The last one and a half weeks were very exciting for me. After all the things I experienced during HiveFest 2025 in Kuala Lumpur, I decided to stay a bit longer in Asia, especially in Thailand. It was a great mix of adventure, nature, and community time. But of course, at some point, every trip must come to an end.
In just a few days, I traveled from the small Thai island of Koh Lipe, then over to Langkawi, and finally back to Kuala Lumpur by plane. I had a short break there to relax before flying home to Europe. But even while enjoying my time in Asia, I had one big thing on my mind: the upcoming Hive Hardfork, version 1.28.3.
This important software upgrade for the Hive blockchain stayed in the back of my head during the whole trip. And when I finally got home, I did not wait long. I jumped right back into action.
First, I rented a small cloud server for around 8,99€ per month to quickly set up a backup witness node. I wanted to make sure everything would run smoothly before touching my main big bare metal server. So I started with the small server, which was running Ubuntu 22.04. I installed the dockerized version of Hive by 
The blockchain replay took around 30 hours on this cloud machine. Yes, I know – that is a long time. But the CPU on the cloud server is a shared CPU, so it is not the fastest. Still, it worked fine, and once the replay was done, I temporarily moved my witness over to the cloud server so I could start working on the big one.
On my main server, there was some cleanup to do. I had not done a deep system refresh in a while, so I decided to stop and remove all running Docker containers and other software that was no longer needed. After that, I upgraded the system from Ubuntu 20.04 to 22.04, and to my surprise, the upgrade went very smoothly. It was all done in around 30 minutes, and everything worked right away.
Then I installed the latest version of Mahdiyari’s docker setup again. This time, I added again the very useful setting:
block-log-split = 1
In the screenshot below, you can see the memory usage before and after the replay with this new setting enabled. The difference is huge – and it shows how even small changes in configuration can make a big impact thanks to the grandios work of the blocktrades team and all other Core-Developers!
Before Upgrading to 1.28.3
Total Available Memory
19,59 GiB out of 62,76 GiB RAM available in Total.
After Upgrading to 1.28.3 - with block-log-split = 1
Total Available Memory
55,54 GiB out of 62,76 GiB RAM available in Total.
Right now, I have my main server up and running again. I also still have the smaller cloud backup server online as a second option, just in case. In the next few days, I am also planning to set up another backup witness – this time on my Raspberry Pi. That one will be more of an experiment. To be honest, I am still missing some basic architecture knowledge on a Raspberry Pi, especially when it comes to compiling and getting the right Blockchain Software Version. But I am motivated to learn more, and I will definitely take a closer look at the guides shared by the BeeDevs here on Hive. They have already posted a good tutorial, and I plan to follow their steps carefully once again.
Now that all the software updates, backups, and replays are completed and my witness is running smoothly, I can finally breathe a little and focus on other things for a while. The system feels stable, and it is always a nice feeling when everything works the way it should.
Fingers crossed for all of us that everything goes well on November 19, when the Hive Hardfork is planned. It is a big step forward for the chain, and I hope everything will go smoothly without any unexpected issues.
Until then, a big thank you for all your support – and let us keep building together.
Support my work by voting for me as a witness!
Hive Witness:
👉 Support using Hive Keychain
Hive-Engine Witness:
👉 Support using Primersion Tool – search for louis.witness
SPS Validator:
👉 Support for louis.sps using The SPS DAO Validator Page
Magi Witness:
👉 Vote for louis.vsc on Magi Explorer
