-
By @qiuxin2012 Pyspark unit test blocking error in the second to last test `SGX=1 ./pal_loader bash -c "cd /ppml/trusted-big-data-ml/work/spark;/usr/bin/python -m pyspark.sql.tests.test_udf"`:
```
…
-
### Description of the problem
root@bai-OptiPlex-5080:/home/bai/examples/pytorch# gramine-direct ./pytorch ./pytorchexample .py
[P1:T1:] error: libos_init() failed in init_exec_hand…
-
Due to large occlum base image (6.22GB), PPML occlum image is close 9.3GB. It's too large for external developer to pull from dockerhub. Need to reduce size.
- [x] [Reduce occlum image size](https…
-
By @qiuxin2012
I check the python and scala failed tests, the python fail happened when operate the datetime or timestamps.
Scala also has similar failed tests in https://github.com/apache/spark/…
-
- [x] API interface (generate & verify)
- [x] JNI & Script Dummy implementation
- [x] SGX SDK based Quote Verifier
- [x] SGX SDK Verifier
- [ ] Open enclave Verifier
- [x] SGX…
-
- [x] Occlum Spark GBT Local 10GB
- [x] Occlum Spark GBY K8S 10GB
- [x] Occlum Spark GBT K8S 50GB with unlimited EPC & Memory
- [x] Occlum Spark GBT K8S 50GB with limited EPC & Memory (300GB EPC)
…
-
- [x] SGX=1 for running in SGX
- [x] Test with non SGX
- [x] Test with SGX
-
By @qiuxin2012 most failure on graphene is because of the inconsistent in graphene filesystem.
Exception 1: java.io.IOException: Failed to list files for dir: /tmp/spark-9cb334e7-94c2-4a0b-9f22-55c3b…
-
- [x] Occlum Spark Local TPC-H 100GB
- [x] Occlum Spark K8S TPC-H 100GB with unlimited resource
- [x] Occlum Spark K8S TPC-H 100GB with 100GB EPC
- [x] Occlum Spark K8S TPC-H 100GB with 50GB EPC
…
-
Thank you for building this example repo. I am trying to run a modified version of `AzureNytaxi.scala` and while building the Docker image from source `bash build-docker-image.sh`, the Docker build ha…