Google Cloud Platform Testnet Nodes Are Broken!

Nodes & Mining Problems
#1

Hello,

I mined on the Testnet successfully for 2-3 months then took a break, I have tried to start again in the last few days, here is what I have encountered so far:

  • Digital Ocean Full Node Deployment (from marketplace) appears to work, however, I setup 10 Nodes, they have been running for over 72 hours now and I haven’t received a single reward. Yes my wallet.dat is correct, yes my walled.pswd is correct, all 10 nodes are synced to the current block height and show a state of “Persist Finished” on NKNX.

  • Google Cloud Platform Deployment (from marketplace) does NOT work anymore. After installing the block height is stuck at around half of the current block height and it will not increase. In fact, monitoring running processes following deployment of the full node I noticed that the NKN Service appears to be stuck in a loop of attempting to update, then running the main NKN Process, then crashing/exiting, then attempting to update again and this continues with (I repeat) no increase in block height.

  • I tried to install on Google Cloud Platform manually following these installation instructions: https://medium.com/nknetwork/setting-up-a-nkn-miner-in-5-minutes-and-run-it-free-for-2-months-with-digitalocean-76bafcd82ae8 (minus the auto-updating part because that never worked for me before anyway). I have followed these instructions so many times that I pretty much know them off by heart now. Following installation and launch the NKN Nodes are stuck at very low block heights and don’t seem to be increasing anymore after several hours, one has: Node block height: 816 another has 2696 - many are stuck on zero.

It is my understanding from what I have read that we will be able to continue earning rewards on the Testnet beyond the release of the Mainnet and we will still be able to convert these to NKN tokens so I hope these issues will be fixed soon.

Thanks

#2

Update: So I did eventually get a couple of rewards through from the Digital Ocean Nodes.

I tried another host and managed to setup Nodes manually which are working fine.

Google Cloud Nodes are still having issues and will not synchronise with the block-chain, they all seem to get stuck at random (low) block heights if I set them up manually and if I use the marketplace deployment they will be stuck at about 50% of the current block height, not sure what is causing that but I assume it’s related to their network, I did make sure all ports were open and I have had working NKN Nodes on GCP before.

#3

I’m personally running a GCP 1-click with f1-micro (1 vCPU, 0.6 GB memory) instance in US-Central zone, and it has automatically upgraded itself to v0.8.3 and mining properly.

My other personal AWS 1-click is working properly as well on v0.8.3. Also auto-upgrade.

My 10 DigitalOcean nodes are working properly as well.

#4

I have tried again with GCP but it’s still not working, I’m assuming you’ve had that instance running for a while because deploying from the marketplace currently just leads to it the node getting stuck in a loop of nknd terminating itself and failing to synchronise with the chain. I’ve tried multiple regions as well.

#6

Stuck at this blockheight: image

Error log: https://pastebin.com/2jCpSWyc

#7

From the logs it seems that your node fails to sync block header, which we never saw before. Do you have the log file when the error occurs for the first time? The one you posted is not the first one and thus do not contain enough information.

#8

Hi,

I have tried everything I can think of in GCP Control panel working with the Marketplace Deployment and also trying to manually install the Node on Debian which is working currently on all other hosts I try. Is it possible Google Cloud Network Engineers performed some traffic shaping/routing which is now defunct following the latest updates?

I’m really frustrated/sick of GCP’s failure to work with NKN at the moment and tired of looking at Terminals trying to fix this :crazy_face: but I will start over with a new deployment one more time to retrieve you the log files, will update with the Pastebin’s later.

Thanks

#9

Also not sure if it’s related: you can also try to launch GCP one-click with a slightly better instance type (g1-small or n1) to see if it’s working properly (but please remember to delete it after testing if you are not intended to keep it)

#10

@ajs @yilun
Maybe occurred OOM during syncBlock on GCP f1-micro (1 vCPU, 600 MB memory)
Try to setting the SyncBatchWindowSize in config.json refer the Config.json complete reference for v0.9 [EN]