John Galea's Blog

My blog on Gadgets and the like

Tessan 65W USBC/USBA charger

Ok, this is a bit of a boring subject, but I spent way to much time looking into this so I thought I would share my research. At this point, my laptops, work and home, and my work phone are charged by USB-C. The laptops needing a 65W charger. When you travel, such as too and from work, you need to take along chargers to support your day. If you need to charge an iPad/iPhone too … well that’s another charger. So I started looking into pocket friendly option. The stock charger from Lenovo comes in at 300g and is in no way pocket friendly.

So what I wanted I couldn’t find, I wanted something like this one, but could put out more like at least 85W, but I settled on this one from Tessan and bought it on Amazon at a price I could live with. So what it does is if only USB-C is plugged in, it pumps out the full 65W into the laptop, which means full speed charging. If you plug into the USB-C and either the second USB-C port or the USB-A port you get 45W into the USB-C for the laptop, which is more than enough for my laptops to maintain and even charge, albeit slower, and then 18W go into the other port. On the iPad mini 5th gen I was able to provide to max power from the USB-A port to the lightning, 5V-2A, which is also the max my iPhone XS can draw. So this power adapter does what I want, weighing in at a SVELT 108g, measuring a paltry 55x50x27mm with a folding AC clip this is VERY pocket friendly charger and perfect for travelling. It get’s warm when pushed but not unreasonably. I had to add the cables I needed as they were not included, but overall, I like it and small enough to be in the pocket!

April 17, 2022 Posted by | Uncategorized | Leave a comment

Time to upgrade my home Vmware ESX server

I’ve been using ESX in the home ever since I repurposed my Pfsense box … I started with an HP desktop, an Elite 8200, and slowly upgrade it, getting it to an i7 from i5 (which as it turns out is not all that useful since processors are rarely the gate, once you get spinning media out of the way, memory becomes the bottleneck) and upgrading it to it’s max of 32G. So many machines max out at 32 I got to the point I either had to start pairing down what I was running, add a second ESX server (something I didn’t want to do) or upgrade to a new server with more memory … At the same time, the machine I had was still running ESX 6.5 which hasn’t been updated since 2016 … an eterntiy.

So I started the process of what eventually became an obsession to get onto VMware ESX 7 to be able to continue to move forward. VMware have changed their approach on ESX, originally it only ran on select approved hardware, then they opened it up making it run on a LOT of different boxes, even if not supported, and then came ESX 7. VMware have intentionally, IMHO, refocused their attention on server grade hardware, inherently dropping support for things that previously worked. This come down to three items, first and foremost network card, then drive controllers and last processors, again with the focus on server grade hardware rather than desktops. The VMware compatibility list is the place to start. I narrowed in on NICs first and quickly found some inexpensive Broadcom NetXtreme 5720 dual port cards. These for some reason would not even allow my HP to power up, more nails in the coffin of this old workhorse. The second most reasonable VMware 7 supported card I found was 7MJH5 Intel Ethernet Server Adapter I350, but be careful this is a PCI 4X slot which is less common than the X1 of the BCM5720.

I looked at true server hardware like an IBM X3550, but most them are bulky, don’t support industry standard parts so repairs could get costly, are noisy, consume a LOT of power, and don’t have a lot of drive space in them. Memory upgrades for them is often a fortune. Shocking how little these go for these days. I looked at a couple workstations like the HP Z240, Dell Precision T3610 (careful the T3600 doesn’t support >64G) which support larger amounts of memory, and sport Xeon processors that VMware support … btw beware VMware does NOT support all Xeons … Ideally I wanted more than 32G of RAM and preferably 128 or greater. What you quickly run into is the only way to get there is with registered DIMMs that again are prohibitively expense. I also looked at Dell PowerEdge T320 but again you have to be VERY careful there are a lot of models under this same name and some of the earlier Xeon processors are simply 4 core and no hyper thread, I wanted more than the 4 core I already had. I did find a place called Delta server store in Scarborough that had a lot of this hardware used in stock, something that’s rare in these pandemic times. You will not believe the INORDINATE amount of time I spent looking at options only to find a catch here, or a catch there. I had become paranoid I’d buy something only to find it wouldn’t work. I probably spent almost 2 weeks on and off looking at options and ruling them out for one reason or another, too noisy, too much power, not enough memory, doesn’t support VMware etc.

The other places I looked at for refurbs was Laptops for less that I’ve dealt with a number of times, as well as Refurb IO.

I tripped over a couple of Asus motherboards based on the X99 chipset that get past the 32G barrier and then found a Asus PX9X79 LE on Facebook marketplace used, and I finally had a great solution to my desire to run VMware 7. Oh btw be careful there are two versions of this motherboard, the WS and the LE (Low End and Work Station I guess, with different specs). The motherboard has 8 DIMM slots, takes standard DDR3 memory, and can go up to 64 of memory. I’d hoped to get to 128, but the price I got this for $300 was a bargain. I double checked, the processor I was getting this with, the Xeon E5-2660V2 10 Core hyper thread IS supported on ESX 7, as is the chipset’s hard drive controller, the NIC wasn’t supported but that’s ok … I had the BCM5720 to put in it.

Loaded up ESX and low and behold it loaded perfectly well minus the usual processor may not be supported in the future warning.

I got a full 64G in and recognized, and was able to use some of the ram from the HP, and if I had to buy memory it was less expensive than server or workstation memory.

The 10 core Xeon is a VMware workhorse, I know eventually memory will again be my barrier, but for now I have 100% room to grow and I’m back on a supported version of VMware, in fact, I was no sooner done than I noticed VMware released an up to 7 … sheesh, can’t catch a break, missed it by days.

I was able to remove the drive where all the VMs were stored on ESX 6.5, in the HP, and just move it on over to the new machine, import the VMs and away I went. This part of the upgrade was shockingly easy.

I had to delete my VEEAM backups and start again but had no issues with VEEAM/VMware 7. Zabbix monitoring was again a bear to get running and I had to reset up all my dashboards as the host and even the VMs were like new to Zabbix, it did not see any of them as simply moved. The joys of NOT running virtual center I suppose.

I completed the upgrade by adding to the max memory the system can accomodate:

April 3, 2022 Posted by | Uncategorized | Leave a comment

Adding watermarks to your images using Lightrom

Almost all social media platforms remove EXIF tags which include owner and copyright information. To protect your images it’s important to add a watermark, a kind of digital signature to them. This does not guarantee anything, and honestly can be edited out unless it is obtrusively placed on the image, and in this case why bother posting? So to start off you can add a simple text watermark, but what I like is to create a graphic of your name. I used FontMe to do this. It takes a bit of playing with size and font preferences to get what you want, but in the end you end up with a graphic that is your signature and looks pretty … It gets downloaded as a PNG file.

Now within Lightroom desktop, once your done editing and ready to export your image scroll down to Watermarking. Here you point it at the image you just created, or you can just use text, click on the drag down box and then select edit watermark:

Now customize your watermark:

Then tell it where to put the watermark on the image (I choose a bottom corner to leave the image unobstructed), lastly save this new watermark preset so you can select it any time you export:

And here’s an example:

March 31, 2022 Posted by | Uncategorized | Leave a comment

Adobe Lightoom Classic vs Creative cloud edition

When I first started learning Lightroom, it was unclear to me the differences between these two product offerings from Adobe, but even more challenging IMHO was understanding the ramifications of choosing one product or the other. I’d like to save you that time …

There are two editions of Lightroom, and their interface, the commands you use and the features differs between them. I personally found Creative cloud edition easier, and more glossy to use, so I started with it, which is what precipitated this post. Desktop edition (also called Classic) has been around a LOT longer and so there are functions in Desktop that are either not yet in Creative cloud edition, or may never be. It’s important when you get started to avoid having to relearn, that you choose the right edition. I waffled between the two, landed on Creative Cloud edition, and then eventually found the error of my ways and had to totally relearn.

