Back in 2014 we installed an inground fibreglass pool and illuminated it with Pentair GloBrites. The vibrant colour from these LEDs highlights the glass resin and creates a magical scene. Unfortunately, the product themselves are prone to a high failure rate. As of 2025, we have gone through 3 sets at over $1500 per set. The first set was covered under warranty, unfortunately the next sets weren’t. I’m also fortunate that I can change the lights myself so we’re able to save on labour but it is still a huge investment for 2 lights.
This spring we unwrapped the pool after its winter slumber and found that another GloBrite has failed. Shocked? No. Disappointed, yes. I researched replacements and have heard relatively good things about the Pentair MicroBrites. I purchased two new MicroBrites as well as the corresponding adapters to fit MicroBrites in to a GloBrite housing. The parts are: “Pentair Adapter Microbrite To Globrite – 618040” and “Pentair EC-620425 MicroBrite Color LED Light 100′”.
For anyone worried if there’s a noticeable difference if you mix-and-match the GloBrites with the MicroBrites, there isn’t. Physically the lights are smaller and the power cable is thinner, but the output matches the GloBrites. I took this picture to show the similarities between the two when using a fixed colour. I did not run a colour-swim.
Hard to tell, but the light on the left is the GloBrite and the light on the right is MicroBrite. The lights in the pool are very close to the lanterns on the patio.
From this angle, albeit hard to see, the MicroBrite is on the left and the GloBrite is on the right. The Globrite is beside the yellow-white floatie. For anyone that is worried about the difference between the two, when using this colour, we’re unable to distinguish the difference.
Hope that helps someone.
A client utilizes Jump Desktop on Mac to access their local Windows servers via RemoteDesktop protocol, port 3389. When trying to connect to their servers they would see the error:
There was an error connecting to the computer. Error Code: 2.50 (0x32)
The client is able to connect to remote/external Remote Desktop servers without issue. After working on the Windows hosts it appeared that they would accept normal Windows RDP connections, steering the issue back to the Mac that is connecting.
After some research it was found that somewhere (MacOS 15.4?) the network security settings for Jump Desktop were changed which prevented Jump Desktop from accessing local servers.
On MacOS, settings, privacy & security, local network, be sure to enable Jump Desktop.
Listening to the needs of customers, Gilson Technologies now offers temporary rental power options in both clean battery and gas-generator technologies.
Battery powered is limited by consumption and is available in 1 kw, 2kw and 11kw options. To calculate need, 1 kw battery will provide 1 kw of power for 1 hour.
Delivery is available in South Western Ontario, Canada for a fee. Please inquire with your needs for proper pricing.
Plugging away on a project where I need to reference the C++ SDK for AWS. I’m working on a small t3a.micro instance and this is how how the commands went:
~/sdk_build$ cmake ../aws-sdk-cpp -DCMAKE_BUILD_TYPE=Debug -DCMAKE_PREFIX_PATH=/usr/local/ -DCMAKE_INSTALL_PREFIX=/usr/local/ -DBUILD_ONLY=”s3″
CMake Warning at CMakeLists.txt:9 (message):
In 1.11 releases, we are releasing experimental alternative building
mode.By setting -DLEGACY_BUILD=OFF you can test our advances in modern
CMake building and provide early feedback. The legacy support is set by
default in 1.11, when you complete build updating scripts please update the
— Performing Test HAVE_ATOMICS_WITHOUT_LIBATOMIC – Success
CMake Error at CMakeLists.txt:225 (include):
include could not find requested file:
AwsFindPackage
[…]
CMake Error at crt/aws-crt-cpp/CMakeLists.txt:140 (include):
include could not find requested file:
AwsSharedLibSetup
CMake Error at crt/aws-crt-cpp/CMakeLists.txt:197 (include):
include could not find requested file:
AwsCheckHeaders
CMake Error at crt/aws-crt-cpp/CMakeLists.txt:198 (aws_check_headers):
Unknown CMake command “aws_check_headers”.
— Configuring incomplete, errors occurred!
