a yum update for any of the following systems we cannot support your system until the system has been completely restored to factory settings.

trixbox CE 2.2

trixbox CE 2.2.1

trixbox CE 2.2.2

trixbox CE 2.2.3

trixbox CE 2.2.4

trixbox CE 2.2.5

trixbox CE 2.2.6

trixbox CE 2.2.7

trixbox CE 2.2.8

trixbox CE 2.2.9

This includes any future releases on the trixbox CE 2.2.x branch

This is due to errors in the trixbox repositories and this leads to corruption to your trixbox system.

The repositories for the upcoming trixbox 2.4 will be updated from what our trixbox sources have said, but for the trixbox 2.2.x repository this is not something that they can currently fix.

HDLC and FCS abort errors with Rhino T1/E1 Cards and PRI

Tuesday, November 20th, 2007

If you are seeing HDLC and FCS aborts in your logs. This is could be due to one of the following reasons:

1:

Make sure that your T1/E1 cards are at the following revisions and have the apropriate drivers installed on your system.

R1T1 -EC and Non-EC models must be at Hardware version 37 or above

R2T1 and R4T1 must be at Hardware version 36 or above

Driver revision must be at 2.0 and above.

If the above criteria is not met then you may see the aborts in your logs.

2:

If you are using CentOS4 and are using SATA drives you may see this issue.

You must use a kernel greater than 2.6.9.

By upgrading to an asterisk distribution with an updated kernel such as CentOS5 will alleviate these issues.

Please try one of the following systems:

Elastix

trixbox CE 2.4 (Beta for Cent 5)

RhinOSterisk (Alpha)

Known Issues:

Tuesday, October 23rd, 2007

As of last testing we have observed the following issues. This list is as of 10/23/2007. Please note user experience may vary and some of these issues may have been resolved in later builds and only apply to the versions listed.

trixbox CE

2.2.3 - genzaptelconfig does not recognize our cards.

2.2.x> the 2.6.9 kernel does not properly implement apic and has poor IRQ managemen. Use of multiple high interrupt devices like Digital Cards in combination with each other, extra network devices or RAID will cause interrupt issues and instability.

2.2.4 Zaptel does not install. You can run yum -y install zaptel-*

Digital cards shipped after 10/05/2007 require a new driver use the 2.2.1 package in the beta directory of our ftp. If you recieved a digital card in a ceros this update has been made.

Zaptel version 1.4.2.1 Does not build correctly. This seems isolated to this version

Some Qwest and Ameritech lines are not sending voltages across correctly. This issue may also occur in other AT&T regions. This issue can be overcome by upgrading your rcbfx firmware to 1.11.

Rhinosterisk 70021008-0917may not have an extensions.conf. You can correct this by rerunning install_amp

Some netgear routers will randomly reboot with qos and vlan enabled causing SIP delays.

zaptel 1.4.2.1

Tuesday, September 25th, 2007

Zaptel 1.4.2.1 Does not compile correctly and is missing calls required by our drivers. This has been observed in this version and appears to be resolved in other versions. If you have the kernel headers, and zaptel installed and linked per our instructions you should be able to simply make and make install. If the dependencies are met and the build fails try a higher or lower version of zaptel and it should clear up.

Known Issues in trixBox CE 2.2.x

Thursday, September 20th, 2007

So there are some issues that are in some 2.2.x builds and they aren’t being fixed so here are the work arounds.

(more…)

Firewall ports and Port Forwarding

Friday, June 15th, 2007

You may wish to open some or all of the ports on your firewall depending on your situation. Some routers do not allow allot of pinholes and may crash or reboot repeatedly, if this occurs contact your router maker for updated firmware. You do not need to open all of these ports only the ones you use. There are 2 types of firewalls, software and hardware. You may have to open these ports and properly forward them in all devices and programs. Many situations have 1 or more hardware firewalls and a software firewall. If after opening these ports you still have issues try placing your system in to the DMZ of the firewall and see if the issue resolves. It is recommended to open only the ports needed as opening to0 many could put you at risk of attack.

Protocol Port(range) UDP/TCP Notes
SSH 22 TCP open for remote secure shell access
HTTP 80 TCP open for remote http access you may also need 143 open
HTTPS 443 TCP open for secure http access you may also need 80 open
SIP 5060 UDP/TCP  
IAX 5036 UDP  
IAX2 4569 UDP  
RTP 10000-20000 UDP  
MGP 2727 UDP  

For more information on Port Forwarding please visit the following URL:

http://www.portforward.com/english/routers/port_forwarding/routerindex.htm

Fake Supervision

Wednesday, May 30th, 2007

If your telco does not provide far end disconnect supervision or if they just dont do it well you may need to “fake” supervision in order for your lines to disconnect properly. You may fake supervision by adding a few lines to your zapata.conf .

Add the following to your /etc/asterisk/zapata.conf

callprogress=yes
busydetect=yes

After doing this restart asterisk. There may still be a disconnect delay but it will be reduced.