So … Creative cloud edition of Lightroom is designed, and intended to be used with Adobe’s cloud offering at the core. I had previously thought, oh well, I can use Creative Cloud edition and just disable syncing to the cloud. Because personally, I am not interested in paying Adobe to host for me, and I don’t want my images in the cloud where they could be stolen, lost or basically beyond my control. In fact, this is WRONG and created a ticking time bomb … You see what Lightroom Creative cloud does is create a cache of images you have edited on your machine and this cache is ONLY emptied by syncing with the cloud. If you disable sync to the cloud what happens is this cache grows, and grows until, next thing you know your hard drive is full. And if you want to move to a new machine, there is NO WAY to backup this cache and transfer it to a new machine. I spoke with Adobe and they confirmed this. So the net result is, if you don’t want to pay Adobe an increasing amount of money to store every image you’ve ever edited, and don’t want to have to manually delete images to constantly free up space on the cloud, then DON’T use Lightroom Creative Cloud edition, use Lightroom Desktop also called Classic. If you are also not planning on using multiple machines, say your desktop/tablet then again, there is no reason to use the Creative cloud edition.

Ok, so now let’s talk about using Lightroom Desktop/Classic. The interface is definitely a little more coarse less refined, but I have also found, once you get use to it, to be more efficient. Classic does NOT copy the original into it’s database so it’s important that wherever you import it into light room from be somewhere the photo is going to stay if you plan on ever editing or exporting it in the future. The good thing about this is the local database stays small and is manageable. I found I had to alter my photo workflow which I will talk about in a bit. Desktop edition can still work with the cloud if you so choose, but you can safely disable this and your photos NEVER go to the cloud, which is exactly what I wanted. Since moving back to desktop, aside from having to relearn, I have found that I could do in Creative Cloud edition that I can not do in desktop, unfortunately, at first, each and every thing you want to do you have to refigure out. The basic steps don’t change but how you do them does …

My photogaphy workflow now looks like this:

  1. Go out, take photos, have fun.
  2. Immediately copy the images off the memory card to avoid loosing anything. I put them on my file server where the data is on a mirrored drive (the data is written to two hard drives, so even the loss of one drive does not risk the data), as well as they are backed up nightly to protect my in process pics. I put the images in a directory with the date and name of the place where they were taken in the directory name.
  3. Tagging images is really important, so the first thing I do is take care of tagging the images, this includes adding owner information, adding GPS location if available (my camera doesn’t support geolocating images) and for my Sigma lens I add the tag because this isn’t properly handled. I’ve scripted all this tagging and the work is done by EXIFTOOL.
  4. One of the more important steps is sorting … not every pic is worth keeping, but for me, few are so bad to delete, so we move the photos we are going to keep/edit/publish into a directory of the same name below the original directory name
  5. I publish my pictures using two self hosted apps, Photoshow, and Photoprism. Each has there strengths and weaknesses, so I use Photoprism for my friends (downloading is disabled to protect content) and Photoshow is behind a password protected web site. So to get ready for editing I copy the saved images into the photoshow directory. For images that were not selected to be saved, I then archive the CR2s. I keep the CR2s because they are much better for editing but they are 4X the size of a JPG so to conserve disk space I move the not selected CR2 into an archive for later deleting.
  6. My editor of choice, as this post says, is Adobe Lightroom desktop edition. I have to say, I love how well this program works, but HATE the subscription model. At this point I can’t imagine not using it.
  7. Publish the photos to photoprism
  8. Last step is to archive all photos, saved and not to a permanent location that is not online, is backed up and I even go so far as to have them backed up offsite for true paranoia. My backup of choice by the way is AEOMI backupper which took me some time to choose, after using Windows backup for a LONG TIME.
  9. I also have scripts that create linked lists based on the items the photo is (which is done by keyword tagging), as well as when/where it was taken.

Editing steps within Lightoom:

  1. Import photos
  2. For my Canon/Sigma lens combination I add Lens Corrections, Enable profile corrections from within Develop. What this does is compensate for the known lens aborrations
  3. For each photo I start with cropping the image
  4. Adjust shadows, exposure, contrast, highlights etc within Develop
  5. Add a keyword to the image. This keyword is SUPER important and allows searching, lists of your subjects etc
  6. Exporting with a watermark for sharing with email blasts and social media. Since all EXIF tags are removed from images on MOST social media watermarks protect your content. I also export in a much smaller size to make them more shareable on email for example and to again protect content.
  7. Exporting saved images at max resolution, without watermarks for copying back to the original directory and publishing.

March 28, 2022 Posted by | Uncategorized | Leave a comment

Lenovo L14 Gen 2 review

Ok so after giving up on Dell’s 2 in 1 Inspiron 14 and Lenovo’s Yoga 7 I decided it was time to just get a typical Lenovo Corporate laptop like I’m use to. It will give me back the trackpoint that I love, and I found a rare model that uses SODIMMs for memory expansion, so I can upgrade the memory if needed all the way up to 64G, which hopefully will give it even more longevity. I again chose to go with the Core i7, specifically 11th Generation Intel® Core™ i7-1165G7 Processor (same one in both the Yoga and the Inspiron 14). I had issues finding the model I want, this was yet another model that Lenovo is having supply chain issues on and delivery was 3 months. So I did a search and found CDW had one in stock, slightly different model, but close enough. There was a price premium over the one on the Lenovo site of $200 which the only differences I could see were an Intel wireless card (instead of a Realtek) as well as a slightly better display at 300nits vs 250. The specific model I bought MT20X1 more specifically 20X100G6US. CDW said they would ship by end of day, and by paying a little more for shipping I would get it the next day by 10:30AM. I was skeptical, but I went ahead and sure enough it arrived by 9:30, impressive. I did find another reseller that had the model, cheaper, but when I did a google search on them there was enough noise that told me to stay away from them and pay the premium … buyer beware. So with that, let’s dig into this laptop.

Port wise, being a full size notebook this has a LOT more ports, full sized HDMI, 1 USBC, 1 USBC PD thunderbolt, 1 USB3 always on, 1 USB 3, a microSD slot (I’m disappointed it’s not an SD slot, means I have to carry around an adapter still) and a Kensington lock slot.

I tried a couple USBC docking stations/hubs and finally landed on a reasonably priced one from a no namer called Intpw on Amazon. I had issues with the network performance on the USB Realtek network card that is in most of these USBC docks. This one uses the AX88179 chipset and the network is rock solid. Power delivery on this also works perfectly as does the HDMI. The SD slot is rough getting in and out of, feels like you could break the card, how hard is it to design a SD card slot?

Processor wise, as mentioned 11th Generation Intel® Core™ i7-1165G7 Processor, which is a quad core hyper threaded, compared to the T450s which is a dual core hyper threaded so lots of processing power. The GPU is also a lot more modern, but unless you play games you wont see much use of the GPU. My World of Tanks games uses it for 3D, but Adobe Lightroom which apparently supports GPUs barely touches it, like <5% according to task manager.

The Wifi 6 adapter is an Intel AX210 which was an upgrade as far as I could tell from the Realtek. Testing shows the antenna/adapter in this laptop are not as good as the Yoga, but on better than the T450s I’m coming from. It is worth noting, and I have to admit I’m surprised there is NO wired NIC in this laptop. I believe you can get a WAN card for it as well.

2.5GHZ in close proximity:
[  4]   0.00-10.00  sec  89.6 MBytes  75.2 Mbits/sec                  sender
[  4]   0.00-10.00  sec  89.6 MBytes  75.1 Mbits/sec                  receiver

[  4]   0.00-10.01  sec  54.8 MBytes  45.9 Mbits/sec                  sender
[  4]   0.00-10.01  sec  54.7 MBytes  45.9 Mbits/sec                  receiver

[ 4] 0.00-10.01 sec 173 MBytes 145 Mbits/sec sender
[ 4] 0.00-10.01 sec 173 MBytes 145 Mbits/sec receiver

5GHZ in close proximity:
[  4]   0.00-10.01  sec   354 MBytes   297 Mbits/sec                  sender
[  4]   0.00-10.01  sec   354 MBytes   297 Mbits/sec                  receiver

[  4]   0.00-10.00  sec   113 MBytes  94.5 Mbits/sec                  sender
[  4]   0.00-10.00  sec   113 MBytes  94.4 Mbits/sec                  receiver

[  4]   0.00-10.00  sec   428 MBytes   359 Mbits/sec                  sender
[  4]   0.00-10.00  sec   428 MBytes   359 Mbits/sec                  receiver

Memory is a Samsung 16GB DDR4 3200MHz SODIMM PC4-25600 CL22, so Lenovo has been smart and left the second DIMM slot available for upgrade. For ~$100 on Amazon I can boost this up to 32G using the exact same DIMM which hopefully will allow interleaved memory.

