steem logo

---

that could lead to a denial of service. Current stable branch (2ee5160) has an appropriate fix. All publicly accessible nodes should be updated.

This bug, however, did not create any risk to the Steam blockchain, accounts, keys, or balances.
Last but not least:
I want to thank you (know who) for reporting and handling it in a professional way.
Good job, as always :-)

Dear witness, please keep your nodes in shape
If you are a witness, it is highly recommended to support the network by running some nodes. Some of us run their services powered by one or many of such nodes, others have complex, distributed infrastructures for their R&D projects, while some are taking their first steps by running their own seed node. In either case, it is crucial to keep them in a good shape.
Security update mentioned above is a good excuse to do some maintenance checks:

do you still have enough storage for your block log and shared memory file?

For now, the main purpose of my node is to help various service providers with the transition from WebSockets to JSON RPC 2.0 requests by giving them a bit more time to adapt.
(Honestly, you guys should have done that a while ago.)

I
I’m writing this because I’m currently experiencing increased load on my node, which is caused not only by the high rate of requests from several big service providers that I know but also by a lot of new ones and tons of distributed, short-lived connections from random end-users.

That causes performance issues and when upstream speed(s) are too loaded to respond on time. Make sure that your software can handle such a situation gracefully.

Please be responsible while using public resources: if you expect a significant number of requests, you should consider running your local node, which would also greatly reduce latency, improve performance, and save your time.

(Seriously, when you need to throw 80 million of get_blocks on a public API node in one week, you are doing it wrong. PRO TIP: we don’t even have 20 million blocks yet.)

But fear not. I’m not complaining, I will scale it up. I just need a little time.

• connections coming to my node, real-time scale
  Another thing is that it’s not really about the performance of hardware as such, but about certain limitations of the current speed and the way it handles RPC connections. For now, I’ve just added the second server running seem to split the load. This issue will be addressed in one of the upcoming steam releases based on appease, which will greatly improve the performance of a single node.

If you are running a project that depends on my node, feel free to contact me when you encounter any issues or you need a dedicated node.

My main, public API endpoint, currently under test, is now working with artificially limited performance, occasionally falling back to external upstream servers. The final version will be powered by multiple separate stemmed nodes optimized for various purposes and microservices to serve you better and faster.

---

▶️ DTube
▶️ IPFS