DIY Eclipse Jump Starter Kit - Pure and Undiluted
Not too many words, but a lot of useful stuff.
Intended for the Hive API node operators, witnesses, and developers.
Except for the video, the video is for everyone:
https://www.youtube.com/watch?v=rOLhoc6vuXc Cool, huh?
# Services
## API node
https://api.openhive.network
It currently runs `hived` at `v1.24.6` and `hivemind` at the latest `develop`.
A lot of fixes and optimizations have been recently introduced, so I try to keep up to date.
During the maintenance mode, it will fall back to https://api.hive.blog
## Seed node
`hived` `v1.24.6` listens on `gtg.openhive.network:2001`
# Stuff for download
TL;DR https://gtg.openhive.network/get
## Binaries
`./get/bin`contains `hived` and `cli_wallet` binaries built on `Ubuntu 18.04 LTS`
Right now it’s: `v1.24.6`
## Blocks
`./get/blockchain`
As usual, the `block_log` file, roughly 300GB and counting.
## Snapshots
`./get/snapshot/api/` contains a relatively recent snapshot of the API node with all the fancy plugins.
There’s also the `example-api-config.ini` file out there.
The snapshot is compressed with `lbzip2`: 280GB
> lbzip2 is a free, multi-threaded compression utility with support for bzip2 compressed file format.
To decompress, you can use simply run it through something like: `lbzip2 -dc | tar xv`
The uncompressed snapshot: 413GB
To use snapshots:
- Get `block_log`, it can be bigger but not smaller than the one used when the snapshot was made.
- Get a reference config file, adjust it to your needs, make sure you don’t affect it in a way that changes the state.
- Get a binary compatible with the snapshot (a newer one is OK only when no replay is required).
- Run `hived` with `--load-snapshot name`, assuming the snapshot is stored in `snapshot/name`
`hived` API node runtime: 722GB (incl. shm 18GB, excl. snapshot)
## Hivemind database dump
`./get/hivemind/` contains a relatively recent dump of the Hivemind database: 50GB
I use self-describing file names such as:
`hivemind-20201105-294e2411-Fc.dump`
Take your best guess.
The custom format is compressed by default and you can use `pg_restore` with `-j 6` (recommended) to restore the database in parallel.
After restoring, make sure you run the `db_upgrade` script.
When restored, database can take 600-700GB
All resources are offered AS IS. Obviously.
# From zero to hero
By using the above resources you can get your fully featured API node running in one working day.
# More to come
Stay tuned.
See: DIY Eclipse Jump Starter Kit - Pure and Undiluted by @gtg