Custom High Fiber & Protein Breakfast “Bars”

This is my take on https://www.allrecipes.com/recipe/231925/high-fiber-high-protein-breakfast-bars/ .

For me I’m doing some work around my blood sugar and have made a custom fiber blend I decided on that I was using in yogurt for a while. Now, I had good results but that is by far NOT the only thing so this does not constitute medical advice.

However, recently I’ve been slacking a bit on my breakfast with this fiber in it so this in theory will allow me to get that custom fiber back into my diet in an additional ‘easier’ option. And by that I just mean I don’t have to mix it in the morning as it involves making ahead of time.

Dry Ingredients

• 60 grams – Instant Oats
• 33 grams – High Protein Rolled Oats
• 36 grams – Epic Protein Powder – Sprouted Living
• 50 grams – Custom Fiber Mix (See Below)
• 1.5 grams – Table Salt
• ~ 1.65 grams – Maple Sugar

Wet Ingredients

• 136 grams – Smashed Banana
• 99 grams / 1 Packet – 100% Mango Baby Food Puree (https://plumorganics.com/products/mangos-baby-food/)
• 41 grams – Chunky No Sugar Peanut Butter

Directions

1. Mix dry ingredients in a bowl
2. Mix wet ingredients in a larger mixing bowl
3. Add dry ingredients to wet ingredient bowl
4. Bake in oil sprayed muffin tins for ~20 minutes @ 375 F

Custom Fiber Mix – Equal Parts:

• Solaray Oat Fiber
• NOW Foods – Acacia Fiber
• Doctor’s Best Prebiotic Powder Sunfiber
• Jonnys Good Nature Organic Ultra High Resistant Starch

Nutrition Information Will Change Depending on Your Exact Ingredients

Lime Cilantro Chicken Taco Filling

This is a very simple ‘taco’ filling that is just different enough from my usual to warrant writing it down.

Ingredients:

• 1 lb bulk chicken breakfast sausage
• ~2 tsp of taco seasoning
• ~ 1/2 to 3/4 cup lime juice
• ~ 1 cup cilantro

Directions:

• Cook the chicken sausage in a frying pan
• Once cooked through sprinkle on the taco seasoning, stir through, and then add the lime juice
  • This is less taco seasoning than normal for a lb of meat and we essentially replace the water with the lime juice
• Once mostly reduced and coating the chicken throw in the cilantro and mix in
• Take off heat and enjoy

Skyline Skillet Lasagna

This is a derivative of my no-bake noodle skyline recipe. Easier, faster, and tasty.

1. In a wide oven safe frying pan bring to boil 1 – 14 oz can of Skyline + 1.25 c ups of water
2. Add 1/2 pound Radiatore Pasta (Others should work you may have to adjust time or amount of water)
3. Boil for covered for 12 minutes stirring ever few minutes
4. Mix 4 oz cream cheese with 4 oz ricotta
5. Drop the cheese mixture into ‘dumplings on the top’
6. Sprinkle a good amount of cheddar cheese on top
7. Put in a 350 degree oven for 6 minutes or until the cheeses are melted
8. Add oyster crackers and enjoy!

Instant Pot Lower FODMAPs Chili

This recipe uses lower fodmap Adzuki dried beans to make it simple.

Ingredients:

• 1.25 cup dried Adzuki beans
• 1 lb ground beef – I used pre-made meatballs because it is what I had at the moment
• 1 cup diced onions
• 1 cup diced green peppers
• 1 tsp garlic powder
• 1 – 28 oz can of crushed tomatoes
• 1.5 Tbsp Chili powder (or to taste)
• 1 Tbsp Worsheshire sauce
• 1 tsp salt
• 1/2 tsp pepper
• 2 cup water

Directions:

1. Sauté onion, green peppers, and garlic powder till translucent
2. Brown the ground beef
3. Add the rest of the ingredients
4. Cook on high for 25 minutes and quick release

Review:

I have never cooked with Adzuki beans before especially not from dry state. I originally cooked on high for 18 minutes and the beans were still crunchy, so I added 6 more minutes on high pressure, I thought still crunchy, so another 10 minutes. Some were still crunchy. I stopped at this point. The bottom had burned a bit when I got to it and I thought some beans were maybe still undercooked. However, after serving out 3/4 of it to freeze the bowl I ate was very good and the beans were fine. So all in all I’m not sure how to cook it for. I set the recipe to 25 minutes because I’m guessing if you keep going and not release the steam that should be good. But pay attention to the beans or maybe try them before hand to see what properly cooked ones taste like. Ha! The directions I had for from dry Adzuki beans in the PC said 15-20 minutes.

