Category Archives: Notepad

Issue installing pyinstaller

Recently while running:

pip install pyinstaller

I encountered the following error:

Failed to build pyinstaller
Skipping wheel build for altgraph, due to binaries being disabled for it.
Skipping wheel build for pefile, due to binaries being disabled for it.
ERROR: Could not build wheels for pyinstaller which use PEP 517 and cannot be installed directly

It turns out that Windows Defender (the built in AV in Windows 10) was blocking runw.exe

The real error was a few lines above:

error: could not open 'PyInstaller\bootloader\Windows-32bit\runw.exe': Invalid argument

A quick unblock from Windows Defender and it worked.

Thanks to: https://github.com/pyinstaller/pyinstaller/issues/3813#issuecomment-575319940 for the hint.

Adding Headset to LG Phones

Recently I added some headsets (Jabra PRO 9450) to a range of LG LIP-8012D and LG LIP-8024D phones.

  1. The electronic hook switch plugs into the back of the handset, with the coloured strip (usually indicating pin 1) facing up. It’s best to remove the power, add the accessory, then plug it back in, otherwise, the phone won’t recognize the accessory.
    A second cable for audio is required from the headset port on the back of the phone to the base use of the headset. (only control signals are sent down the electronic hook switch cable)
  2. The handset requires programming to be told to use the handset and not the speakerphone.
    Trans PGM > 6 > 1.
    Select 0 for headset then Hold/Save.
  3. The handset requires programming to be told to ring in the headset:
    Trans PGM > 6 > 2
    Select 2 for headset or 3 for both, then Hold/Save
  4. That’s about it. Enjoy

Upgrading requests_toolbelt in Zato 2.0.8

After failing with pip and easy_install I came across this post again:

https://forum.zato.io/t/what-is-the-canonical-way-to-add-extra-python-packages-to-my-zato-server/580?source_topic_id=586

There is another way, that of updating versions.cfg and buildout.cfg files but if you are more familiar with pip than buildout then pip install package-name is perfectly fine.

So it turns out it is as easy as:

nano ~/current/versions.cfg
update the version from 0.2.0 to 0.8.0 then run
cd ~/current
buildout

And buildout will do it’s magic.

Sorting 404s in Nginx access Logs

The default Nginx logs can be combined and sorted to gather some useful stats.

  1. Make a working directory and copy current logs over (optional):
    mkdir ~/accesslogs
    sudo cp /var/log/nginx/access.log* ~/accesslogs
    cd ~/accesslogs
  2. Extract logs compressed by log rotate:
    ungzip *.gz
  3. Make a big combined log:
    sudo cat access.log* >> log.combined
  4. Parse the combined log and count the 404’s
    sudo awk '($9 ~ /404/)' log.combined | awk '{print $7}' | sort | uniq -c | sort -rn >> 404sC.txt

Let’s encrypt certbot – moving vps

Sometimes you need to move VPS’s, and you don’t want downtime with the SSL certs.

Well it turns out that you can easily run Let’s Encrypt in standalone mode and use DNS verification to do a once-off cert prior to moving the VPS. – Read the Certbot docs.

But once the migration has taken place, you may also want to return to the validation method that you had been using previously. Two options exist, either a) modify the config files by hand (strongly advised against) or to return to the certbot and get it to update the config itself.

For example here is how I generally get a SSL cert on a particular box:

certbot certonly --webroot -w=/var/www/letsencrypt -d [www.URL.com] -d [URL.com]

And here is how I update to the new config

sudo certbot certonly --cert-name [www.URL.com] --force-renewal -a webroot -w=/var/www/letsencrypt

Thanks to @schoen on this post: https://community.letsencrypt.org/t/how-to-change-certbot-verification-method/56735/3

Zato: Add MSSQL to Outgoing SQL Options

To Connect with MSSQL or SQL Server via Zato 3:

#sudo su - zato
$/opt/zato/current/bin/pip install pymssql
(More info: http://pymssql.org/en/stable/intro.html)
$nano /opt/zato/env/qs-1/server1/config/repo/sql.conf

Add the following lines at the end:
[mssql+pymssql]
display_name=MSSQL
sqlalchemy_driver=mssql+pymssql
ping_query=SELECT 1

???I don’t know if this is required or not, but I still do it.
Deploy to other servers in the cluster:
$cp /opt/zato/env/qs-1/server1/config/repo/sql.conf /opt/zato/env/qs-1/server2/config/repo/

Restart your cluster and MSSQL should appear as an option under SQL Outgoing connections.

LG iPECS Least Cost Routing (LCR)

Recently I had a bunch of trouble with some SIP trunks (read about it at: LG IPECS SIP Trunk Issue) and needed to edit the Least Cost Routing (or Least Call Cost Routing).

The system is pretty basic, landline calls go out of one CO group, calls to mobiles go to another CO group.
There are a couple of tables that control how LCR works: LCR LDT (PGM 221) and LCR DMT (PGM 222).