This laptop came with Win 10 Pro, which I knew, on boot the image was pretty clean with very little installed. It wanted to then do an in place upgrade. Rather than do that I decided to do a completely clean Win11 install and use Lenovo Vantage to update all the drivers. On a clean boot from a clean install the memory use is pretty low for Win11.

SSD speed is again outstanding and on par with the Yoga, for reference the T450s racked up 192/291 MB/s so this is almost 5x write and 3x write. These new M2 drives are really amazing, and I can HIGHLY recommend them over even SSDs, not even to mention the snails pace rotating media yields.

Display wise this is supposed to be a 300 nit display 1920×1080, it’s bright enough, and crisp enough but at best an entry level screen. The touch screen is Ok, but no where near as responsive or nice feeling as the Yoga/Inspiron, honestly I don’t think it needs to be. The webcam includes a privacy screen, something that I believe all should.

Keyboard is ok, doesn’t seem to have a nice a feel as the Yoga, less travel, or so it seems, but all in all it’s fine. It’s backlit and give me the trackpoint as well as a VERY responsive glidepoint. Pretty much as expected.

The laptop came with a 65W USBC charger which again I applaud. If your buying a USBC hub, be sure and buy one that supports USBC-PD (power delivery) and then you have only one thing to unplug to be mobile.

Power management wise there is a setting in the BIOS that allows you to choose to between connected standby and the older style standby. Lenovo refers to it as Linux S3 or the default (which is connected standby) which is calls Windows/Linux:

With powerfg /a you can see the difference in standby modes supported:

Firmware set to S3 Linux only
C:\Users\mtn_b>powercfg -a
The following sleep states are available on this system:
    Standby (S3)
    Fast Startup

The following sleep states are not available on this system:
    Standby (S1) The system firmware does not support this standby state.

    Standby (S2) The system firmware does not support this standby state.

    Standby (S0 Low Power Idle) The system firmware does not support this standby state.

    Hybrid Sleep The hypervisor does not support this standby state.

Set to linux/windows:
C:\Users\mtn_b>powercfg -a
The following sleep states are available on this system:
    Standby (S0 Low Power Idle) Network Connected
    Fast Startup

The following sleep states are not available on this system:
    Standby (S1) 
        The system firmware does not support this standby state.
        This standby state is disabled when S0 low power idle is supported.

    Standby (S2)
        The system firmware does not support this standby state.
        This standby state is disabled when S0 low power idle is supported.

    Standby (S3)
        The system firmware does not support this standby state.
        This standby state is disabled when S0 low power idle is supported.

    Hybrid Sleep
        Standby (S3) is not available.
        The hypervisor does not support this standby state.

At one point somehow hibernate got enabled with some bizarre number of seconds and it was intermittently hibernating from sleep. Took me a bit to sort that out, and then I just disabled hibernate using powercfg. I found resume did NOT work well on Windows 11 from S3, the touch screen wouldn’t work.

Not specific to this model but here are some performance numbers for this class of device. First off using lightroom desktop you will see that the GPU does very little in spite of Adobe claiming they support it. In this case I am exporting and you can see it is 100% CPU bound, uses all cores, and you will also notice the extreme amount of memory used. Adobe recommend 8GB min and 16G recommended.

Here’s another performance graph showing that the GPU hit and maintained around ~70% while playing World of Tanks, and it’s all 3D, of course. Again you can see the amount of RAM used, it shocks me these days how much memory can be CHEWED up … Given a lot of laptops (not this one) are not upgradeable and often top out at 16Gb, it’s one of the reasons I chose this model.

A good sized battery is good, this one is 45,730 mWh according to the command: powercfg /batteryreport, but what is even more important is how long it takes to get back to some power you can use once dead. These new laptops have really come head and shoulders above the old ones in charge speed. This one with power off on the laptop off charged at 1.6%/min which dropped down to around 1.5%/min with the laptop in use. This means it went from 33% up to 70% in 25 minutes, impressive. The 65W power supply over achieved pumping out over 70W, with the charge current ramping down after the battery crosses ~80%.

So all in all … there was a lot to be intrigued about the 2in1s but in the end they were unable to deliver a reliable, well designed platform. This laptop is a solid corporate workhorse that just delivers, is massively expandable from a memory point of view. I think I FINALLY have found my new laptop.

March 10, 2022 Posted by | Uncategorized | Leave a comment

Lenovo Yoga 7i review

After the debacle that the Dell Inspiron 14 turned out to be, I decided to go ahead and buy my first choice, this laptop. Now there were two reasons I went with the Dell, one was price and two was availability. Well in between buying the Inspiron and becoming completely dissatisfied with it, which took about a week, Lenovo resolved availability issues that it previously said was going to be months. There’s an old saying the cost of poor quality lives on a LOT longer than the memory of the low price and when it came to the Dell that would quickly come true. I still liked the idea of a 2 in one, and so the Yoga 7i was the choice … I admit to having a bias towards Lenovo having worked for IBM for years and having their laptops personally for just about as many, in fact I am replacing a Lenovo t450s … The Lenovo was about 30% more expensive than the Inspiron so let’s get going and see if it’s worth it.

The specific model I’m reviewing is 82BH:

Straight out of the box the build quality and materials used by Lenovo are top notch. Lots of metal … And if anyone knows how to make a hinge for a 2 in 1 it’s Lenovo, and sure enough it is perfectly designed, no sharp edges on my lap, Dell should learn … do it right or don’t do it at all. The Lenovo came with an industry standard 65W USB charger which I applaud, thank you!

Size wise the Yoga 7i is

17.97-19.25mm x 356.46mm x 235.65mm weighing in at 3.1 lbs. Vs the Dell Inspiron:

17.95: 321.50 mm 3.31lbs, so comparable.

Jacks on the Lenovo are pretty limited so be aware, 1 USB-3, one USBC power delivery and one USB-C, that’s it (oh and a 3.5mmm audio). There’s no HDMI port, no ethernet, no uSD card nada, so you will be needing a USB dock for full desktop connectivity. This time around the Lenovo (unlike the Dell) was rock solid with my noname USB-C dock that included ethernet, HDMI, and USB3, but be careful a lot of these USB docks are not USB-C power delivery which means you still have to plug in the USB-C power adapter separately, look for USBC-PD …

It became very clear that Lenovo has been a lot more mindful of heat and power consumption, on AC the CPU were appropriately ramped down while on the Dell they were always full tilt resulting in the CPU fan rarely shutting up and the CPU temperature being always VERY HOT, now the user can tweak power settings but I expect the vendor to be more mindful and expect less of the end user. The Yoga has vents on the top so be careful which orientation you sit it on or there will heat issues. Even with full tilt charging temperatures of the laptop were very reasonable, not even hot to the touch. Under load the unit does get warm, especially in tablet/tent mode where airflow is more restricted. I didn’t notice heat as much on my lap …

The unit can be used as a tablet or in tent mode, but it’s not the nicest thing in the hands in this mode, and holding it with the keyboard folded backwards doesn’t exactly feel terrific as you are pushing keys. The sharp edges are uncomfortable at best.

I’ve been overwhelmed by the speed of the new M2 NVM drives but the one is this laptop takes the king of the hill as the FASTEST I’ve ever seen.