The flavor and texture were great. It wasn’t an award winning chili but it was an EASY chili. I ate it with some Cotija I had around and shredded cheese. And as I always like it with crackers. You can add more water at the end if it is to thick for your taste as I think the cooking beans thicken the chili.

I will definitely do it again.

Easy Instant Pot Red Lentil Taco Filling

Ingredients

• 1 cup red lentils
• 1 cup water + 1 cup water
• 1 packet taco seasoning

Directions

1. Add 1 cup water and lentils to the instant pot
2. Cook for 3 minutes on high pressure
3. Mix in the taco seasoning and 1 cup water
4. Simmer until preferred thickness

Review

It ended up thicker than standard taco meat filling and the color is different due to the lentils but since the spice mix is the same the flavor is very very close. I mean you can tell it isn’t meat if that is your first choice but it was still very tasty and I’ll definitely do more of it. While cold it sort of feels more like refried beans but I ate it with just cheese and a sauce I like and enjoyed it. At about $3/lb for red lentils and each lb making about 2-3 batches of taco filling. It is a very inexpensive, filling center of a meal.

Instant Pot Mango Pineapple Pork

Ingredients:

• 1 cup diced onion
• 1 can pineapple chunks with juice
• 1 can mango juice
• 1 tsp garlic powder
• 1/4 tsp seasoned salt
• 1-1.5 lb pork stew meat or cubed other pork
• 1 Sweet Potato – Peeled sliced into 1.5 inch slices
• 1 – 1.5 cup carrot chips
• 1.5 cups snow peas
• 2-3 cups broccoli

Directions:

1. Saute diced onions, garlic powder, and seasoned salt until translucent-ish
2. Pour juice of pineapple can into instant pot set pinneapple cubes aside
3. Pour mango juice into instant pot
4. Add cubed pork
5. Cook for 10 minutes on high pressure
6. Fast release pressure and add the sweet potatoes and brocolli
7. Cook for 2 more minutes on high
8. Fast release pressure
9. Cut sweet potatoes into bite size chunks
10. Add snow peas, carrots, and pineapple chunks
11. Let sit for a few minutes to heat the above
12. Serve over rice and if you want to get fancy you can cook the rice in some pineapple and mango juice

Review

When I cooked it for 8 minutes then 4 the sweet potatoes and broccoli are pretty broken down. So I’d recommend going 10 and 2.

Overall it is very tasty and semi-surprisingly flavorful. Make sure to pour some of the liquid over the rice to get all the flavor out. It isn’t a flavor bomb or over the top amazing but it gives the hint of pork chops with apple sauce through the fruit juice in the dish. I used a wild rice blend to give it more body too.

If you wanted you could 1/2 the liquid, thicken it, or add more spices.

Day 2 Review

Somewhat surprising I really like it on day two. I ate it without the rice and no extra juice added. It is very flavorful and the mushy sweet potatoes and broccoli almost make a mash that goes well with the crunchy carrots and snow peas and the pork. I also added some pre-cooked bacon bits in but that wasn’t a big part of the flavor.

Odroid N2+ Final Build Steps

I’m ready to do the final build.

• Step 1 – Flash eMMC and boot
• Step 2 – Reset all DietPi passwords
• Step 3 – Do a clean install
• Step 4 – Enable WOL
  • nano /boot/boot.ini
  • setenv enable_wol=”1″ (it is 0 by default)
• Step 5 – dietpi-config
  • Language & Region Settings
    • keyboard can only change w/a real keyboard plugged in. ie even on SSH I needed a wireless keyboard plugged in. But also ssh gave me the right charatcter # vs currency pound sign so I didn’t care
  • Security Options:
    • Change Hostname
• Step 6 – Using dietpi-launcher for the next steps
• Step 7 – DietPi-Cron
  • Enable Minutely w/15 minutes (for the auto shutdown script)
• Step 8 – DietPi-software
  • MinIO
  • Mosquitto – MQTT
• Step 9 – Configure MinIO
  • http://IP:9000
  • Create a bucket
  • Create a user w/readwrite permissions
• Confiigure custom shutdown script
  • apt-get install mosquitto-clients
  • Connect to MinIO server and tx file
  • ln -s /mnt/dietpi-userdata/minio-data/[bucketname]/script.sh /etc/cron.minuetly/scriptname
    • no .sh suffix
  • chmod +x script in cront.minutely
  • test with the command
    • run-parts –test /etc/cron.minutely
• Pause For Day 1 until the HDD is hooked up

