🐞 [Bug]: Can't list deployments

[AhmedHanafy725]

Is there an existing issue for this?

• I have searched the existing issues

which package/s did you face the problem with?

Dashboard

What happened?

can't list the fullvm deployments

Steps To Reproduce

1- Deploy a full vm.
2- Wait to be listed after deployment

Note: I don't have any failed deployments

which network/s did you face the problem on?

Dev

version

2.3.2

Twin ID/s

No response

Node ID/s

No response

Farm ID/s

No response

Contract ID/s

No response

Relevant screenshots/screen records

Relevant log output

-

[A-Harby]

Verified, Devnet a9928b8.

Was able to list new deployed VM.

As well as the old ones.

[RoberL72]

not fixed. Ever since we changed to the new dash board I can see all the contracts but none of the 8 VM's deployed

[AhmedHanafy725]

@RoberL72 Can you provide more info? What are the contract IDs? Which network?

[RoberL72]

Van: AhmedHanafy725 ***@***.***> Verzonden: dinsdag 19 maart 2024 20:17 Aan: threefoldtech/tfgrid-sdk-ts ***@***.***> CC: RoberL72 ***@***.***>; Mention ***@***.***> Onderwerp: Re: [threefoldtech/tfgrid-sdk-ts] 🐞 [Bug]: Can't list deployments (Issue #2435) @RoberL72 <https://github.com/RoberL72> Can you provide more info? What are the contract IDs? Which network? — Reply to this email directly, view it on GitHub <#2435 (comment)> , or unsubscribe <https://github.com/notifications/unsubscribe-auth/A6RIJ2TM2EW2PP3RDFSFNVDYZCFK3AVCNFSM6AAAAABEWAFJJGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMBXHE2DINJZHE> . You are receiving this because you were mentioned. <https://github.com/notifications/beacon/A6RIJ2TPI2IDLOUJBWJXAW3YZCFK3A5CNFSM6AAAAABEWAFJJGWGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTTXV3GZO.gif> Message ID: ***@***.*** ***@***.***> >

[Cameron1986labeyrie]

VMs for the following contract IDs are not loading/available in the deployments list under full vms.

Contract : 229142
Contract : 257804

[0oM4R]

VMs for the following contract IDs are not loading/available in the deployments list under full vms.

Contract : 229142 Contract : 257804

@Cameron1986labeyrie could you please provide the logs ?

[Mahmoud-Emad]

@Cameron1986labeyrie

Seems to be that the issue persists, we'll actively work on resolving it. we'll investigate further and keep you updated on our progress. Please continue to monitor the issue for any updates. Thank you for your patience and understanding.

My contracts on devnet:

• 

My listed full VMs on devnet

• 

The mentioned failed Deployments

• 

[Mahmoud-Emad]

After conducting an investigation, it has come to my attention that we previously adhered to a specific pattern for storing machine details in the old playground. As part of the transition to the new dashboard, we introduced a change to this pattern and implemented a fix to update all instances of the old pattern to the new one. However, it appears that the fix is not functioning as expected.

Old Pattern:

New Pattern:

[0oM4R]

Investigation:
the type in deployment data is Fullvm, this can't be listed in fullvm deployment list, and should be migrated and listed in vm deployment list, trying to confirm that the migration happened and deployments appears on vm deployments table

[0oM4R]

I couldn't reproduce deployment with the mentioned deployment data. For now, you should find the missing deployments on vm deployment table.

[Cameron1986labeyrie]

Hey @Mahmoud-Emad It's a farmers' issue that was reported over the support chat, and from what I understood, it was a full VM deployment.

