70M Hive blocks

Managed to capture this on live logs for today's post, with blockchain size hitting 360GB compressed and growing.

...and 3 years

In approximately 6 days would be my 3-year anniversary since I produced the first block (#38707240). Since then, I have successfully signed 3,040 blocks and over 100K transactions contained within them. Huge shoutout to all of you who voted for making this possible, your support is appreciated!

techcoderx.com hiccups

Onto the real stuff, the API server had some weird I/O errors on one of the drives in the ZFS pool that corrupted the DBs as hived and hivemind froze along with PostgreSQL.

It may seem to be a drive failure as the same error occurred on both available M.2 slots even with PBO2 disabled. However, upon research reseating the power supply cables on the motherboard likely fixed the issue.

The ZFS pool had to be re-created and everything replayed.

Last hardware upgrade of the year

To finish off the API server hardware upgrade in the previous post, I made an in-socket CPU upgrade from the 1700 (Zen 1) to a 5700X (Zen 3) for only $187, thanks to the support for Ryzen 5000 CPUs on B350 motherboards (with a BIOS update of course).

The node feels a lot faster now as replays (and index/foreign key creation) took significantly less time to complete and API calls took way less time to respond.

Have been running it for at least 3 weeks with PBO2 -30 with +200MHz offset on all cores, however 2 of the cores crashed today (unrelated to the I/O error above as confirmed), so will be running it at stock going forward as the extra performance aren't worth the extra time needed to further tune the curve optimizer.

Development updates

Definitely an overdue update as the previous mention of it was around a year ago.

While I had unrelated projects in 1H 2022, most of the recent work was on OneLoveIPFS v3. The highly anticipated v3 release are already feature frozen and complete, only things left are likely bug fixes of remote encoders, UX enhancements and finalizing the process of 3Speak cross-platform posting by calling the API to pin the resulting upload to our own IPFS node. On-track for final release by this coming Christmas.

As for the 2.5 year old Alive Protocol (@aliveprotocol), some work will be required to convert HAlive (HLS livestream indexer for Hive) into a HAF app along with implementation of RTMP endpoint for hosting services with Qt app integration.

---

Witness performance

Current rank: 74th (active rank 69th - nice)
Votes: 7,726 MVests
Voter count: 286

Producer rewards (7 days): 44.607 HP
Producer rewards (30 days): 200.129 HP
Missed blocks (all-time): 29 <-- much better on my NUC so far

Server resource statistics

Last recorded 2 days ago before I/O incident.

hived (v1.27, 26f38cd2, all plugins)

block_log file size (compressed): 360 GB
block_log.artifacts file size: 1.6 GB
shared_memory.bin file size: 21 GB

HAF db

Output of SELECT pg_size_pretty( pg_database_size('block_log') );
Database size: 3,200 GB

hivemind (v1.26.1, 28f86d5a)

Output of SELECT pg_size_pretty( pg_database_size('hive') );
Database size: 551 GB

Overall Postgres database

RAM usage: 20 GB
Compressed disk usage: 1.5 TB

---