Forum Replies Created
-
Posts
-
Greg, could you please delete this slice and recreate it?
We had some leftover layer3 connections from testing which were causing the issue. We were able to identify and clear them. It should work now. Please let us know if you still face this issue.
Hi Chengyi,
As discussed on slack, I looked at your slice: e10be783-2ef6-407e-bb9b-14a1103317a9
Slivers for the slice do have the updated lease time on the orchestrator. But the update could not be reflected at the Aggregates as the token used to trigger the renew had expired. I would also try to reproduce this issue on our development setup and work on providing a fix for it.
All the slivers report the error:
Reservation ID: 2b5bf824-8472-4af9-b4c0-b8137fd27f45 Slice ID: e10be783-2ef6-407e-bb9b-14a1103317a9
Resource Type: VM Notices: Reservation 2b5bf824-8472-4af9-b4c0-b8137fd27f45 (Slice SciStream-demo-more-accurate(e10be783-2ef6-407e-bb9b-14a1103317a9) Graph Id:0ae993ab-2775-4b15-befe-3f4e2ec3fb3d) is in state (Closed,None_) (Last ticket update: Unable to validate provided token: ValidateCode.INVALID/Signature has expired) (Ticket events: Unable to validate provided token: ValidateCode.INVALID/Signature has expired)Thank you for sharing the observations. fabric-cli creates FABRIC tokens for the user which are then used to authenticate/authorize with the testbed. The lifetime of FABRIC identity token is 1 hour.
A new token is created if the expiry time is < 30 mins from the current time. The tokens are cached in ~/.tokens.json.
In the above scenario, even though the project id was changed, the cached token may not have been re-requested due to the time-limit as explained above thus resulting in the Create Slice being successful.
We would work on addressing this in the next release. Appreciate your feedback.
fabric-cli currently does a json dump of the Slice object thus resulting in the escaping of the inner json objects. We would work on making this output more parseable in the next releases until then using the FABlib interface which returns the Python objects is recommended.
I’ll add ipysheet in the default image with the next build. Also, you can restart your container from File -> Hub Control Panel -> Stop Container followed by Start Container.
Workaround to copy the following files seems to make conda happy. No luck finding the config parameter.
cp /opt/conda/conda-meta/certifi-2021.10.8-py39hf3d152e_2.json /opt/conda/conda-meta/certifi-2021.5.30-py39hf3d152e_0.json
cp /opt/conda/conda-meta/ca-certificates-2021.10.8-ha878542_0.json /opt/conda/conda-meta/ca-certificates-2021.5.30-ha878542_0.jsonBy default ca-certificates.json exists and is not corrupted. After the first install using the conda, ca-certificates and certifi package gets updated. Even though ca-certificates and certifi packages were updated, conda still seems to be looking for older versions. Thus resulting in any subsequent install failures. I am looking at ways on how to reset conda to avoid such failures. In the meanwhile, could you please share the packages you need to install for your notebook? I can see if I can add them to the default notebook container image.
Snapshot before the install.
I am able to reproduce the error and found this to be a bug in the code. We would address it in the next release.
User should be able to get to the respective node components using the get_slice API:
`status, topology = slice_manager.get_slice(slice_id=slice_id)
if status == Status.OK:
print(f”Properties: {topology.nodes[‘n1’].list_properties()}”)
print(f”Components: {topology.nodes[‘n1’].components}”)
print(f”Component: {topology.nodes[‘n1’].components[‘c1’]}”)
else:
print(f”Failed to get slice topology: {topology}”)`However, I think we have bug which is preventing the components from being listed in the output of the above snippet.
