Mac Does Stuff

Month: September 2017

  • Windows Update Errors (Insider Preview)

    Recently on some of my machines I’ve been getting errors not being able to update Insider Preview versions. I end up with ‘Updates could not be installed’ or similar. Using the Windows Update Fix Tool doesn’t seem to sort it out either. Anyway, I’ve found a process that seems to resolve the issues for me. Have done it a fair few times and it seems to work…So I thought I’d share the process.

    Stop the Services

    Start the Services plug-in – Services.msc (I’m assuming you’ll know how to do this – Windows key+R, enter services.msc).

    Find the Windows Update service – if it’s started, stop it. Set the service to ‘disabled’ for now.

    Services

    Next, fire up an elevated command prompt (I.e. Command from the Start menu, right-click, select run as administrator).

    Run As Admin

    From that command prompt, stop these services:

    net stop bits

    net stop appidsvc

    net stop cryptsvc

    You can copy/paste those in to the DOS prompt if you want.

    Clean up the SoftwareDistribution Folder

    Go in to your Windows folder, and find the ‘SoftwareDistribution’ folder. Rename it to ‘SoftwareDistribution.bak’. Note that if you get an error saying the folder is in use – double check to make sure you’ve stopped the Windows update service. Personally I just normally bin the whole folder – but then most of my units are virtual so I normally snapshot them before doing this.

    Clean up the CatRoot2 Folder

    Go in to your System32 folder in your Windows folder, and find the ‘CatRoot2’ folder. Rename it to ‘CatRoot2.bak’. Now, this can be fussy. This is often locked by the Cryptographic service (cryptsvc) and you’ll find that the CryptSvc often keep starting due to dependencies from RPC and the like. The trick is to wait until you get the ‘This folder is locked’ with the ‘Try again’ button, then type in the ‘net stop cryptsvc’ into the Command prompt, and as soon as it’s finished stopping, hit the try again button. This generally works. May take a few times though – it’s clearer what I mean by this in the video.

    Reset Security Descriptors

    Next, we’re going to reset the security descriptors for BITS and Windows Update. See here for the commands for that. Couldn’t paste them in here – keeps converting some of the stuff to Emoji, and I haven’t worked out how to stop it!

    Reset the Windows Update Service to Manual

    Finally, reset the Windows Update service to ‘Manual’ in the ‘services.msc’ consoler, and reboot.

    You should now be able to re-run Windows Update and hopefully all your updates apply. There’s a quick video of the run through below.

  • 1Gbps Fibre Internet Connectivity

    I live in Central London, and one of the benefits of living here is that the Internet connectivity has always been relatively good. I’ve had 200Mbps down/20+Mbps up for quite a while, and you get used to such performance. The upload speed always seems to be the restrictor though, doesn’t it? Makes things like uploading YouTube videos, and online backups (for example) something you set off, and leave to it. 

    Anyway, I got the opportunity to upgrade to 1Gbps Fibre connectivity from a company called HyperOptic. That’s 1Gbps up and down. I saw it was available, and given the cost at circa GBP50/month, I thought it was a no-brainer. I also kept my existing provider, as something so cutting edge may perhaps not be reliable. I tend to adopt stuff early, and sometimes that can cause some headaches – I couldn’t afford any issues with my Internet as I rely on it so much. So much so I’ve had more than one provider for as long as I can remember. In fact, I wrote about load-balancing multiple connections here:

    Load Balancing Broadband Connections

    So I’ve had the service for a good few months now, so thought I’d share my views. 

    My opinion is a simple one – just wow. It’s been rock-steady from a reliability point of view, and the performance just changes the way you use things. Uploading/download stuff feels much the same as it would as if I were operating over a local network. Just simple things like Windows Insider Preview updates come down in no time at all. It just makes my remote working life far, far easier.

    Online backups now are far more tenable than they were. I use BackBlaze, and I’ve found their service very fast, and utterly reliable. You add in a decent connection and all of a sudden I can now backup all of my important stuff and not give too much thought as to how much I backup off-site. I wonder with the advent of such Internet connections whether services like BackBlaze, who offer unlimited backup, may have to reconsider the price point/offer? Mozy went through that pain, and I believe Crashplan has now pulled out of the consumer market? I shall watch that space. I’ve a ton of cloud storage with my Office365 tenancy and my DropBox area so was considering jumping to that, but the BackBlaze service is just such a configure-and-forget platform it’s well worth the money. Backups you have to do sometimes just don’t happen. Automating those backups is the winner here.

