Skip to content

Instantly share code, notes, and snippets.

@wey-gu

wey-gu/reproduce_v1.md

Last active November 13, 2023 11:14
Show Gist options
  • Save wey-gu/81417425bb1ed39f8b219fc3debc79b9 to your computer and use it in GitHub Desktop.
Save wey-gu/81417425bb1ed39f8b219fc3debc79b9 to your computer and use it in GitHub Desktop.
Download ZIP
NebulaGraph v1 scan data after schema alter
Raw
reproduce_v1.md

Quickly setup env

git clone --branch v1.0 https://github.com/vesoft-inc/nebula-docker-compose.git
cd nebula-docker-compose
docker-compose up -d


docker run --rm -ti --network=host vesoft/nebula-console:v1.2.1 -u root -p nebula --addr=127.0.0.1 --port=3699

Prepare data

# ngql
# https://docs.nebula-graph.com.cn/1.2.0/manual-CN/2.query-language/4.statement-syntax/1.data-definition-statements/create-space-syntax/
CREATE SPACE test(partition_num=1, replica_factor=1);

USE test;

CREATE TAG course(name string, credits int);

INSERT VERTEX course (name, credits) VALUES 100:("cs", 12);

Scan data

nGQL

(root@nebula) [test]> FETCH PROP ON * 100;
===========================================
| VertexID | course.name | course.credits |
===========================================
| 100      | cs          | 12             |
-------------------------------------------
Got 1 rows (Time spent: 7.981/9.31 ms)

StorageClient

ref: https://github.com/vesoft-inc/nebula-python/blob/v1.0/examples/StorageClientExample.py

We could see data is in storaged1:

(root@nebula) [test]> show hosts
=============================================================================================
| Ip         | Port  | Status | Leader count | Leader distribution | Partition distribution |
=============================================================================================
| 172.28.2.1 | 44500 | online | 0            | No valid partition  | No valid partition     |
---------------------------------------------------------------------------------------------
| 172.28.2.2 | 44500 | online | 1            | test: 1             | test: 1                |
---------------------------------------------------------------------------------------------
| 172.28.2.3 | 44500 | online | 0            | No valid partition  | No valid partition     |
---------------------------------------------------------------------------------------------
| Total      |       |        | 1            | test: 1             | test: 1                |
---------------------------------------------------------------------------------------------
Got 4 rows (Time spent: 1.349/2.74 ms)

from compose file:

  storaged1:
    image: vesoft/nebula-storaged:v1
    environment:
      USER: root
      TZ:   "${TZ}"
    command:
      - --meta_server_addrs=172.28.1.1:45500,172.28.1.2:45500,172.28.1.3:45500
      - --local_ip=172.28.2.2
      - --ws_ip=172.28.2.2
      - --port=44500
      - --data_path=/data/storage
      - --log_dir=/logs
      - --v=0
      - --minloglevel=0
    depends_on:
      - metad0
      - metad1
      - metad2
    healthcheck:
      test: ["CMD", "curl", "-f", "http://172.28.2.2:12000/status"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 20s
    ports:
      - "44501:44500"
      - 12000
      - 12002
# python3 -m pip install nebula-python


meta_client = MetaClient([("127.0.0.1", 45502)], 3000, 3)
code =  meta_client.connect()
if code == ErrorCode.E_FAIL_TO_CONNECT:
    raise Exception('connect to %s:%s failed' % (sys.argv[1], sys.argv[2]))
# initialize a StorageClient. Set its connection timeout to 3000ms, retry times to 3
storage_client = StorageClient(meta_client, 3000, 3)
# initialize a ScanVertexProcessor to process scanned vertex data
scan_vertex_processor = ScanVertexProcessor(meta_client)

def get_return_cols(space):
    vertex_return_cols = {}
    edge_return_cols = {}

    tags_name = meta_client.get_tags_name(space)
    for tag_name in tags_name:
        vertex_return_cols[tag_name] = []

    edges_name = meta_client.get_edges_name(space)
    for edge_name in edges_name:
        edge_return_cols[edge_name] = []

    return vertex_return_cols, edge_return_cols

def process_vertex(space, scan_vertex_response):
    result = scan_vertex_processor.process(space, scan_vertex_response)
    if result is None:
        return None
    # Get the corresponding rows by tag name
    for tag_name, tag_rows in result._rows.items():
        for row in tag_rows:
            vid = row._default_properties[0].get_value()
            props = {}
            for prop in row._properties:
                prop_name = prop.get_name()
                prop_value = prop.get_value()
                props[prop_name] = prop_value
            print(props)


space_to_read = "test"
vertex_return_cols, edge_return_cols = get_return_cols(space_to_read)


scan_vertex(space_to_read, vertex_return_cols, True)