The camera does not support facial recognition but the finger print scanner worked VERY well being setup in short order with no issues, head and shoulders above Dell’s. The camera does include a physical privacy shield.

Using a USBC realtek adapter I managed good performance, which is on par with other systems and was stable (the same was not stable on the Dell) (using iPerf3).

On WIFI 2.5GHZ in close proximity to the access point I managed (which is more than double the t450s numbers):

[ 4] 0.00-10.01 sec 173 MBytes 145 Mbits/sec sender
[ 4] 0.00-10.01 sec 173 MBytes 145 Mbits/sec receiver

On 5GHZ in close proximity to the access point I managed :

[  4]   0.00-10.00  sec   428 MBytes   359 Mbits/sec                  sender

[  4]   0.00-10.00  sec   428 MBytes   359 Mbits/sec                  receiver

These numbers are typical. The laptop supports WIFI 6, but honestly I won’t be upgrading my access point anytime soon, my slow internet and wires in the right places means it’s not worth the investment … for now, until I get an itch, sadly knowing my personality of knowing something better is out there is enough to make me want it … 😉

The keyboard is backlit and has nice touch feel … the delete key is where it belongs as are most keys with only the up and down arrows being a little smaller than usual. Sadly, there is no touchpoint so your limited to a glidepoint 😦 I have no idea why Lenovo would eliminate their trademark innovation from this device, much to my shagrin. Overall feel is good, but of course can never compete with desktop keyboards. The touchpad is responsive and smooth.

The display is 14.0″ FHD (1920 x 1080) IPS, touchscreen, 300 nits which is at best average in comparison to the competition, nothing outstanding, but nothing horrible either. Touchscreen is also adequate but not outstanding.

Lenovo have always done an outstanding job of managing the MANY drivers needed to make laptops sing, with them moving to Lenovo Vantage to handle it all and keep everything up to date. And it works well as always.

Processor wise the Processor: 11th Generation Intel® Core™ i7-1165G7 Processor (2.80 GHz, up to 4.70 GHz with Turbo Boost, 4 Cores, 8 Threads, 12 MB Cache) is near the top of the heap with 4 core hyper threading, you have to go a lot higher in price, weight and bulk to go up to the next level 6 (for example i7-10750H ) or 8 core (for example  i7-11800H)… It’s zippy and responsive.

Memory on this model is soldered 16 GB DDR4 3200MHz with no expansion slots, so buy the max you will EVER need. And Windows 11 is memory hungry so be aware. With lots of trimming based on my article on taking back your new laptop a clean start is:

Lenovo include a 65W USB-C charger, from what I saw when the battery was fully charged the laptop was pulling around 30%, so if you need to charge your laptop quickly, the best way is to put it to sleep or powered off. I ran a little test with the battery run down and the laptop shutdown. First off this graph shows time Vs % and as you can see the charging ramps up VERY quickly (around 1.4% per min) right up to around 60% and then starts to slow, by the time it hit 90% it was down to half the rate or 0.7%/hr and then even slower. The same graph with the system on, under light use shows the dramatically slower, and the last curve shows the stock adapter plugged into a USB Green USB-C dock which says it supports OSB-C PD. As you can see in the chart the charging speed is even slower, watching current the PD on this USB-C dock NEVER max’d out the power from the charger, drawing only 50 of 65W. To go from 33% to 90 took 46 mins with the laptop powered off, 70min or 1.5x as long with the laptop in use, and 100mins or 2.1x as long through the USB-C dock’s PD.

What I found even more interesting was to watch the current drawn from the 65W adapter, starting originally at 72.2W or full power output and then dropping off around 80%. I have no idea if it supports it but Lenovo offers a 95W adapter (PN GX20Z46240, or 4X20V24690) for around $80 that could potentially help with quick charging, especially with the power on, although, through the USB-C dock it would make NO DIFFERENCE because the charger never max’d out. This chart shows time Vs % (with the laptop powered off, plugged directly in) as well as time Vs watts drawn from the power adapter.

These new processors are back to supporting connected standby which in the past was a debacle. I shuttered when I saw it in the battery info. I was relieved to see they seem to have worked out the bugs or so I first though. In ~10 hours the battery dropped 5%, or 0.5%/hr, which from a full charge would translate into about 50 hours on connected standby. The second day I watched it I was not as impressed, in 8.1 hrs it dropped 14% or a whopping 1.7%/hr, no clue why it was so different the second time. What connected standby does, is it attempts to simulate how a phone works keeping things like email etc up to date. Older processors went into a VERY low power standby mode that would last for days even weeks. If I had a choice I always prefer the older standby method. But one does not always have choices, or Microsoft takes them away from us … moving on …

And now we find ourselves at the most unpleasant part of this experience the pen … brace yourself, it’s BAD, I mean the worse pen experience I’ve EVER had. I bought the Lenovo digital pen, PN GX80U45010, as recommended by their presales dept. So let’s start out with the documentation which is TERRIBLE. Here’s the doc, it’s minimal at best. Luckily I had the Lenovo Pen app installed or I would be even more stumped, not that it helped much either. There’s a link to bluetooth settings, great so let’s pair the pen … over an hour later I call Lenovo … nope this pen isn’t bluetooth enabled, well thanks for that wild goose chase.

Next up are the buttons … so I’m pushing them and pushing them nothing is happening. Well, as it turns out the ONLY way the buttons work is to push the button and then move the pen close to the screen? WTF? So I ask are there any other pens I can use with this laptop … NOPE, that’s it. And the physicals are abysmal, the buttons are placed such that writing or drawing without pushing them is challenging. Oh and no erase functionality on the top either. The feel in the hand is really not great, nor is the writing on the screen. And no pen clip to hold it in your pocket … do we really have to reinvent the physicals of a pen? Oh and no spare tips either … And now we get to functionality, well … the lag between the pen and the screen draw is worse than my REALLY old Asus Vivonote 8 … Palm rejection is awful and even trying to write with this pen is an act of EXTREME patience. Now I could suspect that there’s something wrong with either the laptop, the pen or both from the youtube videos I’ve seen but man this is definitely a HUGE disappointment. FYI I tried the pen with Microsoft OneNote, Microsoft Whiteboard and Lenovo Smartnote, bad in them all … This is one area the Dell was WAY better.

Ok, so here we find ourselves in a bad situation … I bought the Lenovo Yoga 2 in 1 to use it as a 2 in 1, ie in tablet mode and with the pen, now both aren’t happening, so why would I accept the limitations of the 2 in 1, ie soldered memory? Well … you guessed it, it’s return time AGAIN. Stay tuned for the next installment.

March 3, 2022 Posted by | Uncategorized | Leave a comment

Winter birding in Algonquin Park

I had seen a number of people talking about heading up to Algonquin park to see some unique birds, ones we don’t often, or even don’t, see here in the GTA. A number of people gave us suggestions for where to go and so we decided to venture up. At 3.5 hours each way without running into delays we decided to do an overnight and ended up staying at the Huntsville Inn which is about an hour outside of the park, but was the best we could do, we booked it at the last minute. We were thrilled to find anything, this was over the family day holiday weekend. And so off we went.

Believe it or not, one of the most productive places to go to see birds is the Algonquin visitor center where they have a feeder setup, you can even watch it live on Youtube. There’s a viewing platform that you can look down on the birds, or if you are feeling adventurous there’s an unmaintained path down below the viewing platform that gets you at eye level. But be forewarned, especially in the winter this path can be challenging.

Path down to the feeder warning unmaintained

Inside the Visitor center you will also find a message board for when and when people sighted things.

Message board in the visitors center

Crossbills (which we didn’t see) are apparently common high up in the top of trees as you approach the visitors center. Here’s some pics taken at the feeder at the visitors center.

Driving along HWY 60 which runs straight through the park is a great place to find moose, although we didn’t. If you see cars stopped, that may be why … check it out. We were also told to checkout Openago Lake Rd for moose, but also had no luck. We did however find Canada Jays on Openago Lake Rd! Some have also seen grouse’s along the roadside. Sadly this was the only moose we saw 😉

Another unique inhabitant to the park are pine martens which we were told like to hang out near the garbage cans at Mew Lake campground which is open throughout the winter, as well again at Openago Lake campgrounds. We didn’t find one, but here is an image I shamelessly pilfered from a google search.

