Jump to content

Sukrim

Platinum Member
  • Content Count

    1,938
  • Joined

About Sukrim

  • Rank
    Veteran

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. https://groups.google.com/forum/#!topic/ripple-server/LXH80_aKomQ
  2. Maybe South Korea; Germany definitely does NOT test asymptomatic cases. retep's comments just sound like something that will turn out really badly in a few weeks and such people tend to delete their posts and accounts after that to not take the credibility hit. I try to just not interact with such people, they will self-select out anyways over the coming weeks and months.
  3. Maybe, but also because there is significant lag time between infection and a test result. If everyone stayed at home 24/7 and only sent in a test tube, case numbers would still go up for days while these tests are being processed. Hubei did a "China style" lockdown on Jan 23. at ~400 new daily cases. 18 days later, on Feb. 10 they had close to 3000(!) "new" cases still - 2 and a half weeks after lockdown. If you are citing numbers by the way, it might be useful to link to primary sources such as a ministry of health or similar.
  4. Not your keys, not your coins - not your code, not your wallet.
  5. Hire an expert in the field. In general you can know from each server at which point in time it currently is synced, so you can detect outages within seconds anyways and fail over.
  6. I don't and I want to see data and facts first. Random talking heads on Youtube are not really the kind of source I prefer, especially if they don't provide verifiable information.
  7. When was it "1 validator from a halt" - and do you actually know how many need to be offline for that to happen? Hint: it's far more than 20%...
  8. https://xrpl.org/currency-formats.html#issued-currency-precision "All issued currencies in the XRP Ledger are always divisible down to the minimum amount."
  9. The code for the Data API is Open Source, you can debug the issues yourself there and suggest fixes. Still a bad situation to be in of course and it seems to be a bit neglected by Ripple in general.
  10. Firstly: RippleNet is NOT the same as XRPL. You are likely talking about the latter.
  11. https://xrpcharts.ripple.com/ "XRP Ledger Trade volume" (https://xrpcharts.ripple.com/#/trade-volume is also a nice view) Yes. IOU issuers can control how "divisible" their issuances are (https://xrpl.org/ticksize.html), but your example should always be possible.
  12. Unknown unknowns. If you acccept random user input to a server that will deserializes that input, process the deserialized data and give a response based on internal state there's a ton of things that can go wrong. I don't think that rippled is bug free and I doubt anyone at Ripple thinks that either. This means the next best thing is to limit the things that your validator is exposed to. Ideally it only connects ot a single second server that serves as a "canary" (so if someone manages to crash servers remotely, or worse, then your validator is not affected) or as a "proxy" (any malformed request will already be rejected by the first server, your validator will only see valid messages unless there's an issue that manages to replicate over the whole peer network). Monitoring wise, you can already do a lot better than hoping that users complain after a while by the way, including automatic recovery if you use something more fancy like running in Kubernetes. Yeah, thanks a lot for that on both accounts (pushing internally and the domain verification part).
  13. If you're not on the recommended UNL de-facto currently you could turn off validation anyways as you have no say at all in the consensus process. Alternatively you should strive to get on that list (like apparently Tiffany did) and I certainly hope that servers on that list are not used to process API requests on the side. Ripple still operates more than one validator and did so since the start of the ledger by the way. I hope either https://github.com/ripple/rippled/issues/2601 gets implemented or there are enough validators out there that Ripple in their special process deems worthy to end up on their list. Maybe even third party list operators with a bit more transparent processes enter the scene?
  14. I kinda raised these points about 3 years ago by the way and complained again about a year later: Still there are no publicly verifiable criteria to get on or off that list... There is no "The UNL". A UNL is a private setting on each individual server, but by default it is set to a list that is auto-updated directly and only by Ripple with no clear criteria on who gets on it or why. Unfortunately Ripple also controls the distribution channels of the server software and thus they also control the defaults that are being set (instead of - for example - forcing every server operator to make a concious choice about settings). By far most server operators don't change defaults by the way and few guides on how to set up a server even explain how that would work, so in practice the Ripple Validator Registry is by far the most used one in the network at the moment.
×
×
  • Create New...