Back in October of 2020, I announced that I have upgraded my node api.deathwing.me to serve faster and with fewer issues. So far, in the last 5 months, it has done a great job, but... why not more?
So today, a couple of hours ago, I switched the DNS records for api.deathwing.me, changing the node once again.
| Node Specifications |
|---|
| CPU: AMD Ryzen 9 5950X |
| Memory: 128GB DDR4 ECC RAM |
| Storage: 8TB NVMe (in RAID0), 50TB Potato Disks |
Basically, every single thing (other than RAM, unfortunately, that's a hard cap for now) have been (nearly) doubled. Double the single-core performance and about 1.5x multicore performance as opposed to the previous node.
Here are the benchmark results from @fullnodeupdate (sourced from @ausbitbank's hive.ausbit.dev) alas, the only benchmark that the server fails (and others do too) is get_config, which I still couldn't really wrap my head around as to why.
That's all for today! Hopefully, the node should perform even better than before.
Thank you for reading through the post.
Thanks for having the best node on chain.
yay potatoes
Congratulations @deathwing! You have completed the following achievement on the Hive blockchain and have been rewarded with new badge(s) :
Your next target is to reach 350 posts.
You can view your badges on your board and compare yourself to others in the Ranking
If you no longer want to receive notifications, reply to this comment with the word
STOPCheck out the last post from @hivebuzz:
One of our latest internal servers here has pretty much the same config (except we don't have the 50TB worth of HDDs, that sounds more like one of our file servers). Are you leasing this in the cloud or did you buy your own?
Hey @blocktrades,
Potato disks are not HDDs, it was a joke that I made during my first announcement of the node. Regarding the lease/building, it's currently a lease. Though it's definitely in my plans one day to basically build one on my own and then colo it in somewhere. Unfortunately, PC/server part costs in Turkey are quite steep.
At the very least, I'd like to lease several servers in different datacenters/countries and Anycast it. In an ideal world, I'd love to have a server in NA, EU and SEA.
Potato disks:
Ah, that's too bad, was thinking could make a nice backup for our image server. It's pretty amazing what you can build in file servers now. We just built one with 8 16TB in a RAID configuration (90TB effective space after 2 disks worth of space allocated to error correction) and a 1 TB NVME drive as a cache. So 50TB didn't sound that unreasonable :-)
Hmm, it's really not that hard to get a very big fileserver with NVMe caching, just expensive.
Hopefully, one day. If I get to Top 20 and Hive goes up, the first thing I'd like to start with is having more nodes, then who knows, maybe we'll have decentralized image servers too, same with our API nodes? :)
Please spin up a full node in NA(preferably in Chicago). It sucks with most things being in EU.
BW costs are quite high in NA, driving up server prices. Unfortunately, investing in EU this much prevents me from investing in NA for the time being