Pine Marten

Here’s what the garbage cans look like your looking for, look in the trees near the cans.

Garbage cans at the campgrounds

We did a lovely 1.5km walk along Spruce bog which is home to beaver, spruce grouse, and pine martens (none of which we saw), but we did find a beautiful Canada Jay as well as a pair of Common ravens, which by the way are quite common in Algonquin unlike in the GTA.

Common raven

The Algonquin logging museum trail was suggested, but we didn’t get to it, a black headed woodpecker had been seen there, which would have been a real treat but no joy …

Ragged falls is less than 1KM in from the parking lots and is quite pretty in the winter, but aware this is a one way trail and can be quite slippery in the winter so take care.

Ragged falls map

If your looking for even more to do in the area there’s a unique ice skating trail at Arrowhead provincial park. In the end we loved are time and counted 4 new to us birds, it was well worth the trip! Thanks to Cerian Ware Segal, Dale Darcy, and Elizabeth Szekeres who’s posts from Halton/Peel birds on Facebook guided us on the trip! Here’s a link to gallery of everything we saw on day one, day two to be added when they are done, sorry 😦 Day 1 we took 886 photos, almost 24G worth … 

February 24, 2022 Posted by | Birding, Uncategorized | Leave a comment

Dell Inspiron 14 2 in one laptop review

I’ve been living on a Lenovo T450s for over 2 years now as my main machine. I’ve done a few upgrades, more RAM and more SSD but it was getting a little long in the tooth, I probably ought to have thought of a new machine sooner rather than upgrade this old one … I bought it on refurb, so it wasn’t new even when I got it … I’m a Lenovo fan I’m one of those few people that actually like and use the trackpoint … But what really put the nail in the coffin of the t450 was the announcement that Windows 11 would not be supporting the processor in it. At some point a company has to decide how long to support older hardware especially when trying to move the platform forward for things like security, I’m not thrilled, but I get it. At 8G of ram (after upgrades) my T450 was doing ok, well until I discovered that Adobe actually recommend 12G or more for Lightroom, which has become my photo editor of choice. The processor in the t450s is a Core i5 5300 dual core hyper threaded. All of this was to frame where I am coming from …

So I looked at a number of options including buying another refurb. The Lenovo Thinkpad X1 Yoga 3rd trickled to the top for coming with a Core i7, includes an active pen and comes with a 2560×1440 display with reasonable brightness at 500 nits compared to most around 300. The large bezel, outdated processor and outdated network card left me wanting more from this upgrade.

On the top of my list was a Lenovo Yoga 7i, unfortunately availability was terrible at MONTHs to deliver, is not memory upgradeable, does not come with a pen, display is 1920×1080, so sadly this dropped off the list.

And so, without further ado, we move onto a review of the Dell Inspiron 14 2 in 1 5410. Acers ended off the list for the way they mess around with keyboard placement, that drives me nutz btw. Ok let’s start out with specs …

First off, when you get a new laptop the first thing I HIGHLY recommend you do, is DECLUTTER the absolute travesty of crap Microsoft, and your hardware vendor preload for ALL THE WRONG reasons.

I chose the 14″ rather than the 15 simply because I use it in laptop mode infrequently and when I do I don’t want the weight so this comes in at

Dimensions & WeightFront Height: 16.32 mm (0.64″)
Rear Height: 17.95 mm (0.71″)
Width: 321.50 mm (12.66″)
Depth: 211.35 mm (8.32″)
Starting Weight: 1.5 kg (3.31 lb)*

The unit is made up of a combination of plastic and aluminum. The overall feel of the unit is good and feels well made.

This can come with, and I chose the Processor 11th Gen Intel(R) Core(TM) i7-1195G7 @ 2.90GHz, 1805 Mhz, 4 Core(s), 8 Logical Processor(s) which for now is pretty near the top of the heap in the mobile field. Adobe Lightroom really pushes the processor during edits, import and exports. From the t450s this is a HUGE improvement going from Dual core to quad core not to mention other improvements in the process. So all in all this ought to be a big welcome boost. This processor gets hot, so when pushed and when on AC the fan comes on and is noticeable. I run mine in tablet mode on my desk with a keyboard and so the airflow is restricted and the fan is on more than I would dream of.

Display wise this is only 1920×1280 and not the brightest or crispest displays on the market at only 300 nits brightness. Compared to the T450s it is brighter and crisper, but not by a lot. The touchscreen is way more responsive, and the Dell supports a pen, more on that in a bit. The display has a nice small modern bezel.

The Dell supports twisting the display around to use it like a tablet, thus the 2 in 1 … similar to the yoga. I wasn’t sure how much I will use this feature so we will see, but it’s intriguing. Not long after receiving the laptop I discovered a HUGE design flaw, airflow to the processor comes from the bottom grill of notebook and flows out the top of the vents close to the screen. If you use this device in tablet mode with these vents facing down the laptop quickly over heats. I went to bed one night with the laptop in this orientation and the unit over heated, shutdown, started up and ran diags and then left itself on, overheating until I found it the next morning. Absolutely idiotic and something that Dell should have anticipated. Now that I know, and once you know it’s obvious, I will attempt to avoid this orientation, but this is definitely a WTF were they thinking moment. I thought I had this figured out, and then I continued using it in tablet mode on my desk and two more times the unit overheated and shut down, it would appear that tablet mode is not meant for constant use, and since this is the way I intended to use it, I will be returning it. Fortunately speaking with Dell I should be getting a full refund, a shame really …

The hinges on these types of 2 in 1s are always challenging from a design point of view, and in this case Dell missed the mark, when on your lap, not being used in laptop mode the display protrudes down below the bottom of the tablet leaving a sharp uncomfortable edge, it would seem they don’t expect this to be used on a lap? WTF … again.

Unlike some Dells this unit does NOT have a trackpoint and uses a glide point, something I reluctantly tolerate. The keyboard has better than average feel and key placement is for the most part the usual with only a few exceptions. The fingerprint scanner is part of the power button (really WTF … again) and those are where the delete ought to be. The Up and down arrows aren’t where they ought to be or the normal size. The keyboard is backlit. The finger print scanner is supported by Windows hello, so can be used to logon, the camera is NOT supported by Windows hello so cannot be used to logon. All in all the keyboard is what I would call better than expected.

Memory wise this unit has two SODIM slots and can be expanded to 16GB according to Dell, and 32G according to Kingston, while the processor datasheet says up to 64g. It’s DDR4 3200 and from comes from Dell ONLY in 8G (which is populated stupidly IMHO as two 4G, yup another WTF … to do any upgrade you have to throw out one of the 4Gs) and 12G (8GB/4G). There is no 16GB option or I would have chosen it. DDR4 BTW clocks in at 37.5gb/s compare with DD3 in the T450s 25.5GB/s or a performance boost when outside of processor cache, which is uncommon, of 46%. I do like how upgradeable and flexible this unit is, compared with a LOT of the competition with soldered on RAM.

One of the things I wanted in my new laptop was a M2 PCI SSD, I’ve seen the performance benefits and there are so many times the hard drive is THE gate. This one comes with a Model NVMe KBG40ZNS512G NVMe KIOXIA 512GB which is blazingly fast. This is compared to 269/299 MB/s on the SATA SSD of the T450 so more than double the speed. This is a HUGE improvement, and of course random access of an SSD is head and shoulders above a rotating disk, I could NEVER go back to rotating media.

