starrocks: [FE] stream load failed with "no 100-continue header" under HTTP/1.0 protocol

Steps to reproduce the behavior (Required)

$ curl -0 -i -kv  -H "label:auto_test_stream_load_table_basic2" -u root: -T heap-after.bin -X PUT http://172.26.95.190:8030/api/auto_test_stream_load/auto_test_stream_load_table_basic/_stream_load
> PUT /api/auto_test_stream_load/auto_test_stream_load_table_basic/_stream_load HTTP/1.0
> Host: 172.26.95.190:8030
> Authorization: Basic cm9vdDo=
> User-Agent: curl/7.81.0
> Accept: */*
> label:auto_test_stream_load_table_basic2
> Content-Length: 1379507030
> 
< HTTP/1.1 200 OK
< content-length: 113
< content-type: application/json; charset=UTF-8
< connection: close
< 
{
  "Status": "FAILED",
  "Message": "class com.starrocks.common.DdlException: There is no 100-continue header"
}

Expected behavior (Required)

Real behavior (Required)

StarRocks version (Required)

3.3.6

Affected version:

  • v3.3.6

About this issue

  • Original URL
  • State: closed
  • Created 2 years ago
  • Comments: 21 (11 by maintainers)

Most upvoted comments

either downgrade to http 1.0 protocol with curl -0 or manually add -H 'Expect: 100-Continue' to the curl command. The tpcds script/library needs a revise.

Success! I expand curl command in py-file with -H 'Expect: 100-Continue' and it works now.

Thank you a lot, @kevincai !

@dimkas78 can’t repro it. How did you launch the allin1 container?

@kevincai, sure, I changed only mysql_host row

[starrocks]
# for mysql cmd
mysql_host: 127.0.0.1
mysql_port: 9030
mysql_user: root
mysql_password:
database: tpcds

# cluster ports
http_port: 8030
be_heartbeat_port: 9050
broker_port: 8000

# 0 means the pipeline engine automatically adjusts the parallelism
parallel_num: 0
concurrency_num: 1
num_of_queries: 1
sleep_ms: 500

[broker_load]
broker: starrocks
broker_username: hdfs_broker
broker_password:

hadoop_home: /tools/hadoop-2.7.7

# data size per broker load job: 500M
max_bytes_per_job: 524288000

# format and column separator
file_format: orc
column_separator:

# job properties
# The maximum tolerance rate of job. 0 ~ 1
max_filter_ratio: 0
# timeout
timeout: 14400

@dimkas78 how did you setup your sr cluster? vm-based or docker allin1 or k8s env?

It is docker-allin1 on top of brand-new Ubuntu 24.04.

I have bare metal home server, so I install fresh Ubuntu 24.04 to empty disk, add containerd, curl, python3, pip, then run docker command from https://docs.starrocks.io/docs/quick_start/shared-nothing/#launch-starrocks and go to TPC-DS script.