[Komal Thareja]

There was a missing configuration on RUTG due to the maintenance being lifted from one of the hosts there. Please check again now; the interfaces should be visible on your VMs.

Thank you for reporting this issue and allowing us to address the misconfiguration.

Thanks,

Komal

1 user thanked author for this post.

Garegin Grigoryan

[Komal Thareja]

Thanks,

Komal

 

• This reply was modified 1 year, 6 months ago by Komal Thareja.

[Komal Thareja]

Hi Nirmala,

Are you a member of multiple projects? If so, could you please try the following and see if this helps?

From the portal Go to Experiments -> Projects and Slices, choose the specific project, click on Slices under that project.

Please let us know if this is still an issue.

Alternatively, you could renew the slices from Jupyter Hub with one of the following options:

Option A: Slice commander

• Open a terminal, type slice-commander
• type ls to list your slices
• cd to your slice and then type renew <days>

Option B: Notebook

• List your slices using notebook Start Here -> List All Slices (available under Managing Slices)
• Renew your slices using notebook Start Here -> Extending a Slice Reservation (available under Managing Slices)

 

Thanks,

Komal

 

[Komal Thareja]

Could you please share your slice id? ID you shared earlier is your Project ID.

Thanks,

Komal

1 user thanked author for this post.

Garegin Grigoryan

[Komal Thareja]

Hi Garegin,

I suspect you are using ubuntu image for your VMs. Please note for ubuntu, the interfaces are not up by default.

Please install net-tools using the following command:

apt install net-tools

You can then verify the interfaces via the command: ifconfig -a

Thanks,

Komal

[Komal Thareja]

Hi Laura,

Please check your slice on Portal. I can confirm all the resources requested by your slice are provisioned and are in Active State. I suspect /home/fabric/work/fabric_config/ssh_config is not setup correctly.

Could you please check if you see any errors in /tmp/fablib/fablib.log ?

Also, please try following steps:

• Remove the file /home/fabric/work/fabric_config/ssh_config
• Run the notebook jupter-examples-1.6.1/configure_and_validate.ipynb
• Recreate your slice

Thanks,

Komal

[Komal Thareja]

Hi Sunjay,

As you correctly pointed, only the contents of the work directory persist across container restarts. I would recommend setting up your local Jupyter Environment for any customized experience as indicated here: https://learn.fabric-testbed.net/knowledge-base/install-the-python-api/#install-jupyter-in-the-virtual-environment

We are working on providing a containerized access to Jupyter as well where users can launch the container on their desktop/laptop and use that. This should be available soon and would enable user environment customization.

Thanks,

Komal

[Komal Thareja]

Portal currently display overall site TACC disk usage which is combined disk space on all the hosts at TACC. Control Framework determines possible candidate nodes for your VM on the basis of the resources requested.

Your slice is requesting for a VM with CX5, which in case of TACC are only available on tacc-w4 and hence it is trying to allocate it on tacc-w4 but fails due to no disk space. I would recommend using a different site than TACC.

Also, we are working on improving the Resource Usage display to show per Host level information.

Thanks,

Komal

 

[Komal Thareja]

Hi Nishanth,

The error Insufficient resources : [disk] implies that there is not enough disk available on the host on which your VM is being requested. Looking at your slice, following VM requesting a ConnectX5 is being rejected as it maps to tacc-w4  There is not enough disk available on tacc-w4 to accomodate your VM hence the failure.


