Users can use **following new options** to configure memory management:
_\# The memory management switch in HugeGraph. Options: off-heap, on-heap, disable._
**memory.mode=off-heap**
_\# The maximum memory capacity that can be managed for all queries in HugeGraph._
**memory.max_capacity=1073741824**
_\# The maximum memory capacity that can be managed for a query in HugeGraph._
**memory.one_query_max_capacity=104857600**
_\# The alignment used for round memory size._
**memory.alignment=8**
**Detailed docs**: https://github.com/apache/incubator-hugegraph/wiki/%5BMemory-Management%5D-GSoC-2024-Final-Report
---------
Co-authored-by: imbajin <jin@apache.org>
* chore: upgrade sofa-rpc to latest version (2024.8)
* fix conflicts
* downgrade sofa-rpc to 5.12.
* fix: replace the wrong dep usage for sofa-rpc
* feat: support multi RPC desc type for users
* chore: disable sofa-rpc server by default
HugeGraph supports the parent-child edge feature, meaning that an Edgelabel can have a subordinate type. Using the bank transfer graph as an example, transfers may include person-to-person transfers (person-to-person), person-to-company transfers (person-to-company), and company-to-company transfers (company-to-company). These three different types of transfers share a common operation, transfer.
In actual business scenarios, it is often necessary to retrieve all transfer edges with a single query. Currently, competitors can only manually split the transfer edge types, perform multiple queries, and then aggregate the results. With HugeGraph's parent-child edge feature, it is possible to query the corresponding person-to-person transfers, person-to-company transfers, and other sub-edge types, and also conveniently and efficiently retrieve all transfer-related edges at once.
PS: The parent-child edge feature for the cassandra and scylladb backends has been temporarily disabled through the store feature.
> Code formatting will be done separately in the next PR.
Related to:
- https://github.com/apache/incubator-hugegraph/issues/745
- https://github.com/apache/incubator-hugegraph/issues/447
**TODO:**
After this PR merged, change the README in commons & pin a [issue](https://github.com/hugegraph/hugegraph-hubble/issues/371) to notify all users/devs to know the context (also update the repo description...)
Finally, we may need start a discussion for marking commons repo in achieved status
As title, change limit:
- vid max to 16KB
- eid max to 64kb (128k as backup)
- property max to 10MB (keep consistent)
fix#1593#2291
---------
Co-authored-by: imbajin <jin@apache.org>
subtask of #2265
When introducing hstore, server-core needs corresponding modifications. Except for BytesBuffer.java in hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/backend/serializer/, other files should be consistent with the pd-store branch.
---------
Co-authored-by: imbajin <jin@apache.org>
benny066567
VGalaxies
imbajin
vaughn
John
HaoJin Yang