Closed parasj closed 2 years ago
I suspect this issue is caused by far too high parallelism for object store queries. This may be fixed as we switched to RBAC authentication.
Tracked by SKY-224
_____ _ ____ _______ _ ___ _ _ _____ / ___| | / /\ \ / / ___ \ | / _ \ | \ | || ___| \ `--.| |/ / \ V /| |_/ / | / /_\ \| \| || |__ `--. \ \ \ / | __/| | | _ || . ` || __| /\__/ / |\ \ | | | | | |____| | | || |\ || |___ \____/\_| \_/ \_/ \_| \_____/\_| |_/\_| \_/\____/ Will transfer 1152 objects totaling 69.76GB from azure:eastus to azure:westus VMs to provision: 1x azure:eastus, 1x azure:westus Estimated egress cost: $1.40 at $0.02/GB fake_imagenet/train-00000-of-01024 => train-00000-of-01024 fake_imagenet/train-00001-of-01024 => train-00001-of-01024 fake_imagenet/train-00002-of-01024 => train-00002-of-01024 fake_imagenet/train-00003-of-01024 => train-00003-of-01024 ... fake_imagenet/validation-00124-of-00128 => validation-00124-of-00128 fake_imagenet/validation-00125-of-00128 => validation-00125-of-00128 fake_imagenet/validation-00126-of-00128 => validation-00126-of-00128 fake_imagenet/validation-00127-of-00128 => validation-00127-of-00128 Using overridden docker image: ghcr.io/skyplane-project/skyplane:local-7ab4bd091242ee5f5832a558525190ee Instances will remain up and may result in continued cloud billing. Remember to call `skyplane deprovision` to deprovision gateways. Storing debug information for transfer in /tmp/skyplane/transfer_logs/20220628_011227/client.log ✓ Installing gateway package (2/2) in 12.34s ✓ Assigning gateways permissions to buckets (2/2) in 4.55s 🚀 69.76GB transfer job launched Warning: It can take up to 60s for role assignments to propagate on Azure. See issue #355 Transfer progress (1152 of 1152 chunks) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 69.8/69.8 GiB 1.6 GB/s 0:00:00 Compression saved 0.47% of egress fees AzureCliCredential.get_token failed: Failed to invoke the Azure CLI AzureCliCredential.get_token failed: Failed to invoke the Azure CLI AzureCliCredential.get_token failed: Failed to invoke the Azure CLI AzureCliCredential.get_token failed: Failed to invoke the Azure CLI ... AzureCliCredential.get_token failed: Failed to invoke the Azure CLI AzureCliCredential.get_token failed: Failed to invoke the Azure CLI AzureCliCredential.get_token failed: Failed to invoke the Azure CLI AzureCliCredential.get_token failed: Failed to invoke the Azure CLI AzureCliCredential.get_token failed: Failed to invoke the Azure CLI AzureCliCredential.get_token failed: Failed to invoke the Azure CLI AzureCliCredential.get_token failed: Failed to invoke the Azure CLI ✓ Verifying transfer (1152/1152) in 186.15s
fixed in #505 and #507
I suspect this issue is caused by far too high parallelism for object store queries. This may be fixed as we switched to RBAC authentication.
Tracked by SKY-224