SSD drive speed

Sound quality is terrible, way worse than my T450s which, honestly I thought wasn’t possible. Gun sounds and explosions in World of tanks sound way less satisfying. I’m shocked how bad this was …

I bought the Dell pen they recommend for this unit a PN350M, and it was ok but wasn’t the best feel in the hand. Now, I was shocked to find out the Microsoft surface pen worked perfectly including bluetooth support, battery level, button configuration, eraser functionality everything (within OneNote). Sweet! Writing on the screen feels good, just the right amount of resistance, no delay at all. The pen did NOT come with the required AAAA battery …

Dell Active Pen – PN350M

Port wise this comes with 1 USBC (includes power delivery), one USB3, a full size HDMI, a barrel plug for charging with the proprietary adapter, and a microSD slot. I hate that they didn’t use USBC for the charging, but oddly you can use a USBC charger if you have one and charge speeds are similar. This is by no means a quicky charge, but not slow only, at 65W you only have so much juice for charging the system takes up about half or more of the 65W anyway … The USBC port was unstable, I could not get any of the three USBC-docks I have to work, all of which worked on other machines, now it’s possible the overheating was the culprit, I don’t know, but this definitely contributed to the feel of an unstable machine.

One of the most unique things I was shocked about is Dell have included something called Dell Mobile connect, powered by Screenovate, this gives you complete access to the iPhone, contacts, messages, pictures, and remove desktop support of the phone. I had no clue this was even possible on an iPhone!

Overall there was a LOT to like about this machine, sadly, even more to hate with so many WTF moments. So back it went to Dell …

February 15, 2022 Posted by | Uncategorized | Leave a comment

Taking back control of your new Windows PC

Ok, you got yourself a brand spanking new PC, and out of the box your excited … awesome, that new PC smell? 😉 Ok it’s important to get started controlling what is running and taking memory, processor and disk from you without your knowledge (or consent). Pretty much every tech company thinks they know better than you what you need, I mean I know they are trying to do their rendition of “let’s make the PC great again” but heh. You decide, some of this may seem out there … and maybe not. If you want to see the fruits of your efforts you can note the amount of memory used from a clean start from task manager, right click the start menu, select task manager. You can also take note of the amount of drive space as well.

The first place I recommend you start to to UNINSTALL anything you don’t think you need. Now a word of caution, some preloaded stuff once removed can not be added back, so tread somewhat carefully. Start out by going to Control Panel Uninstall a program.

Here you will find a list of STUFF a combination of Microsoft and your vendor loaded “for you”. Some of this stuff the vendor PAID to have installed, you know a “Free Trial” or a “One year subscription”. Ya gone. There’s a tendency to think heh I might try that one day … a year later you still haven’t … Moving on

Next up go to the start menu and click all apps and this will show even more stuff that Microsoft, and or your vendor “Loaded for you”.

You can right click on top of most of it and say uninstall. Some are a little more challenging, especially Microsoft bundled stuff.

Next up head on over to the Settings (right click the start menu and select settings) and click Apps, Startup and you will see the list of CRAP that vendors think you need started for you … Prune this by disabling as much as possible.

Some apps are cagey and don’t put themselves here so they end up putting themselves in as a service. You can deal with these by right clicking on the start menu and select run, and then type services.msc. Here you will see the autostart setting for services, if you have an app that just keeps auto starting have a look in here. You can change it to either auto start or disabled if you don’t EVER want it.

Getting there, now let’s deal with some Microsoft bloatware that I know I don’t want and even if I do I can reinstall it … First off Microsoft teams, you can find it in the start menu, right click uninstall … bye bye. Next up head over to your taskbar and right click, taskbar settings and clean stuff up, get rid of chat, if you don’t have a touch screen get rid of the virtual keyboard and if you don’t have a pen get rid of pen and ink, and if you want your start menu back where it belongs click taskbar behaviors and put it back on the left instead of the default center, ya this ISN’T a Mac and putting the start menu in the center won’t make it so.

If you’re like me you don’t want the extra clicks on the right click menu for context, another of the Microsoft knows better suite.

So to get rid of it from an administrators command prompt, and reboot

reg add HKCU\Software\Classes\CLSID\{86ca1aa0-34aa-4e8b-a509-50c905bae2a2}\InprocServer32 /ve /d "" /f

Another way to find stuff you don’t need is head back to settings (right click start menu select settings) then apps, apps and features, here too you can do uninstalls.

From settings you can also see windows installed stuff from settings, apps optional features. Here you find more USELESS stuff.

Next up, Microsoft just love XBOX, so much so totally removing EVERYTHING to do with it is next to impossible. I started playing a game and all these distractions came up, click this to record, click that to do that, oops I’m now dead because I was distracted by all this silliness … Head back to settings, gaming XBOX Game bar and turn it off

One of the Microsoft apps called Your phone only works with Android, so I don’t want it, however Microsoft decided to NOT allow you to remove it … Start an elevated command prompt (click the start menu, type cmd, then right click and select run as admin)

And then type powershell … then run this command

Get-AppxPackage *Microsoft.YourPhone* -AllUsers | Remove-AppxPackage

Ok, it’s time to see how we did and if we missed anything. Let start task manager (right click start menu and click task manager) and then start resource manager:

from here you can see what is taking up memory and how much each is taking.

If you want to see how much your time has saved you do a clean reboot and then re note the amount of free memory and disk space you now have. For me I saved well over two gig of ram and took back SOME control over my new PC. Thanks Microsoft, Adobe and in this case Dell. Time well wasted. Grrrrr

February 12, 2022 Posted by | Uncategorized | Leave a comment

Zabbix monitoring

I started this little rabbit hole as a curiosity of what open source monitoring could do for me, rather than coming at it from addressing something I was looking for. The more I got into it, the more I discovered how much rich information I could get about my environment. Googling Open source monitoring you find a lot of choices. Which will fit your needs really depends on what your environment looks like and what you want out of the monitoring. You quickly narrow down to three types of monitors. As an amusing anecdote, when I mentioned I was looking into monitoring tools, my GF looked at me with a concerned look and asked … what are you monitoring 🙂 Well you know, your only paranoid if it turned out you weren’t being watched, otherwise your wise 🙂 I digress … back to this post and types of monitors …

Basic up down functionality. For this, internally, I have been using Monitorr which is free. What I like about this app is it provides a one stop easy way to see what’s up, and you can provide a link to go directly to that service, super handy. What I don’t like is it’s basic ya the port is listening, no notifications and if a port is down it takes an INORDINATE amount of time to show the page. Honestly for the convenience and prettiness of having all my internal services listed on one page I will keep using it despite it’s limitations. It is running in a container so costs next to nothing in resources. I’m sure there are other ones like this but I haven’t found them.

For external monitoring I use the free Uptime robot which has a notifications, a web dashboard, as well as an iPhone app. It does what I need …

The next type of monitors are based on SNMP or Simple Network Management Protocol. SNMP has been around for a very long time now, and for the large part is considered, old, archaic, insecure and basic in the data it provides. SNMP is disabled on almost everything so getting this working takes a bit of time on each and everything you want to monitor. It starts with installing SNMP, making sure the firewall is open to allow SNMP in, and then configuring SNMP for things like community settings. In linux there is a fantastic tool called snmpwalker that can help you debug if SNMP is setup and working or not.

snmpwalk -v 2c -c public

I played a bit with Cacti which is one of the easiest ones to get running under Ubuntu, it’s a single install that does EVERYTHING for you. There are no agents for Cacti, so SNMP was what it used, Cacti was fine but I could already see wanting more info that it offered. My time with Cacti was limited, so this is by no means in depth, but it does work and is quick and relatively easy to setup.

The next kind of monitoring is based on an agent, a piece of code you manually install on each and everything you want to monitor. Agent based monitoring will ALWAYS get you the most complete amount of information possible. Agents will only be available for certain types of devices so you will need to check first.

