Updates once a week? I don’t think it’s possible, but I’ll keep trying.
Same with “Hive Power Up Month”. So far so good.
(it's like recent "gtg's abbrs." but with more meaningful name, so I don't have to explain the title)
Hivemind
I’m not a 2nd layer wizard, but I’m doing my best to help with Hivemind’s QA. Maybe you can't see that on the outside, but Hivemind is improving its underlying tech a lot.
OpenHive-Network Github
Improved CodeQL workflow. It's used for code scanning, which then lets us improve code quality and security.
Performance
A lot is going on this battlefield. While reviewing the latest changes I’ve found some issues that were preventing hive sync
to catch up with the head block in reasonable time (#162
)
Testing
- Full hive sync tests. I’m hoping for the next stable release to be ready soon.
- Testing upgrade procedures, between
47a41c96
and currentdevelop
(12efa972
) - Testing SQL Serializer and HAF. Still a lot work to do but both projects behind those buzz-words are very promising.
Updated snapshots and dumps
I’ve just made available for download a recent hivemind database dump and a hived API snapshot.
https://gtg.openhive.network/get
If you wonder about sizes:
- Blocks - 385G
- Hived API Snapshot - 352G (uncompressed 518G)
- Hivemind's database dump - 54G
Optional hived flavor
Please note that hived-v1.25.0
that is available for download is a general-purpose binary that is good for all use cases, i.e. for seed-node and fully featured API node.
However, on some of my nodes I’m running a different, witness-specific flavor of hived
build, that was created with -DSTORE_ACCOUNT_METADATA=OFF
.
It makes it incompatible with API nodes, as it has all account’s metadata records empty, but it still has everything that’s needed for seed nodes and witness nodes, and results in slightly smaller shared_memory.bin
file.
Miscellaneous
- Small fix for docker building instructions
- Revived
cli_wallet
meta-issue that I’ve mentioned in the last update, which includes recently added issues: - Identified issues with building latest develop in certain environments
Hive public seed nodes
One of the hard-coded public seed nodes was recently decommissioned so I’ve updated the list and while I’m at it I’d like to ask around for new/updated entries.
Here’s a current list:
api.hive.blog:2001 # blocktrades
seed.openhive.network:2001 # gtg
rpc.ausbit.dev:2001 # ausbitbank
hive-seed.roelandp.nl:2001 # roelandp
hive-seed.arcange.eu:2001 # arcange
anyx.io:2001 # anyx
hived.splinterlands.com:2001 # aggroed
node.mahdiyari.info:2001 # mahdiyari
hive-seed.lukestokes.info:2001 # lukestokes.mhth
seed.liondani.com:2016 # liondani
hiveseed-se.privex.io:2001 # privex
This list of nodes is available in doc/seednodes.txt
(for scripting) and hardcoded into hived
binary.
Such a list doesn’t have to be long, few entries are enough for a node to quickly discover and connect to dozens of others, but it’s essential to have a reliable set of publicly available ones.
You don’t have to be a witness to run such a node, however, because witnesses’ job is to ensure security and reliability of the Hive network (and they are compensated for it) it’s expected that they run them.
If you run such node and your node is missing on that list please let me know in comments or on OpenHive.Chat - I’ll either add it to the list or add to the waiting queue (to ensure its reliability before adding to repository)
Best practices for public seed nodes:
- It has to be owned and operated by a well known Hive user (either individual or organization)
- It needs to have a fully qualified domain name that the given Hive user owns and is in full control.
- Preferred port is
2001
- It should not be the same node as the witness node, although it can serve as a backup node in emergency situations.
- It can be the same node as an API node.
- At least 100Mbps network and 99% uptime.