Hmm, that seems odd. I have a note on what commands to run and these commands ran fine on another server only yesterday. I removed the cloned aws-cdk-cpp folder and tried again. Same thing. Maybe I’m missing something else?
I started stepping through my documents and noticed:
:~$ git clone –recurse-submodules https://github.com/aws/aws-sdk-cpp
Cloning into ‘aws-sdk-cpp’…
remote: Enumerating objects: 1081754, done.
remote: Counting objects: 100% (1215/1215), done.
remote: Compressing objects: 100% (630/630), done.
remote: Total 1081754 (delta 609), reused 772 (delta 411), pack-reused 1080539 (from 3)
Receiving objects: 100% (1081754/1081754), 633.97 MiB | 19.19 MiB/s, done.
Resolving deltas: 100% (857146/857146), done.
Updating files: 100% (135908/135908), done.
Submodule ‘aws-common-runtime/aws-crt-cpp’ (https://github.com/awslabs/aws-crt-cpp.git) registered for path ‘crt/aws-crt-cpp’
Cloning into ‘/home/ubuntu/aws-sdk-cpp/crt/aws-crt-cpp’…
remote: Enumerating objects: 11729, done.
Receiving objects: 100% (112048/112048), 577.64 MiB | 24.84 MiB/s, done.
fatal: fetch-pack: invalid index-pack output
fatal: clone of ‘https://github.com/awslabs/aws-lc.git’ into submodule path ‘/home/ubuntu/aws-sdk-cpp/crt/aws-crt-cpp/crt/aws-lc’ failed
Failed to clone ‘crt/aws-lc’ a second time, aborting
fatal: Failed to recurse into submodule path ‘crt/aws-crt-cpp’
Ha! I hadn’t noticed that before. The git clone failure doesn’t show up in red and I was busy doing other things and didn’t notice the failure. The next command that I would run would be the cmake above, so of course it failed because the git clone failed. Now how to fix this?
Based on a memory related comment here, I decided to increased my AWS EC2 instance size from a t3a.micro to a t3a.small. And that was it. The git clone now works fine and everything compiled.
Check here for odds and ends that we’ll be selling. Please ignore the mess, it’s a work in progress.
Not something I would normally post about, but I’ve gone through several Sonicare toothbrushes (toothbrushii? teethbrush?) over the past few years. They always fail the same way; the metal rod that connects to the toothbrush head seems to break off inside the unit and creates a horrible buzzing noise. When the last unit failed I took it apart to see where / why the unit failed and discovered that a simple small screw had come loose. Of course at that time I had broken the toothbrush so badly that I couldn’t fix it.
Fast forward 6 months later and a new brush did the same thing. There are videos floating around on YouTube that explain how to get this toothbrush apart. Once it’s apart the internals look like:

This screw here:

Needs to be tightened. That’s it!
I was having an issue where I’d send an email from my work account to my personal account and when the email arrived in Outlook it would contain the email signature, but on my iPhone the signature would be missing.
I created a new signature in CodeTwo’s Email Signature platform and set a plaintext version of it to be simply:
“Professional Webinar Attendee | My Work”
I then sent plaintext from my work account laptop to my personal account. What arrived in Mac Mail / Outlook Mail is the plaintext email with signature. What arrived on my iPhone was again missing the footer. I then resent the message from my work laptop to my personal account and I manually typed the words, “Professional Webinar Attendee | My Work” as the signature in the outgoing email. What arrived in my inbox on Mac Mail and Outlook Mail was:
“Professional Webinar Attendee | My Work
Professional Webinar Attendee | My Work”
On my iPhone I only had:
“Professional Webinar Attendee | My Work”
So I figured that there’s some encoding change that is happening between the signature that I manually typed and the signature that is added by CodeTwo. Luckily I host my personal email on a cPanel server, so I grabbed the email file from my cPanel server and compared the email in NotePad++. There is absolutely no encoding change in the plaintext. It literally is:
“Professional Webinar Attendee | My Work<LF>
<LF>
Professional Webinar Attendee | My Work<LF>”
Well this doesn’t make sense. How does the iPhone know that the second line doesn’t belong? The only way it could know that would be if it compared the incoming email to the outgoing email. Turns out that both accounts (work and personal) are installed on my phone. I removed the work email account and powered off my phone. 10 steamboats later I powered it back on and sent a new test from my work account on my laptop to my personal account. TWO signatures as it should be on my iPhone! I added my work account back on to my phone and sent a new test. ONE signature, same problem as before.