I first landed on Checkmk and the amount of data provided out of the Windows agent was extensive and I was impressed. It integrates well with VMware ESX by setting up an account that it logs on and gets loads of info from. It plays OK with Docker, but the way it’s done is problematic, it relies on the docker container ID (which is a 12 digit code) and while you can add an alias the container ID is shown in more places than the alias and your left trying to figure out which container ID is which container. For something you want to glance at and glean from this is BAD. By the way, the easiest way of finding the container ID that you need to add it to Checkmk is to use docker ps command, something they didn’t mention in their documentation. I also found the Windows agent to be inconsistent in my environment. It loaded well into one machine and refused to work on two others. I got to the point where I just plain gave up. There is no FreeBSD agent so monitoring on Pfsense is limited to SNMP. Checkmk if it works in your environment and the agents work is a great choice, well other than the Container ID silliness … Installing Checkmk is no small feat, on docker it’s a number of containers you bolt together. Their so called appliance doesn’t include checkmk? WTF. Installing it on ubuntu went ok, but involved. Leave lots of time and be prepared for lots of googling.

Update: I also looked into PRTG … PRTG is one of the easiest to get working, it runs on Windows and you simply run the installer and let it do it’s magic. I first installed it in a VM on ESX and it was not stable at all in spite of giving it more than ample resources. I moved it to an old laptop, believe it or not a Pentium N3450 with 4G of memory and it ran like a charm on Windows 10. The code then does a scan of your network, which I found this less helpful and provided a LOT of clutter to clean up such as items on your network there is little to no point monitoring. And every point you monitor has a number of what they call sensors, maybe memory/cpu, drives etc. Here’s the kicker … the free version of PRTG allows only 100 sensors … sounds like a lot? Well … my Windows file server had over 100 sensors on it’s own? So I could monitor 1 machine? Now I could work at pairing down to the bare necessities, but what’s the point. My VMware ESX server, which BTW was trivial to setup, ended up with 21 sensors, including one sensor for each VM. So as you can see you will reach the free limit in a New York minute. And the cost to buy it is $1799 (US I assume) for 500 sensors. I ended up at 188 after some pruning. I could probably get it down to 100 by monitoring NOTHING but my two servers. There’s a iPhone/Android app that is excellent. On Windows hard drive SMART parameters were supported out of the box, but no reading of the temperature on the drives, or the temperature of the system. On ESX SMART was NOT supported. So there’s lots to like about PRTG, well except their sensor limit and price. NOT HAPPENING …

Next up I moved over to Zabbix to give it a look, and spoiler alert, since this article is about Zabbix it’s the one I like the MOST, hands down. First up, let me save you some time, there are two versions of the agent, zabbix_agent and zabbix_agent2. There are places the first gen work well, and others ONLY the 2nd gen of the agent work. I’ll show that in a bit …

Update: 2/18/2021 Well, after struggling to get this all setup perfectly I discovered a number of errors I made. I’m placing this in the middle of the post to help you avoid my errors. Zabbix has auto discover agents that prepopulate if you have things setup right. If you have manually created hosts, and their name is the same as the real name, then this will get in the way of the auto population, host name and visible name MUST be unique. So, before you spent a lot of time on things, and get in the way of auto populate. And if you’ve created dashboard on an item you ultimately will need to rename to get out of the way of auto population, and you want auto population for the best results, you want to get this right or your redoing your dashboards. Be sure and change your internal network to the right seubnest in configure, discovery. Ok back to the post.

Zabbix includes an appliance that is completely setup ready to play, based on CentOS, and MySQL, it takes only 10G of drive space and 4G of RAM. It’s a good thing the appliance exists because the documentation for how to install 6.0 and 5.4 on Ubuntu immediately runs into a roadblock that the Zabbix repository doesn’t have a release file and won’t install. Overall documentation of Zabbix is good, but not amazing. I found myself using other peoples documents to supplement what was lacking from and just plain wrong on Zabbix, which, fortunately, there’s lots of. This guide for example tells you how to install agents and was helpful, albeit stale. Zabbix has a free and opensource version so perfect for in home use like mine. The appliance uses DHCP for easy setup but I converted mine over to a static IP, agents ONLY talk to the IP of the Zabbix server so having it on DHCP wouldn’t be my choice. Zabbix can also do SNMP based monitoring for anywhere the agent isn’t available. I played with the appliance for a while before I discovered for whatever reason I couldn’t backup the appliance up using VEEAM, I think the import of the appliance went badly and I didn’t know it, so I installed it manually. Because Ubuntu wasn’t happening I chose CentOS 8 stream version. The guide to install Zabbix left out critical steps, installing MySQL and installing/configuring Apache and then lastly configuring the Zabbix front end. They just kinda assume you can figure that out on your own. Googling reveals how to do both. So what can Zabbix do? A LOT … let’s talk about each device being monitored …

I am moving the VMware discussion up in the post, because getting the VMware host right, will get the VMs right and save you a LOT of time. So there are four VMware templates that have auto discovery that are the key to getting this working, they work with both ESX stand alone as well as Virtual center. Zabbix logs into VMware, grabs data through an API and populates data for you. To do this first of all create a userid on VMware, but it only needs read access.

In my case I have only one ESX host in the environment so I am going to directly configure the templates. The alternative is to add the manually add the host, add the hypervisor template and put the variables Zabbix needs to log on to ESX with. Go into each of the 4 templates and define:

I was stumped for quite a while because I used a browser and checked out https://ip/sdk and it didn’t work so I thought it was wrong, it isn’t. Once the templates are set, do all 4 be patient and wait. And remember to NOT have any manually entered hosts with hostname/visible name that will block auto population. Now wait an INORDINATE amount of time … Zabbix discoveries run once and hour, and once discovered populating all the data for a host takes a while. All in this process can take hours. So set it up, and LEAVE IT ALONE. You can work on other things within Zabbix as long as you don’t get in the way. Once the ESX host is auto discovered here is what it precreates:

The alternative way to make this work if you have more than one ESX host is you can manually create the ESX host, but I would use a name that won’t conflict with auto discovery. The host will look like this:

Now click macros and your going to define the same variables as above my manually adding them, and they are:

{$VMWARE.URL} (remember this is https://ip/sdk

If you have manually added the host you can check on status in the latest data for that host, eureka, you can see Zabbix logging into the esx host!

Ultimately the way this ends up working, if auto population works, and if the moon aligns with venus, and the Zabbix gods are feeling nice to you, is a host will get created that has the name set to the system UUID of the vmware machine, something that is challenging to get, and the visible name set to the actual hostname within vmware. And then once the host is working, then the guests will start showing up, and they are created with the hostname as the VM UUID and the visible name set to the VMs name. Here’s an example:

Now you will notice this is NOT modifiable, so if you have additional templates you want to add to that VM I would suggest create a new host that is yours to add as you want.

You can also manually create a host and enable SNMP on ESX to get more info … but only 13 of the 37 datapoints actually work this way, so not a lot of extra info this way.

This takes a LOT of time for the magic to actually happen, but with it working you have have access to a lot of VMware guest related info, you will still get even more info by installing an agent on the VMware guest, which is next up …

We install the Zabbix 2 agent on my docker Redhat VM container host so that Zabbix can properly monitor all of the stats from memory, disk use, network use, cpu etc at the container level. To install the agent add the Zabbix official repository and then install the agent 2 using yum.

rpm -Uvh
dnf clean all
yum install upgrade zabbix-agent2

The ONLY miss I have is that if a container shuts down cleanly there is no notification. If it exits with an error code you get informed. If it were to flat out crash, well all bets are off. So from this point of view, this isn’t perfect … but good enough, a theme that you will hear more often then you ought to … The Zabbix 2 agent on Linux is COPIOUS in the amount of data you can glean, quite impressive. You MUST use the Zabbix 2 agent if you want docker information. Here’s what the configuration looks like (I’m on Zabbix 6 release candidate):

Redhat with docker setup on Zabbix agent 2

From a Windows point of view my file server is running Server 2019 and I had issues getting the Zabbix agent 2 to do ANYTHING. So I took a step back to version one of the agent and got scads of data. Download and install the agent from Zabbix’s web site. Doing some reading it appears the windows agent does NOT collect system, or CPU data and does NOT support drive SMART? Wow that’s a bit of a miss 😦

Windows host setup with Zabbix agent

Pfsense is well supported, you simply install the Zabbix package on your Pfsense firewall and then setup your host on Zabbix. Pfsense in case you are unaware uses FreeBSD. I found a couple Pfsense unique templates but the ones I tried didn’t work, so just using default FreeBSD which provides most OS level info, but good enough …

There are MANY templates available to help monitor a LOT of stuff. If you can download an XML file of the template it can easily be imported and then added to a host. The amount of stuff out there is mind boggling. For example I found a template for Nextcloud, Plex, PiHole etc … A lot of these other templates are very stale, some of which work, some not … It’s a mixed bag as one should expect. Most of my stuff is running in containers, so the best way to configure this is to add the templates to the manually added host for the container host, which for me is a VM, this avoids getting multiple monitor notifications when something is wrong at the container host level.