Reservation ID: 478b2a91-5a02-4cf0-9bcd-de04c3b873ea Slice ID: 30f9fb42-37be-420f-899e-082a41bfb735
Resource Type: VM Notices: Reservation 478b2a91-5a02-4cf0-9bcd-de04c3b873ea (Slice Traffic Listening Demo TACC(30f9fb42-37be-420f-899e-082a41bfb735) Graph Id:a58b7bc7-55d6-42e9-b457-5a8a32ebebc9 Owner:nshyamkumar@iit.edu) is in state (Closed,None_) (Last ticket update: Insufficient resources : ['disk'])
Start: 2024-06-05 17:55:24 +0000 End: 2024-06-06 17:55:23 +0000 Requested End: 2024-06-06 17:55:23 +0000
Units: 1 State: Closed Pending State: None_
Predecessors
Sliver: {'node_id': '9a579143-79b2-44fb-bacb-e6a5db4da3bf', 'capacities': '{ core: 2 , ram: 8 G, disk: 1 G}', 'capacity_hints': '{ instance_type: fabric.c2.m8.d10}', 'image_ref': 'default_ubuntu_20', 'image_type': 'qcow2', 'name': 'TACC_node4', 'reservation_info': '{"reservation_id": "478b2a91-5a02-4cf0-9bcd-de04c3b873ea", "reservation_state": "Closed"}', 'site': 'TACC', 'type': 'VM', 'user_data': '{"fablib_data": {"instantiated": "False", "run_update_commands": "False", "post_boot_commands": [], "post_update_commands": []}}'}
Component: {'node_id': '670d117f-19ac-477b-bff7-36ac4e90107a', 'details': 'Mellanox ConnectX-5 Dual Port 10/25GbE', 'model': 'ConnectX-5', 'name': 'TACC_node4-pmnic_2', 'type': 'SmartNIC', 'user_data': '{}'}
NS: {'node_id': 'adeede90-a808-45a6-8e1e-8c8de7a4ee6e', 'layer': 'L2', 'name': 'TACC_node4-TACC_node4-pmnic_2-l2ovs', 'site': 'TACC', 'type': 'OVS'}
IFS: {'node_id': '2f9a52b4-3108-48f2-b0f9-e0ccd7716cdc', 'capacities': '{ bw: 25 Gbps, unit: 1 }', 'labels': '{ local_name: p1}', 'name': 'TACC_node4-pmnic_2-p1', 'type': 'DedicatedPort', 'user_data': '{"fablib_data": {"mode": "config"}}'}
IFS: {'node_id': 'b6c42c3e-a570-4ed1-b633-607e90777f34', 'capacities': '{ bw: 25 Gbps, unit: 1 }', 'labels': '{ local_name: p2}', 'name': 'TACC_node4-pmnic_2-p2', 'type': 'DedicatedPort', 'user_data': '{"fablib_data": {"mode": "config"}}'}


Thanks,

Komal

[Komal Thareja]

Hi Nishanth,

This issue has been fixed for a while now but is only available in Beyond Bleeding Edge Container.

Could you please use that? This should be available in the pypi with the next release.

Thanks,

Komal

[Komal Thareja]

Hi Acheme,

We investigated the possibility of enabling UEFI mode for users but encountered issues where GPUs do not function in that mode. Consequently, we have opted to maintain updated firmware to mitigate these errors for users. Could you please rerun your experiment and inform us if the error persists? I am available to collaborate with you on upgrading the firmware and addressing the issue.

Thanks,

Komal

[Komal Thareja]

The updated network model has been deployed and the maintenance is complete.

[Komal Thareja]

Hi Emmanuel,

It is not possible to recover a deleted slice. Apologies we may not be able to recover your data. However, you should be able to request renewal of an expired project though.

Thanks,

Komal

[Komal Thareja]

To clarify, requesting two VMs is acceptable. However, requesting VMs with GPUs and SmartNICs in the mentioned slice is invalid because none of the hosts have SmartNICs and GPUs available on the same host.

[Komal Thareja]

Hello Khawar,

Your slice is requesting 2 VMs. This is unsupported configuration. On UTAH, we have two hosts each with 3 GPUs but none of them have a dedicated CX-6. So your slice configuration is seemed unsupported.

Also, I checked, all 6 RTX-6000 GPUs are in use. Please note that the resource usage displayed on the portal may be outdated by 30 minutes.

• n1 – with RTX-6000 GPU and dedicated NIC CX-6
• n2 – with two RTX-600 GPU

We do have ongoing work for users to identify such invalid slice configurations using fablib API. This should be available soon with the upcoming Release 1.7. We also plan to provide host level resource usage details to user in 1.7 that may help with this too. Hope this helps!

Thanks,

Komal

[Author]

Posts