[tzumby]

I'm running into a similar problem, trying to list the deployments with this command used to work, but now it returns an empty array:

  const result = await gridClient.machines.getObj(vms);

This is the payload used to create the VM:

  const vms = {
    name: '608e9aa00666',
    network: {
      name: '9cec1820a56f',
      ip_range: '10.238.0.0/16',
      addAccess: undefined,
      accessNodeId: undefined,
      myceliumSeeds: undefined
    },
    machines: [
      {
        name: '52839ffc800d',
        node_id: 312,
        disks: [],
        qsfs_disks: undefined,
        public_ip: true,
        public_ip6: undefined,
        planetary: true,
        mycelium: false,
        myceliumSeed: undefined,
        cpu: 2,
        memory: 2048,
        rootfs_size: 0,
        flist: 'xxx',
        entrypoint: '/sbin/zinit init',
        env: [],
        ip: undefined,
        corex: undefined,
        solutionProviderId: 2,
        zlogsOutput: undefined,
        gpus: undefined
      }
    ],
    metadata: '',
    description: 'xxx node'
  };

[Mahmoud-Emad]

Hello @tzumby :)

I just deployed a machine from the grid-client and I was able to list it from the dashboard

this is the machine data I have used

  const vms: MachinesModel = {
    name,
    network: {
      name: "wedtest",
      ip_range: "10.249.0.0/16",
    },
    machines: [
      {
        name: "tobetested",
        node_id: 11,
        disks: [
          {
            name: "wedDisk",
            size: 8,
            mountpoint: "/testdisk",
          },
        ],
        public_ip: false,
        public_ip6: false,
        planetary: true,
        mycelium: false,
        cpu: 1,
        memory: 1024 * 2,
        rootfs_size: 0,
        flist: "https://hub.grid.tf/tf-official-apps/base:latest.flist",
        entrypoint: "/sbin/zinit init",
        env: {
          SSH_KEY: config.ssh_key,
        },
      },
    ],
    metadata: "",
    description: "test deploying VMs via ts grid3 client",
  };

I named the machine tobetested, and I would like to share that the deployment was successful!:

Here's a screenshot from the dashboard showing the deployed machine:

Please make sure that you are connecting to the dashboard by the same account wallet secret you used when you deployed the machine from the grid-client

As you can see, the machine tobetested is listed on the dashboard. To ensure a smooth experience:

1. Connect to the dashboard using the same account wallet secret used during machine deployment.
2. Enable the Show All Deployments switch to view all deployments.
3. Make sure you are connected to the same network on both the dashboard and the Grid client.

[tzumby]

Hey @MohamedElmdary , I can already see them in the Dashboard, please read my comment, they don't show up when I try to call getObj

[Mahmoud-Emad]

Hey @MohamedElmdary , I can already see them in the Dashboard, please read my comment, they don't show up when I try to call getObj

If that is the case, I think you need to send the name of the deployment not the vms interface

  const name = "newVMS";
  const grid3 = await getClient(`vm/${name}`);

  const vms: MachinesModel = {
    name,
    network: {
      name: "wedtest",
      ip_range: "10.249.0.0/16",
    },
    machines: [
      {
        name: "testvm",
        node_id: 11,
        disks: [
          {
            name: "wedDisk",
            size: 8,
            mountpoint: "/testdisk",
          },
        ],
        public_ip: false,
        public_ip6: false,
        planetary: true,
        mycelium: false,
        cpu: 1,
        memory: 1024 * 2,
        rootfs_size: 0,
        flist: "https://hub.grid.tf/tf-official-apps/base:latest.flist",
        entrypoint: "/sbin/zinit init",
        env: {
          SSH_KEY: config.ssh_key,
        },
      },
    ],
    metadata: "",
    description: "test deploying VMs via ts grid3 client",
  };

const deployed = await client.machines.deploy(vms);
const deploymentDetails = await client.machines.getObj(name);

[tzumby]

Let me check this, I'm pretty sure i had both name and vms as parameters there with the same result.

[tzumby]

Yes, confirmed that const deploymentDetails = await client.machines.getObj(name); does not work either

[Mahmoud-Emad]

@tzumby Any errors? logs?