What is less impressive is how difficult it is to setup a simple ping test, and a simple is your web server up? Both of these were quite challenging to get working, I’m just relieved after DAYS of pulling my hair out that it is working. Here’s a sample ping test I’m using to monitor a site to site VPN connection and alert if it goes down.

Simple ping test setup

Once a host is setup, you can see if data is being received from that host and what it looks like by clicking on the host on the monitoring screen and selecting latest data, here’s an example of what you will see, since there are values, it’s communicating. It can take a few minutes for data to flow be patient:

Latest data coming back from a host

With anything you setup there are triggers, sometimes these triggers are something you don’t care about, or your getting too many false positives. From the configuration screen for hosts, you can see the “items” that are being collected and the triggers that are being actioned upon:

Items and triggers

Clicking on triggers will allow you to see what the settings are and you can disable things you don’t care about. In my case, I disabled slow responses over the site to site, and only worried about whether it’s up or down.

Alerting is the crown jewel, if something goes down, or is in peril, ya I kinda wanna know. Getting this going was a LOT more challenging than I expected. You first go to your media types and configure them, for email, you will need an SMTP server (in the Administration media types). Next up you have to enable for each userid which media types are selected (in Administration users). Last but not least you have to enable trigger actions (in Configuration, actions trigger actions), phew finally. Sure there isn’t something more I need to do? You know like a yes, no yes? Moving on … Once configured if you trigger a failure you can see the email was sent on the problems page. Once the issue is resolved you will be sent a message as well. This really is the reason to monitor, and it works VERY well …

With monitoring in place the next step is dashboards and in this the tool is amazing. You can totally flexibly setup a dashboard with any of the detailed data coming from any of the hosts giving you a clear picture of what’s going on in your home devices, such as all CPUs in the home, or all network being consumed etc. Really quite flexible and easy to setup.

And with that, this two week long rabbit hole is done for now. I’m sure I will continue to tinker with stuff and see what else I can add but it’s in place, monitoring, logging and notifying! I actually am IMPRESSED with Zabbix!

Update: So with a LOT of tuning I have Zabbix working correctly. Now one of the things it has done is given me longer term performance metrics from everything running in the home. Here’s the revelations that came out of that data:

  1. In spite of running 6 VMs and 13 containers on VMware ESX 6.7 the number one limiter is memory once an SSD relieved the rotating media issue, CPU is almost NEVER the limiter, so spending money upgrading the processor is frivolous, although, extra cores makes multi tasking through so many applications more seamless.
  2. The number one thing that pushes my drives to the limit is running disk based backups.

Userdata customizations

Update: SMART/Temperatures

So I decided it was important enough for me to monitor temperatures and SMART not covered off by Zabbix, so I wrote my own scripts. Zabbix is VERY easily extensible to add whatever you want. I wrote a script based on this command to check the status of the SMART of ALL drives in the system. I did the same for ESX and fed it back to Zabbix through the Windows file server: Here’s the pertinent code

For Windows to check the SMART status of the drive:
wmic diskdrive get status
- this gives a pass fail with I parsed and passed back number 0 or the number of drives failing

For ESX I used plink to pass a command from Windows to ESX and check SMART (and then add the name of all of the hard drives you want to monitor after the -d)
esxcli storage core device smart get -d
- I parsed it to get the drive status and temperature both of which were passed back to Zabbix

Now getting CPU/System/Drive temps was a LOT more complicated. I found a program OpenMonitorReport that worked on my system and gave me access to everything I need, CPU temp, System temp, drive temperatures etc. This code sadly takes to long and always timed out to Zabbix, and I had issues writing to temp files for parsing, so I wrote a script that simply loops and grabs the data, and then parses out the necessary data. This is then written to txt files that the Zabbix agent grabs and reports back.

Update: I decided to relook at the way of getting CPU/system temperature from a Windows system and found a better way. I looked at the web interface for Open Hardware monitor and while this would work it would be a lot of code to parse it. So instead I found Openhardware monitor creates entries into WMI that I can query from powershell. Now getting the syntax right was BRUTAL, took HOURS. However, this now doesn’t rely on a looping batch file, which is running anyway for the next topic ESX CPU/system temps. So I added Openhardware monitor to the startup of the Windows system. The I run the following powershell commands and feed the results back to zabbix. Now it’s worth noting, that Open hardware monitor does not work on all systems, and if it is closed or dies then so does this method.

To get the CPU temperature:
get-wmiobject -namespace "root/OpenHardwareMonitor" -query "select * from sensor where name='CPU'"|select-objec
t "value" |  select -last 1| Select-Object -ExpandProperty Value

To get the system temperature:
get-wmiobject -namespace "root/OpenHardwareMonitor" -query "select * from sensor where name='System'"|select-ob
ject "value" | select -last 1| Select-Object -ExpandProperty Value

to get the drive temperatures
get-wmiobject -namespace "root/OpenHardwareMonitor" -query "select * from sensor where name='Temperature' AND P
arent='/hdd/$drive'"|select-object "value" | select -last 1  | Select-Object -ExpandProperty Value

Both my ESX server and file server have a UPS, so I wanted Zabbix to monitor it, this would provide a way of knowing if the power in the house went out. The hubs, and modem are on the UPS as well so there is a chance to notify. This turned out to be quite easy using powershell. For ESX I have a VM running that monitors the UPS anyway (and shuts down the ESX host if the power is out long enough), so it was easy to add this to that VM.

$batterystatus=(Get-WmiObject Win32_Battery).BatteryStatus[$which_batt]
$batt=(Get-WmiObject Win32_Battery).EstimatedChargeRemaining[$which_batt]
Write-Output "$batt"

Monitoring the system and CPU temperature of ESX was a LOT more challenging. ESX does not support the sensors on the host I have, and the sensors are not exposed to any VMs so I was left with a challenge. So I found a USB temperature sensor that had two sensors in it on Amazon. One sensor is in the dongle and one is a remote sensor. So what I did was use a USB extension cable to put the dongle inside the ESX host, and then took the remote sensor and McGiver’d it in against the CPU frame to give an approximate CPU temperature.

The device comes with Windows code that constantly logs the two temperatures into a CSV at an interval you can set. From here it was a simple matter to parse out the results and feed them back to Zabbix. To insure the file doesn’t grow indefinitely or roll over, from time to time I clean out the CSV programmatically leaving only the last entry. To do this I wrote a powershell script to mimic the unix tail command.

powershell /command "get-content %1 |select-object -last 1"

code to parse the CSV
if exist D:\pcsensor\TEMPerX\1.csv (
    rem file exists
) else (
 echo 0
for /f "tokens=*" %%i in ('powershell /command "get-content D:\pcsensor\TEMPerX\1.csv |select-object -last 1"')
 do set return=%%i
for /f "tokens=1,2,3,4,5 delims=," %%a in ("%return%") do set date=%%a&set internaltemp=%%b&set cputemp=%%d&set
set cputemp=%cputemp:~0,-1%
echo %cputemp%

Pfsense also allows you to easily grab the CPU temperature for passing back to Zabbix. I created a shell script on Pfsense that has the following code:

sysctl -a | grep "dev.cpu.0.temperature"|sed "s/dev.cpu.0.temperature//"|sed "s/C//"|cut -c 3-

I added items and triggers within Zabbix for all this added data.

February 11, 2022 Posted by | Uncategorized | Leave a comment