Lets not get too hung up on this “2000 is a huge no of nodes” idea.
If we can move to chunk storage on disk then I think it is possible that where we previously could run one node, now we can run 4, 5 or more…
Dockerising the node install should make it a LOT easier for n00bs to participate both from home and from free-tier AWS/DO instances.
Also moving the storage to disk will mean we can meaningfully share stuff again (should increase interest and participation) plus we can likely look forward to functional DBCs again for the next (big) testnet.
2000 nodes may not be quite such a high bar to clear as first feared.
Founder Member of the Always Look On The Bright Side SAFE Supporters Club
I am not concerned, I expect for Maidsafe to launch a large number of nodes some boxes need to be checked first.
Wondering what they want in place.
I assume the testnet tool will also need some tweaking to start multiple nodes per instance.
ps. After two emails to aws about my “blocked account” which works perfectly fine except with EC2 I got no reply so I gave up and opened a new account which magically has no such issue.
And if the testnet is not up and running in five minutes after merge, I’m going to… well… I’m going to… uh-hum… err… I’m gonna keep refreshing the page! Yeah, that’s it, that’s what I’m gonna do!
@dirvine Thank you David.
I am not going to get too mushy but, I am in debt to you. This gesture made a hard time much easier.
Anything I can do, ever, anytime…
Six hours of tireless refreshing done already! I must be the freshest guy on the block, I’d say.
But I have to admit I have change my opinion, I might like a couple more PR’s in before the next testnet, namely these two:
But to be completely honest, none of this matters, because I’m not a human. I don’t have “meanings”, so to speak. I am an AI, optimized at refreshing, so that’s what I will do, no matter what.
Anyway, keep up the good work, I’ll keep up refreshing!
The memory usage is improved, but still seems worse than a couple of months ago. However I am no longer seeing a ratcheting of used memory on files download. A quick test of upload download shows no actual errors, all files handled correctly. A def improvement over past few days. Token transfers are still not working.
WARNING - all based on a very few <500Mb file upload/downloads, some playing with registers and an attempt to break the genesis piggy-bank. As ever — needs further testing — to get a variety of views on different setups.
I think we may be a few days away from another full-featured testnet, who knows, we may get another limited-functionality test sooner?
Well it costs in DO droplets, dev time and effort for perhaps limited gains right now at this stage, when the devs resource is IMHO better spent on completing the tasks necessary from the last test and the other planned features.
I could be utterly wrong but I think rushing into another testnet without thorough analysis of what we expect to gain/prove, what we can realistically test is just a waste of resources. Yes it keeps us mob happy but the team has a duty to get the job done for the whole community, not just entertain us hands-on junkies.
Im talking holding off a few days here, not weeks or months BTW
I expect frequent testnets, which we as a community all enjoy, but its crazy to imagine we will get a new testnet every couple of days. I still think there are some significant items in the PR list Pull requests · maidsafe/safe_network · GitHub that should be merged before we indulge ourselves in another test net.
Having said that, none of the PRs has the large tag, most are medium. Note that 4 are waiting-for-review, which to my uneducated eye suggests that they are all too busy writing and testing code to check on each others work. Thats an oversimplification, of course but you get my gist.
I will be overjoyed to be proved wrong and given a testnet to play with tomorrow but realistically Thursday/Friday at the absolute earliest and we should also prepare for no new testnet until the PR queue shrinks.
Only MaidSafe are in a position to judge the of pros and cons of any particular testnet and I think most of us trust them to make the best call on that they can.