    Anyway, performance. Below is the current performance I’m seeing using the DSL Reports Speedtest. I’m connected over 1Gbps wired-Ethernet at this point. 

    Wired

    Seriously, how impressive is that? For comparison, this is the same test but over WiFi.

    WiFi

    In this case, it’s my WiFi not keeping up with the connection – which is fair enough. I use a Netgear R8000 AC3200 router and am in general very happy with the performance.

    Anyways, want to see this stuff working? There are a couple of videos below you may find interesting. I’m incredibly satisfied with the service. Fantastic performance, and a very reasonable price – what more can you ask for?

    Oh – as a quick side note – if you do consider the service remember they use Carrier Grade NAT (CGN). CGN can cause issues with stuff like VPNs and the like. My work VPN won’t work through it for example. They can give you a static IP address too – I spoke to them online about it, and it was sorted in minutes.

  • Skype for Business inbound Ringtone

    I ran in to an interesting problem a few weeks ago with regards to callers to a Skype for Business platform. Some users were not getting the ‘ring ring’ when dialling in. I.e. The ringtone or ringback as it’s often called.

    What made it harder to locate was the fact that it wasn’t all users, and it wasn’t all of the time. I.e. It seemed random across a user DDI range, and for different callers. It took a lot of logging and reading.

    Here’s the thing. I never did spot a real instance of it actually happening. If you can’t see an event how do you trace it? 

    By a complete fluke I was using my Skype consumer client – and I was logged in to my Australian account, rather than my normal day to day. Guess what – no ringback tone. The experience is dial – hear nothing for a while – person answers.

    This was predictable and reproducible. Also found the issue dialling in from the US.

    This explains the randomness of the event, and made me feel happier about my log-reading skills. 

    So, the scenario is a SIP trunk terminating on a Sonus SBC, and a SIP trunk from the Sonus SBC to a Skype for Business mediation server.

    Investigating a failed call and one that worked however yielded exactly the same call behaviour. You see the 100 trying, the 183 with SDP…and the SIP conversation happens exactly the same way. So it can’t be our end then, right?

    So off to the carrier I go with a list of stuff that isn’t happening. They’re still investigating. 

    In the meantime, there is a way to force ringback on the carrier – I.e. Make sure the carrier is providing it. It’s fixed the issue for us in that all users now always get the ringtone/ringback or whatever you want to call it. So I thought I’d share how to do it – some people may find it useful. 

    Essentially we’re going to change the 183/SDP messages to the carrier to 180 ringing. You can see the full list of SIP response codes here. 

    So, how do we do it? Well, the Sonus can apply message translation rules to routes – so you can change one SIP message to another for incoming calls. In our case, we’re going to change 183’s to 180.

    Let’s have a look at how this was done.

    The first thing we’re going to do, is to define the translation in the ‘Telephony Mapping Tables’. You get to this in the ‘Settings’ part of the Sonus configuration:

    TMR

    Expand the ‘Message Translations’ section and add in a translation. In my configuration, the translation looks like this:

    MTR

    The important bit is the incoming message type, and the outgoing message type. We’re going to convert 183 Session Progress to 180 Ringing.

    Once you have set up the translation, you need to find your incoming route.

    Incoming Route

    We have multiple SIP trunks on this one, from two different providers. Select the one from your provider, and edit it. In there you’ll see the option to select your message translation.

    Edit Routes

    Once you apply it, you should see a change in behaviour on the inbound calls. An example from before the change is shown below – you’ll see the 183 conversation.

    Call with 183

    After you implement the change, you’ll see we send the 180 Ringing. This is causing the service provider to deliver the ringback to the calling party.

    Call with 180

    Now it could be that the service provide will nail why they’re having this behaviour when being called by certain countries – at which point I can take this configuration off.

    If you’re having issues with no ringback however, this brute force approach to asking your provider to deliver may give you a solution.