[Mahmoud-Emad]

@tzumby

I just used your payload and was able to deploy/list the deployment

[
  {
    version: 0,
    contractId: 109538,
    nodeId: 11,
    name: '52839ffc800d',
    created: 1715170810,
    status: 'ok',
    message: '',
    flist: 'https://hub.grid.tf/tf-official-apps/base:latest.flist',
    publicIP: null,
    planetary: '302:9e63:7d43:b742:b2a4:5133:63ba:6b5',
    myceliumIP: '',
    interfaces: [ { network: '9cec1820a56f', ip: '10.238.2.2' } ],
    capacity: { cpu: 2, memory: 2048 },
    mounts: [],
    env: {},
    entrypoint: '/sbin/zinit init',
    metadata: '',
    description: 'xxx node',
    rootfs_size: 0,
    corex: false,
    gpu: []
  }
]
'================= Getting deployment information ================='
disconnecting

However, I faced some errors and I'm not sure if you faced some of them or not

1. The filst should be a valid URL for a real flist not xxx
2. The other errors were about the availability of reserving a public IP and the deployment name that was taken by you, as you see I just added one more char to fix it
3. The env should be an object {} not an array []

[tzumby]

Yes, I redacted those fields out for the flist and the name. There's no errors. The deploys are going successfully, it's just that the deployments returns an empty array. In the logs I can see the contract ids and everything, it seems like there's something wrong with getting them after deploying

[tzumby]

Did you try this and successfully returned the deployments ?

const deploymentDetails = await client.machines.getObj(name);

[Mahmoud-Emad]

Yeah as you see in the previous comment I listed it

[
  {
    version: 0,
    contractId: 109538,
    nodeId: 11,
    name: '52839ffc800d',
    created: 1715170810,
    status: 'ok',
    message: '',
    flist: 'https://hub.grid.tf/tf-official-apps/base:latest.flist',
    publicIP: null,
    planetary: '302:9e63:7d43:b742:b2a4:5133:63ba:6b5',
    myceliumIP: '',
    interfaces: [ { network: '9cec1820a56f', ip: '10.238.2.2' } ],
    capacity: { cpu: 2, memory: 2048 },
    mounts: [],
    env: {},
    entrypoint: '/sbin/zinit init',
    metadata: '',
    description: 'xxx node',
    rootfs_size: 0,
    corex: false,
    gpu: []
  }
]
'================= Getting deployment information ================='
disconnecting

[Mahmoud-Emad]

Issue update

Work done

a new migration function added to the Dashboard to migrate the old Fullvm deployment types to vm

[Mahmoud-Emad]

Issue update

Work done

• Updated the import of 'BaseModule' in 'migration.ts' to be local dir instead of the dist one to fix the un-exported issue
• Flat the array of the contracts then loop over them all
• Updated the moduleName after listing the 'Fullvm' to 'vm' again to list the other deployments
• Pushed the newly changed contract to the 'newContracts' array to avoid the lag of gql
• Updated the 'moduleName' in 'base.ts' to take 'Fullvm' if the 'moduleNames' is undefined

[Mahmoud-Emad]

issue update

Work done

• Replaced the 'exact' search with 'contains' algo by adding 'includes' when searching on the 'projectName'.
• Introduced a new field 'updatedAt' in the 'GqlNodeContract' type to store the updated time in it.
• Added a condition on checking on the 'updatedAt' field to replace the newly added contract with the Gql contract.

[khaledyoussef24]

can you mention where to find the latest fields of updatedAt and search to verify ?

[Mahmoud-Emad]

can you mention where to find the latest fields of updatedAt and search to verify ?

it's defined in the GqlNodeContract interface, you can deploy a machine using the gridClient/single_vm.ts script to see it.

[khaledyoussef24]

verified on staging
version : 408184e
changed the name of the single vm (type,project) in the machine files from vm to Fullvm
and deployed it using the grid client.

on the staging worked fine after @Mahmoud-Emad changes:

while in the dashboard it did not load(normal devnet):

https://app.testlodge.com/a/26076/projects/40893/suites/217765?expand_section=370856#case_3783024