You upgraded the what now?
You may have noticed that @duplibot has not posted in a few days. Let me show you why, in the most direct terms possible.
# curl -s https://api.steemit.com --data '{"jsonrpc": "2.0", "method": "get_version" }' | jq '.result.blockchain_version'
"0.19.4"
It looks like it's upgrade season around here.
All of my work relies on a local instance of SBDS that I maintain, basically the entire blockchain in a MySQL database around 450GB and growing rapidly. To support that even more I deployed a new dedicated server, my own upgrade, a couple of days ago with these specs.
- Dual Intel Xeon 2.13Ghz Processors
- 16GB DDR3 ECC RAM
- 2TB SATA
- 1 Gbit Network
Nothing big by today's full node standards, but definitely a solid machine for the sort of work I'm pushing at it.
SBDS syncs the blockchain to this database by connecting to a public RPC node. I've been connecting to api.steemit.com though I have tried literally every other public RPC node I could find. I can't confirm for sure, but it does appear most of these other public nodes rate limit requests because I can't pull down more than about 1 block per second, sometimes less, while api.steemit.com has pushed blocks to me in the thousands per second on a few occassions.
In any case, api.steemit.com itself was upgraded as they pushed this public RPC node to a Release Candidate version of steemd without annoucement or any notice whatsoever. Or, that I have found at least, please correct me if I missed this. This version rolls out AppBase, which is a significant rearchitecting of everything, with obvious breaking changes.
This is unquestionably a rant-worthy topic, but I will forgo that indulgence today.
Private node for syncing SBDS
This upgrade rolled out while my new server was syncing and everything broke. This is why I haven't posted the past few days, I can't get my hands on the data. I've been again trying various other public RPC nodes and they work but the blocks barely trickle in. Even before the upgrade, api.steemit.com was unreliable on a good day with everything sort of day by day.
All of this makes it clear I should deploy my own node to query these transactions. I've played around with steemd enough that I'm more than comfortable doing this, but I don't want to deploy more than I need. A full RPC node is insane today and far out of my reach, but I have read that you can also deploy on SSD instead of relying on massive amounts of RAM. I got some time in today working on a new deployment but there's a few outstanding questions I'm hoping some of you may have insights on.
My primary question is what do I need to set for enable-plugin and public-api in my config.ini for SBDS to make its calls? I've tried walking through the code to answer this question but didn't get too far.
Also, what issues should I look out for running this on SSD? I'm only finding passing references to this and no posts explicitly talking about deploying steemd on SSD drives so I'm sort of shooting in the dark right now.
In due time I'll figure all this out, but time is always the issue so I'm hoping to get some insights here to move me forward a little quicker.
When will you post again!?
While I'm trying to solve my own data issues I'm again turning to the public infrastructure available. There's a few public SBDS instances that are usually at least a day behind, but still available for older content so I'm now looking at ways to use those as an interim solution.
I've started work collecting some comments around 6 days with pending rewards nearing payout and I'll have that post up shortly.
As I've mentioned before, I'm continually learning more and obviously there are a few bumps in the road to overcome but I'm ready for the challenge and looking forward to working hard on the fight against spam and unearned rewards!