Now we could see it's scanned:

In [38]: scan_vertex(space_to_read, vertex_return_cols, True)
    ...:
############# scanned vertex data #############
{'name': 'cs', 'credits': 12}

Reproduce issue

ref: https://docs.nebula-graph.com.cn/1.2.0/manual-CN/2.query-language/4.statement-syntax/1.data-definition-statements/alter-tag-syntax/

Now let's alter schema:

USE test;
ALTER TAG course ADD (address string);

Now, the version in data was not matched with it was in metad:

In [39]: scan_vertex(space_to_read, vertex_return_cols, True)
############# scanned vertex data #############
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[39], line 1
----> 1 scan_vertex(space_to_read, vertex_return_cols, True)

Cell In[3], line 9, in scan_vertex(space, return_cols, all_cols)
      7     logging.error("Error occurs while scaning vertex")
      8     break
----> 9 process_vertex(space, scan_vertex_response)

Cell In[37], line 2, in process_vertex(space, scan_vertex_response)
      1 def process_vertex(space, scan_vertex_response):
----> 2     result = scan_vertex_processor.process(space, scan_vertex_response)
      3     if result is None:
      4         return None

File ~/.local/lib/python3.8/site-packages/nebula/ngStorage/ngProcessor/ScanVertexProcessor.py:60, in ScanVertexProcessor.process(self, space_name, scan_vertex_response)
     58         tag_name = tag_id_name_map[tag_id]
     59         default_properties = row_reader.vertex_key(scan_tag.vertexId, tag_name)
---> 60         properties = row_reader.decode_value(scan_tag.value)
     61         rows[tag_name].append(Row(default_properties, properties))
     62 else:

File ~/.local/lib/python3.8/site-packages/nebula/ngData/data.py:105, in RowReader.decode_value(self, value, schema_version)
    103     properties.append(self.get_double_property(field, value))
    104 elif property_type == SupportedType.STRING:
--> 105     properties.append(self.get_string_property(field, value))
    106 elif property_type == SupportedType.TIMESTAMP:
    107     properties.append(self.get_timestamp_property(field, value))

File ~/.local/lib/python3.8/site-packages/nebula/ngData/data.py:167, in RowReader.get_string_property(self, name, value)
    165     val = str(value[self._offset:self._offset+strLen])
    166 else:
--> 167     val = str(value[self._offset:self._offset+strLen], encoding='utf-8')
    168 self._offset += strLen
    169 return Property(SupportedType.STRING, name, val)

TypeError: unsupported operand type(s) for +: 'int' and 'NoneType'

Instead, let's try with db_dump:

wget https://github.com/vesoft-inc/nebula/releases/download/v1.2.1/nebula-1.2.1.el7-5.x86_64.rpm
rpm2cpio nebula-1.2.1.el7-5.x86_64.rpm | cpio -idmv
docker cp usr/local/nebula/bin/db_dump nebula-docker-compose_storaged1_1:/usr/local/nebula/bin
docker exec -it nebula-docker-compose_storaged1_1 sh

# in storaged1

./db_dump --space=test --db_path=/data/storage/nebula/ --meta_server=172.28.1.3:45500

And we could scan the data:

sh-4.2# ./db_dump --space=test --db_path=/data/storage/nebula/ --meta_server=172.28.1.3:45500
===========================PARAMS============================
mode: scan
meta server: 172.28.1.3:45500
space: test
path: /data/storage/nebula/
parts:
vids:
tags:
edges:
limit: 1000
===========================PARAMS============================

[vertex] key: 1, 100, course value: cs, 12,
===========================STATISTICS============================
COUNT: 1
VERTEX COUNT: 1
EDGE COUNT: 0
TAG STATISTICS:
	course : 1
EDGE STATISTICS:
============================STATISTICS===========================
Time cost: 77 us

Add limit=0 to scan all instead of sampling data:

sh-4.2# ./db_dump --space=test --db_path=/data/storage/nebula/ --meta_server=172.28.1.3:45500 --limit=0
===========================PARAMS============================
mode: scan
meta server: 172.28.1.3:45500
space: test
path: /data/storage/nebula/
parts:
vids:
tags:
edges:
limit: 0
===========================PARAMS============================

[vertex] key: 1, 100, course value: cs, 12,
===========================STATISTICS============================
COUNT: 1
VERTEX COUNT: 1
EDGE COUNT: 0
TAG STATISTICS:
	course : 1
EDGE STATISTICS:
============================STATISTICS===========================
Time cost: 102 us
@wey-gu
Copy link
Author

wey-gu commented Nov 13, 2023

db_dump ref: https://docs.nebula-graph.com.cn/1.2.0/manual-CN/3.build-develop-and-administration/5.storage-service-administration/data-export/dump-tool/

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment