(Did you know that already ran a witness node 😱4 years ago? I was trying to learn as much as I could about this life... was so epic back then 🤦‍♂️. Then I got persuaded to go to EOS as a block producer, which didn't go anywhere after 6 months. Still, I learned a lot and I don't regret it because it was a super worth experience!)

Back with a HIVE witness node

This time operating via @atexoras.witness 😎

And what an amazing way to start 2022, right?... Still, lots to go, since it's not meant (for now) to be a public node for RPCs or as a seed node, but it does make you feel kind of free and empowered on being part of this community.

I HAVE turned ON the following plugins...

~$ cat localhive/datadir/config.ini |grep plugin |grep -v '#'
plugin = webserver p2p json_rpc
plugin = account_by_key_api condenser_api
plugin = witness
plugin = rc
plugin = block_api network_broadcast_api rc_api
plugin = state_snapshot

I will be playing with a couple of these in the near future and make adjustments as I go. Lot's to learn as there are new things I didn't follow the development.

You can vote for it HERE if you wish. The link uses a great tool developed by @cadawag that makes it super simple to vote for witnesses, no matter where you are. Remember to support its witness too if you like what he has done.

No blocks signed so, that's why there is no version, but I am running 1.25.0

~$ curl -s --data '{"jsonrpc":"2.0", "method":"condenser_api.get_hardfork_version", "params":[], "id":1}' localhost:8090

{"jsonrpc":"2.0","result":"1.25.0","id":1}

And enough RC mana, although it's something I will start monitoring in case it falls below 50%...

~$ curl -s --data '{"jsonrpc":"2.0", "method":"rc_api.find_rc_accounts", "params":{"accounts":["atexoras.witness"]}, "id":1}' localhost:8090

{"jsonrpc":"2.0","result":{"rc_accounts":[{"account":"atexoras.witness","rc_manabar":{"current_mana":"932981261795","last_update_time":1640917200},"max_rc_creation_adjustment":{"amount":"5709043939","precision":6,"nai":"@@000000037"},"max_rc":"933121590307"}]},"id":1}

What's this helping the community with?

For now, let us say that we have another copy of the HIVE DB in New Zealand, in sync, further promoting decentralization and the security of this blockchain. The main objective, for the time being, is to help me (no seed/RPC available yet, sorry 😭) with my other witnesses' initiatives, on Hive-Engine and DLUX.

Still, I need to go through a couple more "works" into it but saying that I managed to get this node running in 4 days, including downloading (with axel, tks @apshamilton) the entire chain from @gtg (link here), prepare a VM, replay the entire thing, and make some local tests... I would say, it felt pretty damn great. Still not perfect but I guess it's what I could do with what I have.

I intend to help others that need to go through the process, so, if are attempting to build a local node like me (for now), reach me out anywhere. Will help with what I can/know or direct you to others that know more than me.

Next year I need to buy a new CPU (because now things are getting exciting). All is going to be supported by my own fricking limited resources! I am here because I like it, want to contribute to and be a part of it... Although it costs money to run these things, be rest assured.

Witness status

The code is way more efficient today than 4 years ago 😎... the community team contributing to further developing this really did a Mega Hyper job! Well done everyone.

8GB of RAM (that could even go to 4 or 2GB) - for witnesses, RPCs, and non-history nodes it's enough (wait for it) if you use the shared memory file in NVMe like me. Just to give you a hint, active memory is less than 1GB... so most of it is for caching situations.
6 vCPUs (based on Zen2 AMD cores) - overkill... 4 is more than enough! - I might reduce...
shared memory file... (currently 20GB) on an NVMe (max 50-60k IOs observed during replay), could go more... but needs more frequency because the replay process is single-threaded. The common alternative to this is to create a ramfs and host the shared memory file there... memory is more efficient on IOops than NVMe, hence faster!
block_log in local ordinary rotary disks using LVM with a 3-stripe array (it needs some IO at the beginning of the index building, but after it, its requirements are as low as any disk can handle, as far as I could see).

Experience...

The replay took almost 3 days... but I moved the block_log into and NVMe for the index rebuild, then back to the rotary disk. That single-threaded(?) process that does most of the replay takes most of the time.

Overall, the experience of re-learning things was not difficult. I had already some experience, so for others that don't it might be harder, but if you have some help is not a big deal. This time of the year is also not the best time to build things... LOL, I can't blame anyone. But doco could improve, there are lots of old stuff and copy-paste jeopardy that for the new user will be dust in the eyes. Maybe this is something I can help with (will be evaluating that), but otherwise, let me know.

Documentation and Resources used

During the setup, I have used mostly ecency searches, some witnesses guides, and HIVE documentation and source code. Here are some of the useful links I consulted with:

Possible TodoS!

For only 4 days of work, I am pretty happy. But there is some stuff I need to complete still...

Add Jussi reverse-proxy to the node.
Add failtoban and test some profiles (not sure if there are some already there)
Allow specific witnesses nodes "to talk" seed/RPC to my node (via IP whitelist) - if you wish to be whitelisted, let me know where you are communicating from.
Explore certificate-based options for the above situation...
Others I might be missing out on...

Upgrades on the horizon... more like wishes!

CPU... (500-2000 NZD) I NEED LOTS MORE! I wish I had only one of my work threadrippers... 😐
Grow to a 128GB RAM (1200-1400 NZD) hypervisor and move the shared memory file in ramfs
New firewall/router (1000-3000 NZD)
Upgrade ISP connection (moderate expensive but might become viable with other projects)

2023 Scope "dream" wishes

If this can even be a possible scenario, this would be the objective! This is not yet a plan, just a dream wish... let's see what will change when we get there!

Hardware:

Dual 10Gbps Firewalls (with HA)
2x 6-10Gbps ISP links
4x Hypervisors (32 core, 64-threads each and 256-512GB RAM)
2x GPUs/accelerators
10Gbps core switch(es) (1 or 2)
6-12Gbps SAS FC Storage (shared across the 4 hypervisors)
8Gbps or more FC HBA's
x2-x4 NVMes, some SSDs, and some HDDs...
Some bigger UPSs for minimal support
Find a place at home to put all this without getting divorced 🤣

Brainstorming services:

2x HIVE witness/RPC nodes with HA availability
1x HIVE seed node
2x HIVE history nodes
2x Hive-Engine witness nodes
1x Hive-Engine snapshot node
2x DELUX nodes
1x 3Speak node (encoding + IPFS)
1x Monitoring page for the services state (public)

I don't even know how much this would cost right now... but the point is to say that if in possession of this hardware, I could make it happen, just by myself. Would just need a bit of help on the last item (webpages are not my style) 😅

Wish you the best for 2022!

Stay safe, if possible near your loved ones, and enjoy the last day of the year the best you can. I am going in first then most of you (in a few hours)... so, see you on the other side. HIVE IN!

(source)

🍀 For a great 💪 cause check this quick post about #tourettes 🤯 call for ⚡ action!

It will not take more than a few seconds of your time and potentially make many lives happier. Thank you for even considering reading it.

🖐️HIVE5's!

Have fun, good luck on all the gaming, and lots of teaching "how to play" to new players.

🤝 Twitter - @forkyishere 😈 (@forykw dark side)

For #news about the #HIVE blockchain, and other stuff. I am following all users!

In addition, if you are looking for a nice place to either reach out, share or just have a great time, come along to @atexoras.pub gatherings. We welcome everyone on the HIVE blockchain.