Geth or Mist do not sync any more [duplicate]

Question

I am new to ethereum but managed to successfully start Geth on 2 of my machines , and could also install MIST. Now, since several days, one of my machines does not sync with the testnet anymore. I start Geth using geth --testnet console I see these lines

0228 19:31:57.086460 eth/backend.go:191] Protocol Versions: [63 62], Network Id: 3
I0228 19:31:57.086640 eth/backend.go:219] Chain config: {ChainID: 3 Homestead: 0 DAO: <nil> DAOSupport: true EIP150: 0 EIP155: 10 EIP158: 10}
I0228 19:31:57.089224 core/blockchain.go:217] Last header: #524061 [ac6dd60d…] TD=166615159323361
I0228 19:31:57.089241 core/blockchain.go:218] Last block: #524061 [ac6dd60d…] TD=166615159323361
I0228 19:31:57.089248 core/blockchain.go:219] Fast block: #524061 [ac6dd60d…] TD=166615159323361

then starting server , the endpoint opens, but admin.peers gives me an empty answer [] It is not a time sync problem as i tried to sync with ntp but without changing the clock The production net seems to work as have imported blocks.

Stopping geste and starting MIST browser has the same result.

Does anybody has an idea what I could do to restart the testnet ? what do I have to save in order to not loose my test accounts ? (i guess the UTC--xxx files from testnet/geth/keystore.)

My other machine the latest block (on the testnet) is 588099. It started with

Block synchronisation started
Synchronisation failed: block download cancelled (requested)
imported block 588100 and then 
Peer xxxxxx: header #596654 broke chain ancestry

would appreciate any help

Answer 1

There's currently (still?) an attack on the Ropsten test network:

Ropsten testnet is under kind of attack? What can we do?

what do I have to save in order to not loose my test accounts ? (i guess the UTC--xxx files from testnet/geth/keystore.)

Yes, back up your ~/.ethereum/testnet/keystore/ directory before doing anything else. (Obviously it's only the testnet, but it'll potentially save you future hassle.)

Does anybody has an idea what I could do to restart the testnet ?

There's a set of instructions for users of Parity on the thread I posted above, which basically performs a soft-fork from just before the spam attack started.

As you're using Geth, not Parity, you could try reinitialising the chain to use the altered ropsten-revert.json file that was created by the Parity folks. I think using the init command in Geth will give the same result as the --chain flag in Parity. Note that I haven't tried this in Geth, so your mileage may vary.

• Remove your current ~/.ethereum/testnet/chaindata folder
• (I don't know how to restore from a snapshot, like the Parity case. Perhaps this isn't even possible on Geth.)
• Initialise with the new .json file: geth --testnet init <path/to/ropsten-revert.json>
• Start Geth with whatever options you normally use, e.g.: geth --testnet --fast --rpc --rpcapi db,eth,net,web3,personal --cache=1024 --rpcport 8545 --rpcaddr 127.0.0.1 --rpccorsdomain "*"