Closed akashin closed 3 weeks ago
I finished generating a GCP disk and image with desired state, it is available at crt-benchmark-state-28-06-2024
.
Some useful stats for when we decide to repeat the state generation:
n2d-standard-16
instance in just under 40 hours.locust -H 127.0.0.1:3030 \
-f locustfiles/ft.py \
--funding-key=$KEY \
--fixed-contract-names --num-ft-contracts=1 \
--num-passive-users=5000000 \
--max-workers=8 -u 1 -r 1 --headless --processes 8
data
folder) takes 1.2TB on diskTo use this state, one must use --fixed-contract-names
, --num-ft-contracts=1
and --num-passive-users
flags.
For the future record, here is how the shard size looks like:
This is a follow-up to https://github.com/near/nearcore/issues/11359.
We now know how to generate the initial state but still need to start using it in the benchmarks. Concretely, I think we should follow the steps suggested in https://near.zulipchat.com/#narrow/stream/308695-nearone.2Fprivate/topic/Creating.20GCP.20disk.20image.20with.20initial.20NEAR.20state/near/445814332 for creating a new GCP image and then using this image in the Terraform configuration of CRT benchmark.