Enhance nodes IBD

[GitGab19]

This PR drastically improves the UX when using the tool on mainnet.
If it's the first time the tool is run on mainnet, the update-mainnet-chainstate.sh script will download the latest snapshot from our VPS, which serves a new snapshot every hour. It will do it also if more than 24 hours have passed since the last time the tool has been used in the mainnet (since it would take a lot of time to sync 3 pruned nodes together).

If we still want to support testnet3 (which I wouldn't), I think we should replicate this PR also for that since it takes a lot of time to sync it. Otherwise we can just support mainnet and testnet4 and open an issue to remove testnet3 from the options in the run-benchmarking-tool.sh and wherever it's mentioned.

Close #38

[Shourya742]

Concept ACK. Just few nits.

How many buffer snapshots are we keeping on the server? For example, if it takes me about 2 hours to download a snapshot, the if condition if wget $BACKUP_URL -O $BACKUP_FILE && wget $BACKUP_HASH_URL -O $BACKUP_HASH_FILE; then is gonna download the whole backup_file first, then the backup_hash_url to verify its authenticity. If we’re storing a lot of buffer snapshots, that’s fine (as long as it doesn't overload the server). But if we’re only keeping a few, like the last 5 hours’ worth, it could be a problem.

[Shourya742]

Getting some socket hung up issue:

[GitGab19]

Concept ACK. Just few nits.

How many buffer snapshots are we keeping on the server? For example, if it takes me about 2 hours to download a snapshot, the if condition if wget $BACKUP_URL -O $BACKUP_FILE && wget $BACKUP_HASH_URL -O $BACKUP_HASH_FILE; then is gonna download the whole backup_file first, then the backup_hash_url to verify its authenticity. If we’re storing a lot of buffer snapshots, that’s fine (as long as it doesn't overload the server). But if we’re only keeping a few, like the last 5 hours’ worth, it could be a problem.

Atm we're keeping last 3 hours. How many do you suggest to keep?

[GitGab19]

Getting some socket hung up issue:

Was the SV1 part working in this case or not?
I'm not experiencing that

[Shourya742]

rt-ACK. If we can have the IBD downloaded once and then mount it as a volume for the template provider or node containers, that would be ideal. Everything else looks good.

[Shourya742]

ACK

[Shourya742]

tACK 30066b2. Snapshot are getting refreshed for new downloads