I had some other work users send test email to my personal account with the work email still active on my iPhone. Signature shows up fine. I send from my work account to my personal, signature is missing again.
I truly believe the iPhone is comparing email from sent to received and only showing me the sent email. I’d bet it’s doing that for “efficiency” so that it uses half the storage space on emails that are sent to itself on the same device. Probably Apple’s logic.
CodeTwo worked fine. It was me all along.
Had a good four-hour fight trying to remove a domain from an Office365 Tenant belonging to a client. The tenant has 9 domains that are entwined with various email addresses for various employees. The client sold 3 businesses and is working with the purchaser to migrate the domains and email to their own instance. Total, 81 users were being migrated off.
All of Microsoft’s instructions show that it’s a 2-minute job to sign in to the admin portal, go to settings, domains and simply remove the domain. But we all know life with Microsoft isn’t that easy.
Using the Settings, Domains, (domain-to-be-released), Users field, I could see the users that needed to have their domains changed to the <tenant>.onmicrosoft.com domain. Click a bunch of check boxes, remove domain, and some of the users would disappear, but I was stuck with 25 users that I could not release.
Went through each user account and found peculiarities that I cleaned up. Email address’ that could not be converted from the domain-to-be-released because a similar email address existed on another domain in the same tenant. Think sales@contoso to sales@<…>. Manually corrected those but still could not release the domain.
Working with some Powershell scripts, I was able to find 25 users that still had the domain in their accounts but it was under the read-only property IMAddresses (or IM Addresses).
Using the Powershell command:
Get-MsolUser –DomainName <domain-to-be-released>
I was given a list of the 25 users. Looking closer at the list, I noticed that those 25 users were all Exchange Online (Plan 1). Looking around, Reddit popped up this gem:
Manually worked through those 25 accounts and was able to release the domains. Hopefully this helps someone else.
Added two new meross MSS510 smart switches to my own home as I’m enjoying some simple home automation. Wiring was simple and the connection to my home automation system was seamless. Both switches joined my wifi without any issue — but afterwards I realized that I added them to the wrong SSID as I have two SSIDs. Simple enough, I’ll reset and re-add them to the proper SSID.
Reset both switches and added the first switch to the proper SSID. Easy. The second switch though, “could not connect to network”. Ugh. Weird, because they’re in them same junction box.
Reset the switch multiple times.
Restart my phone multiple times.
Turned the breaker off and on again. Nothing.
Checked the firewall DHCP leases. Both switches are shown based on their MAC addresses. Both switches are also on the proper SSID, but the switch that isn’t working has an IP address from the previous SSID, not the current SSID. Again, weird.
Reset the switch multiple times.
Restart my phone multiple times.
Turned the breaker off and on again. Nothing.
Wait a second. What’s the scope of the DHCP leases on the proper SSID? Yup, that’s the issue. When I previously defined my DHCP lease scope in the firewall, I had limited it to 20 IPs. This device, on that new SSID, would be device 21. Adjusted the scope to 25 devices, reset the switch and now we’re up and running. Switch worked fine all along. It was a firewall issue in the end.

We’ve recently acquired a 3D printer for rapid prototyping and short run production.
Offering over 16 different filament colours, we’re able to fulfill most orders complete with multi-colour jobs. Prices are negotiable with factors, such as; print time and amount of filament used per print. Volume discount available for multi-piece orders. No extra charges if print time is longer than expected. Quickest turn around for supplied .stl files.
Maximum print size per print is 25cm x 25cm x 25cm.
Also offering 3D object scanning of most objects or people for a physical model!
Recent Comments