Power Consumption Update:
—-This only includes the board and eMMC no keyboard or external storage

• 2.5 W while powered on and idle
  • higher then 2.2 reported but I’m measuring at the wall so this include PSU draw
• 0.6-0.7 W on suspend
• < 0.5 W while poweroff — the current meter I have plugged on only reports <0.5 W so I can’ get more accurate reading with this particular meter

Odroid N2+ Hardware Check

I got my board a few days ago. Snapped on the eMMC module preloaded with Ubuntu MATE 18.04 LTS and booted up. Once I figured out the network cable was bad it worked straight out of the box with SSH support.

Since then I’ve walked through the hardware tests I wanted to check with an official OS before I try my 3rd party OS of choice: dietpi

1. IR receiver works – Checked b/c it was bent upon arrival
2. RTC works
3. Suspend Works
4. WOL works — This is AMAZING to me
5. USB 3.0 works — doesn’t shutdown and reboot (original test was with USB 2.0 which threw me for a loop for a bit)
6. Ethernet gets 900 Mbps
7. MinIO / FTP speeds aren’t 100% of dd disk speed but way better than R Pi. On a flash drive that depending on the write test got 25 – 35 MB/s I was getting 22 MB/s through MinIO and FTP.

I also measure power consumtion with a usb stick in the device at about 3.2 W idle though I think it is on performance govenor at the moment. I tried turning off cores to save power but didn’t see any result immediately. Also, that number uninclues the PSU as well.

At this point I’ve done all the proof of concept tests. Next up once I get the right device I’ll flash the eMMC with dietpi or if I can find an extra microsd card I may try that first and see how dietpi hangs out.

I’ve also done alot of searching of ‘other’ things to do with this powerful little board — short version it is way way way overkill for what I want to do. ie lots of extra resources around. But the reality is:

a) I already have main server doing the NAS, media streming (web player), backups, tools, etc so most of my needs are covered
b) I don’t want to host outside services on my backup server for security reasons
c) I live in a small apt so I’m pretty strigent on projects to the need only category
d) My ASUS router has enough features for what I like to do there
e) I don’t game, tor, torrent those sort of class of options
f) I’m one person so I’m not needing solutions that cover more than me — ie keepass is fine I don’t need to self host a web version for mutiple people to get to. KISS.
g) I want to to turn it off so Pi-Hole isn’t great to run on it!

The only 2 things I’ve come up with possibilities are Pi-Hole and Kodi. If Kodi did the big streaming services I’d use it but it doesn’t and the other server does media serving so it might as well do playback at the same time too and not add another device.

It was designed to do one thing and do it well and that is the most likely scenario even if the hardware has so much more potential.

If I think of something I really want to do with it I could likely swap it out for a Odroid C4 model for ~$70 preshipped. Considering the N2+ cost about $100 preshipped that $30 doesn’t really feel like it makes a difference.

If anyone has some good suggestions for the N2+ let me know. Mostly looking for software based options but I’m a simple person when it comes to my home life. Small, compact, less items, less power rule the decision making — and to boot I know I don’t use shiny things so unless I have a big reason it will just languish and isn’t worth doing. Not on hardware anyway – VMs are easy to clean up.

MinIO Final Hardware Choice – Odroid N2+ 4GB

Why the hardware change? Twos phrases: Wake On Lan & True Sleep

The best power saving mode is the one where computers only run when needed and are in standby all the other times. I figured out “finally” the Odroids (C4, HC4, and N2+ at least) support WOL. Which when combined with my MQTT shutdown program I wrote will allow my main server to run scripts before and after the backups to:

1. Wake up the back up server
2. Write status (Started/Stopped) of back up jobs

This will allow the N2+ to go into truse system sleep after the backups and be woke up before each backup. AWESOME.

To boot the idle power consumption if left on is LOWER than a Raspberry Pi 4 supposedly but a good 1-2 Ws. Around 2.2 W for the board itself.

