Issue(provider): [DO - Digitalocean] - Droplet create event failed, please try again

[rinor]

Are there any new known issues on DO that prevent successful nanos deployments?

For some reasons the instances (droplets) created by nanos images fail to be created/instantiated properly.

➜  ops image create boringtun-cli -c config.json -t do -z nyc3 -i boringtun   
do image 'boringtun' created...
➜  ops instance create boringtun -t do -c config.json
do instance 'boringtun-1673602491' created...

---

• config.json

  {
  "Program": "boringtun-cli",
  "Args": [
  "--foreground",
  "--disable-drop-privileges",
  "--verbosity",
  "info",
  "wg2"
  ],
  "Env": {
  "WG_CONFIG_FILE": "wireguard.wire"
  },
  "Files": [
  "wireguard.wire"
  ],
  "Dirs": [
  "var"
  ],
  "BaseVolumeSz": "10G",
  "RunConfig": {
  "Klibs": [
    "tun"
  ],
  "UDPPorts": [
    "51820"
  ],
  "CanIPForward": true
  },
  "ManifestPassthrough": {
  "tun": {
    "wg": {
      "ipaddress": "10.0.0.1",
      "netmask": "255.255.255.0",
      "up": "true"
    }
  }
  },
  "CloudConfig": {
  "Platform": "do",
  "Zone": "nyc3",
  "BucketName": "ops",
  "Flavor": "s-1vcpu-1gb"
  }
  }

  ---

  Note:

• Tried also to create the droplets manually from the ops uploaded image and still the same result.
• Tried both formats .qcow2 and .raw
• Tested with qemu versions 5.2 and 7.2
• Tested with Nanos version: 0.1.43, nightly, master branch

[francescolavra]

There are no known issues with running Nanos on DO, in fact I just tried creating a Nanos image and then a droplet, and everything seems to work fine: My droplet is running an example program instead of boringtun, but I don't think this makes a difference, it seems to me you encountered a problem with DigitalOcean. I'm using sfo3 as zone instead of nyc3, but this shouldn't make a difference either. I also tried creating the droplet manually from the uploaded image (also tried with different instance flavors), and it worked flawlessly. Tested with Nanos 0.1.43 and nightly. Perhaps there was a transient issue in DigitalOcean when you were using it? Your error message says "please try again in a few minutes".

[rinor]

Ok, thanks. Will try other zones and apps jic. Maybe I'm missing smth in the config file.

Perhaps there was a transient issue in DigitalOcean when you were using it? Your error message says "please try again in a few minutes".

It's 3 days in a row now.

Thank you.

[eyberg]

it could easily be booting and dying very fast - if you haven't tried a simple hello world there first I'd prob do that to see if you could even do that before tackling the wireguard app

[rinor]

Indeed started fresh again today.

• tested some simpler apps and they work perfectly.
• tested wireguard-go and it works as expected

Will get back to checking boringtun-cli, just to understand what the issue is and if it's related to nanos.

Thank you.

[rinor]

It turns up that on DO, the tun interface comes up as wg3 instead of wg2 and while wireguard-go can autoswitch to that wg3 even when instructed to expect wg2, boringtun-cli needs the exact name.

So indeed the instance was fast failing hence DO deleted it.