Mismatch between portal hardware availablility and resource availability

[lam...@gmail.com]

Hi,

I have been trying to instantiate some nodes for compute heavy experiments. However, I see that there's mismatch between the resource availability information provided by resource availability and portal hardware availability pages. I have attached screenshots showing ibm and r6525 nodes as an example.

It seems like the experiment instantiating process uses the same information as portal hardware leading to instantiation failures.

Could you please look into this? Also, for near future, which information should I use while instantiating experiments? 

Thanks,

Srivattsan

[Mike Hibler]

I believe that the latter (portal hardware page) does not take into account
nodes type reservations, so the former is probably more accurate. However,
to ensure you get the more popular node types, you should make reservations
(https://www.cloudlab.us/resgroup.php) for the types you need. Nodes with
GPUs, lots of RAM or NVMe, or 100Gb interfaces tend to be almost constantly
in use.

> --
> You received this message because you are subscribed to the Google Groups
> "cloudlab-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email
> to cloudlab-user...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/cloudlab-users/
> 9494532a-5d54-49c5-b78a-94243d5d83c2n%40googlegroups.com.