That is the main factor that made me go YES this is the right solution for me. Here are the other supporting factors in no particular order:

1. BARREL JACK FOR POWER – omg so much better than USC C in my opinion
2. Full sized HDMI port – no new cable to buy “Just” for setup
3. Cheap case available – $7 at expensive american reseller
4. Small “Enough” – Bigger than a Raspeberry Pi form factor but only wider by about 1.2″ wider. Not “stupid small” like the Rock Pi 4+ Solution but also still small enough and usable form factor
5. ARM Cores – powerful AND energy efficient: 4x A73 + 2x A53. This is same number of ‘big cores’ as Raspberry Pi 4 but with more energy efficent or powerful cores PLUS 2 little cores that are used on the Raspberry Pi 3 (RP3 has 4 cores so only 1/2 the power). All those cores for LESS power the Raspberry Pi 4 at idle and at load. While I don’t NEED the procesing power it is nice to have to grow into if I want
6. Cryto Extensions! While after doing lots of reading I don’t think this matters as much BUT the N2+ has the arm crypto extensions so that is another leg up over Raspberry Pi 4
7. Form factor — all plugs are on the back of the board I need to use. So much nicer than side plugs like other Raspberry Pi style boards
8. Supports eMMC modules up to 128 GB so fast ‘local’ / OS storage
9. Cost – All in the new setup cost me about $150 w/16 GB eMMC

The only downside to the N2+ is I’ll still be using USB 3.0 drives but here is the reality of that:

1. I already have a right sized one with the backups on it so it should be easy to just move it
2. The max speed of USB 3.0 spinning drive is still WAY faster than the Raspberry Pi B+ model I’m using. I get about 2.5 MB/s currently and in theory I should be able to get around 80 MB/s with USB 3.0 and Gigabit Ethernet (That should in theory move 1100 GB in 22 minuts vs 13 hrs. esp since realistically it took over 24 hours b/c the pi or HDD throttled)
3. Even 50 MB/s would make the backup take 38 minuntes vs 13 hours
4. Complexity is actually less – NVME and arm is costly and expensive for speed gains I can’t use (just size gains and SATA boards are farther between
5. USB 3.0 is PLENTY fast for a backup computer data drive

There is a small possibilty the USB 3.0 ports will not work on the N2+ based on some forum reading but….

1. I think it is has been resolved
2. WAsn’t seen with a single data drive
3. It was most often with mutiple drives at full bore speeds — Gigabit speeds of 100 MB/s is lower than that maximum as it stands

Why not the Rock Pi 4 Plus any more?

Main reasons:

1. Availablity — I’m still waiting for it to become available and now that I’ve ordered something I can stop thinking :)
2. NVME Cost
3. No WOL – RTC Wakeup is less dynamic
4. I liked eMMC onboard for the same cost but it also means I had to decide on eMMC AND RAM size when buying – super minor but aware

Total cost of buying everything but NVME was about the same as the N2+ so that is a wash

Out of these NVME was the one that lead me to look around again and no WOL is the one that sealed the deal. NVME compatibility with the Rock Pi 4 Plus was hard to figure out. I nailed down the drive I wanted to use due to cost and performance: SK Hynix P31 1 TB but I wasn’t sure it would work and no real way to find out but order it and true. All the ones that seemed supported were MUC more power instensive and made it not feel worth it. Yes, NVME storage would have made the server package clean and TINY (~3.5″ x 2.5″ x 2″) if it worked it would have cost me ~$130 more for a 1TB model or $200+ for a same sized 2 TB model. The cost addition and uncertainity of it working or not would bring the project up to about $300 and for a backup server I’m running on a RP1 B+ right now felt a bit high but…

….Wake on Lan. When I realized the Odroids all had TRUE sleep and TRUE WOL everthing changed. The RTC wakeup was a good option that in theory should work but still involved scheduling the server to wake up and putting other logic in places that would have added complexity and uncertaining into the process. WOL has always been ideal option. And in the end that is what made me full force go back to Odroid models. And to boot the N2+ was available today so I could commit.

Why not the Odroid C4? Or HC4?

Frankly, either would work fine.

Let’s start with the HC4. It is a C4 with built in SATA ports and the really is the ideal board I was looking for but there weren’t many good cases for the setup I wanted. The offical toaster form factor I don’t like and I really only wanted a single 2.5″ drive. I found some options but they had the heat sink sticking out a hole in the side, required 3d Printing, and maybe editing of the 3d prints which I would be doing for the first time. Or fabricating my own enclosure. A minor detail is the RTC clock batter is on the OLED display board which I wanted the RTC but not the OLED display — I possibly could have wired a battery into the header but it would take some figuring.

C4:

This is a more minor set of reasons.