LCR LDT – Leading Digit Table – PGM 221
This is where the ‘patterns’ are setup, so that the phone system knows what to do with the call, based on the starting digit(s).
Check out this screenshot for a full view:

These are setup under Index 1,
Then in subindex 0-6 are setup for landlines, 7 is setup for the mobile calls.
For each of the Subindex 0-6, it looks at DMT 01
For the Subindex 07 (Mobile calls), it looks at DMT 02

LCR DMT – Digit Modification Table – PGM 222
[To be honest, I have no idea what is actually happening here, except for the reference to the CO/IP Group].
Index 1 routes via CO/IP Group 1
Index 2 routes via CO/IP Group 2
– Index 2 also has an Alternative DMT Index (3) – This is used when all the lines in CO/IP Group 2 are busy
Index 3
– Routes via CO/IP Group 1.

Check out this screen for a full view:

Seems to work alright.

LG IPECS SIP Trunk Issue when WAN IP Address Changes (How to start to troubleshoot VoIP Issues with Wireshark)

Recently I had an issue with a couple of SIP trunks configured on an LG IPECs system when the WAN IP Address Changes.
TL;DR Version – I was getting a SIP2.0 502 Bad Gateway Error, turns out that the SIP provider (sip.iboss.com.au) had a setting that restricted the IP address that was being used to initiate the connection.

Basic Setup:
Router -> Switch <- IPECS The switch is a Cisco POE managed switch, with Voice Traffic on a VLAN. All the IPECS phones connect to the switch. The IPECS unit and the Router also connect into the switch. SIP trunks were setup on 2x of the CO lines. I was swapping the router from ADSL to NBN (vDSL). How to begin to troubleshoot the VoIP Connection issue. When I switched the router over (NBN), and tried dialling out on the phone they would say something like "Normal Call Clearing", and not ring. Yet, when I switched back to the original router (ADSL), it would work as normal. A lot of Googling suggested using Wireshark to inspect. How to start to troubleshoot VoIP Issues with Wireshark: Step 1) Install Wireshark on a machine Step 2) On one of the Switch ports, enable 'Mirror Mode'. On the Cisco web UI this was under Admin > Diagnostics > Port Mirroring. – Note: because VLANS were in use, I needed to mirror a VLAN, but first I had to remove the port from the VLAN, then choose that VLAN in the Port Mirroring Setup.
Step 3) Fire up Wireshark, choose the LAN connection that is connected to the Mirrored port and start the capture.

It’s always nice to see what something looks like when it ‘works’, so I first setup the ADSL router, hooked up wireshark and started the capture, I tried calling my mobile and answered it, then hung up, then stopped the capture.
Then in Wireshark you can click on ‘Telephony’ in the top menu and select SIP and will show you the connection, then click on ‘Prepare Filter’ and it will limit the main screen down to the packets that were involved in the phone call. Pretty Cool.
Closing the Telephony screen, if you right click on the first packet in the list (The Info column should say something like “Request: INVITE sip…….”), go down to Follow, then UDP Stream and a somewhat human-readable stream appears. I copied that screen out into Text File (notepad, notepad++, etc) for use later.

I then swapped over the router, did the same process as above and copied the new flow of data into another Text File.
I then compared the two files side by side using WinMerge.

The first part, looks ok. It’s natural that the IDs would change between calls.

Scroll down a bit and o dear, here is where we hit a problem.

From there it a was a case of trying various configurations. I swapped around routers (SIP ALG causes all sorts of issues with SIP lines), I tired 4G connections, finally I rang the VoIP provider – who last week told me that there wasn’t any IP address security restrictions – I was informed (I think by asking the right question and being blunt – “Are the VoIP Trunks tied to an IP Address?”) that yes they could change the configuration for me.

Side note: To make things a little easier in the testing, I copied the config from the iPECS into a Free Open Source Softphone – MicroSIP. I also had a secondary SIP account to test with a completely different provider, so my test process went something like this:
Bring up WAN connection on Router:
– Bring up Working secondary SIP account – call mobile. – Call connects and works. Router is working OK.
– Bring up Primary SIP account – call mobile. – Call works / fails.
– Change something (firewall rule, WAN connection, entire router)
– Repeat process.

By doing it this way, I have a ‘known’ test (Secondary SIP Account) that I can try, before trying the ‘unknown’ (Primary SIP Account)

Secondary Side Note: The LG iPECS has a field for “Firewall IP Address”, I believe that this is better described as “WAN IP Address”. It’s changed under ‘System&Device IP’ PGM 102 / PGM 103. It appears to use this in the SIP packets. I updated this before doing my WAN testing. Interestly, even with it set to the ‘Wrong’ IP address, things still worked?

Safe to say, I’m currently hating knowing this much about VoIP packets right about now…