So, the latest is, my local node actually managed to upload a ~700Mb file, so I suppose the problem was with the instance, maybe runnng out of memory now that the client is making most of the work? Interesting…
Anyhow, if you manage to get this, here’s a gem for reggae fans, but not only…
I tried most of the recently uploaded files and could get none of them.
Also added several new files and could only get one of them back, so it is still breathing but not really usable.
I got 0.40 running from home, the network is on its last legs though so switching out one cloud instance at a time seems an unnecessary headache at this point.
Thanks to everyone that participated, I believe begblagandsteal wins the the prize for most persistent data!
Going to put my nodes to bed now.
I’ll need to sit down and look at my DO account, it only updates every 24 hours or so. That said it all depends on how many instances, the size of the instances and obviously how long it runs. So no easy answer.
I felt shitty taking it down without any notice last night but my view is that I would rather not waste money once it is crippled and rather spend that money on a next iteration.
Maidsafe’s default is 4vcpu 8gb but for this last network I used 2vcpu 4gb which is much cheaper, I do wonder if that is a bad idea and is it detrimental to performance. Surely they choose what they do for good reason.
If we were to use 4vcpu8gb as a standard it becomes less feasible for me to spin these up on a regular basis, also I am not sure how many cloud instances are really needed. I had 14 to 18 nodes with the last test but is that enough? Will more be beneficial, IDK. (@joshuef any insight into minimum requirements would be greatly appreciated)
I am torn about accepting any contributions as I will end up feeling obligated, much more so if the BGF is involved @Southside.
With relatively short runs I am not yet anxious about the cost.
I will however think about posting a donation address with the next one thank you.
Please post your costs so I can work out if its something I can put up for a few days.
If IF the BGF get involved then its likely that a few of us could share the load and no-one need feel a fantastic burden because there will be 3 or 4 others there read to help keep it going. Just thinking out loud here…
I don’t know what a 4vcpu8gb from DO costs but I doubt its as much as $1k/mo. Given the resources of the BGF and the benefits of a continually running (code allowing) testnet then this seems like an awful lot of bang for relatively little buck in the terms of the BGF resources.
Well this should be our ultimate aim. Correct me if I’m wrong on this…
To get a viable test network we need 7 Elders plus some adults. We can really only run one node each from a home ADSL connection. So unless we can organise a minimum of 8 of us, we are stuck with DO or AWS to kickstart it. Of course we should be looking to get as much participation as possible and to ditch the cloud provider (and the expense) ASAP.