1. Availbiltiy for one. It is in production and dont’ know when it will get shipped.
2. Slightly less powerful with 4x A55 cores so just less growth but also realistcally enough power.
3. Power Consumption is better than the N2+ at Idle by 0.4W
4. Sometimes needs powered cooling unlike the N2+ with it’s giant passive cooler
5. Cost is better than the N2+ for the same 4GB Ram
6. Supports WOL too
7. Form factor has side plugs

In the end I think it was close between the C4 and N2+ as the draw back for both is pretty simliar b/c they both use Amlogic chips. If I am honest in the end it was probably the idea of having a powerful board for $30 usb more that I could get now with big passive cooling with all the plugs on the back side and onboard RTC and battery that swayed me. All these small things for $30 extra…. yeah I think it is worth it.

N2+ Powersavings

Now with the N2+ I think out of the box it shouldn’t be that much more power consumption at idle than my RP1 B+ (2.2W vs 1.2) but I would like to drop it further if I can but frankly if I can do sleep + WOL my yearly power usage will be well below my current setup b/c I’ll be spinning down the extternal HDD as well. I’ll measure it but I expect my current setup to pull about 4W on idle right now.

I’ll see if turning off the cores I don’t need on the N2+ also saves power. For MinIO only usage which is mostly what I’ll be doing I expect a single A73 and single A53 core will be plenty – in fact a single A53 core may be plenty. So if turning off cores saves power I’ll do that too and if it doens’t I will not.

TL;DR

N2+ has all the features I want out of a board minus on board/in case large storage for a price I’m willing to pay over an Raspberry Pi 4 to get Sleep and Wake on Lan options.

MinIO + Rock Pi 4 Plus — Power Saving

If you read my other posts you now I reall am into making things not use power when they don’t need to. It is where this home server / NAS project started 14 years ago. It isn’t about money and it is about conservation – not that my powering down my Raspberry Pi will save the Earth single handed-ly but it also seems like the most ethical way to use computers if we are going to use them.

My main server I accomplish this with Lights Out software. My Raspberry Pi backup server was low enough I didn’t worry about it but this time around I’m thining about it more. I’m adding cores, adding power, and frankly new options so many years later. So while I don’t have a Rock Pi 4 Plus in my hand yet here is my initial ‘good enough’ idea plan:

1. Mosquitto on DietPi
2. Find CLI tool to write actions from the main server via Duplicati
3. Write “Start” and “Finish” for each back up job in a script set in Duplicati at the beg/end of each backup
4. On Rock Pi 4 Plus — application that reads all the backup start stops. WHEN it sees a stop and nothing started for 15 minutes — shut down.
5. RTC wake up at 8 am each day — or whenever the server backups should start.
6. Set the server to wake each day at the same 8 am time too in bios?

In theory this means both machines will come on together daily. Backup daily. And turn off about the same time together.

Traditionally I haven’t turned the server on if I am not using it so I may have to do some math to see if the main server being on for those minutes every day eats up the backup server power usage or not.

It would be best if I could use wake on lan to wake the Rock Pi 4 Plus but I’m pretty sure that doesn’t exist / work. The above is a good idea for shutdown logic I think and start up logic the RTC at least gives a option but I’ll still look a bit more for a more elegant option.

Update: I played a bit with MQTT and this will definitely work pretty easily.

On the main server before/after each job run this command:

mosquitto_pub.exe -h [insert server IP/hostname here] -m “Started” -t backups/[backup name here] -d -[insert username here] -P [insert PW here] -r

change Started to Stopped after the job is done

then on the Rock Pi 4 Plus on a 15 minute cron job run:

mosquitto_sub.exe -h [IP] -t backups/# -u [username] -P [pw] [-C 5]

the -C says close after 5 responses. So it should match the number of backup sets I have.

Explaination:

I’m using MQTT with retain — so the status of each ‘backup’ is kept on the server. Eack backup is in it’s own ‘folder’ so Backup/1, backup/2, etc. So each backup job has an active status of Started or Stopped at all times.

The Rock Pi 4 Plus will check those statuses every 15 minutes. If ANY say Started it will just exist. If they ALL say Stopped it will wait 15 minutes and check again. If they are still ALL stopped it will put the device into sleep mode.

That is my working theory right now. Coupled with an RTC wake.

I’m considering how/when to clear the ratined messages so I could say — turn on and WAIT until at least one backup has started. Then start checking to shutdown. The idea being that the main server isn’t always on and this would make sure to catch the next backup if the main server was off. Not as perfect as 100% synced backups but off a lot of the time AND make sure the backups happen.