honeymoose
/
hapi-fhir
mirror of https://github.com/hapifhir/hapi-fhir.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'ja_20200206_multitenancy' of github.com:jamesagnew/hapi-fhir into ja_20200206_multitenancy

This commit is contained in:
jamesagnew 2020-04-21 20:55:09 -04:00
parent 6226381595 cc2b7452be
commit 7efd8a9f6e
3 changed files with 15 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
hapi-fhir-docs/src/main/resources/ca/uhn/hapi/fhir/docs/server_jpa/partitioning.md
View File

@ -6,9 +6,9 @@ Partitioning allows each resource on the server to be placed in a partition, whi
Partitioning is designed to be very flexible, and can be used to achieve different outcomes. For example:
* Partitioning could be used to achieve **multitenancy**, where there are multiple logically separate pools of resources on the server. Traditionally this kind of setup is desired when each of these pools belongs to a distinct user group / organization / customer / etc. (a "tenant"), and each of these tenants should not be able to access or modify data belonging to anther tenant.
* Partitioning could be used to achieve **multitenancy**, where there are multiple logically separate pools of resources on the server. Traditionally this kind of setup is desired when each of these pools belongs to a distinct user group / organization / customer / etc. (a "tenant"), and each of these tenants should not be able to access or modify data belonging to another tenant.
* Partitioning could also be used to **logically separate data coming from distinct sources** within an organization. For example, patient records might be placed in one partition, lab data sourced from a lab system might be placed in a second partition and patient surveys from a survey app might be placed in another. In this situation data does not need to be completely segregated (lab Observation records may have references to Patient records in the patient partition) but these partitions might be used to create security groups, retention policies, etc.
* Partitioning could also be used to **logically separate data coming from distinct sources** within an organization. For example, patient records might be placed in one partition, lab data sourced from a lab system might be placed in a second partition and patient surveys from a survey app might be placed in another. In this situation data does not need to be completely segregated (lab Observation records may have references to Patient records in the patient partition) but these partitions might be used to support security groups, retention policies, etc.
* Partitioning could be used for **geographic sharding**, keeping data in a partition that is geographically closest to where it is likely to be used.
@ -35,22 +35,22 @@ At the database level, partitioning involves the use of two dedicated columns to
* **PARTITION_ID** – This is an integer indicating the specific partition that a given resource is placed in. This column can also be *NULL*, meaning that the given resource is in the **Default Partition**.
* **PARTITION_DATE** – This is a date/time column that can be assigned an arbitrary value depending on your use case. Typically, this would be used for use cases where data should be automatically dropped after a certain time period using native database partition drops.
When partitioning is used, these two columns will be populated with the same value on all resource-specific tables (this includes [HFJ_RESOURCE](./schema.html#HFJ_RESOURCE) and all tables that have a foreign key relationship to it including [HFJ_RES_VER](./schema.html#HFJ_RES_VER), [HFJ_RESLINK](./schema.html#HFJ_RES_LINK), [HFJ_SPIDX_*](./schema.html#search-indexes), etc.)
When partitioning is used, these two columns will be populated with the same value for a given resource on all resource-specific tables (this includes [HFJ_RESOURCE](./schema.html#HFJ_RESOURCE) and all tables that have a foreign key relationship to it including [HFJ_RES_VER](./schema.html#HFJ_RES_VER), [HFJ_RESLINK](./schema.html#HFJ_RES_LINK), [HFJ_SPIDX_*](./schema.html#search-indexes), etc.)
At the time that a resource is being **created**, an [interceptor hook](#partition-interceptors) is invoked in order to request the partition ID and date, and these will be written to the resource.
When a new resource is **created**, an [interceptor hook](#partition-interceptors) is invoked to request the partition ID and date to be assigned to the resource.
At the time that a resource is being **updated**, the partition ID and date from the previous version will be used.
When a resource is **updated**, the partition ID and date from the previous version will be used.
When a **read operation** is being performed (e.g. a read, search, history, etc.), a separate [interceptor hook](#partition-interceptors) is invoked in order to determine whether the operation should target a specific partition. The outcome of this hook determines how the partitioning manifests itself to the end user:
* If all read operations are scoped by the interceptor to only apply to a single partition, then the partitioning behaves as a **multitenant** solution.
* If read operations are scopes to all partitions, then the partitioning is simply partitioning the data into logical segments.
* The system can be configured to operate as a **multitenant** solution by configuring the partition interceptor to scope all read operations to read data only from the partition that request has access to.```
* The system can be configured to operate with logical segments by configuring the partition interceptor to scope read operations to access all partitions.
# Enabling Partitioning in HAPI FHIR
Enabling partitioning on the server involves a set of steps.
Follow these steps to enable partitioning on the server:
The [PartitionConfig](/apidocs/hapi-fhir-jpaserver-model/ca/uhn/fhir/jpa/model/config/PartitionConfig.html) bean contains configuration settings related to partitioning within the server. To enable partitioning, the [setPartitioningEnabled(boolean)](/apidocs/hapi-fhir-jpaserver-model/ca/uhn/fhir/jpa/model/config/PartitionConfig.html#setPartitioningEnabled(boolean)) property should be enabled.
The [PartitionSettings](/apidocs/hapi-fhir-jpaserver-model/ca/uhn/fhir/jpa/model/config/PartitionSettings.html) bean contains configuration settings related to partitioning within the server. To enable partitioning, the [setPartitioningEnabled(boolean)](/apidocs/hapi-fhir-jpaserver-model/ca/uhn/fhir/jpa/model/config/PartitionSettings.html#setPartitioningEnabled(boolean)) property should be enabled.
The following settings can be enabled:
@ -67,7 +67,7 @@ This interceptor can implement the hooks shown below.
## Identify Partition for Create (Required)
A hook against the [`Pointcut.STORAGE_PARTITION_IDENTIFY_CREATE`](/apidocs/hapi-fhir-base/ca/uhn/fhir/interceptor/api/Pointcut.html#STORAGE_PARTITION_IDENTIFY_CREATE) pointcut must be registered, and this hook method will be invoked every time a resource is being created in order to determine the partition to create the resource in.
A hook against the [`Pointcut.STORAGE_PARTITION_IDENTIFY_CREATE`](/apidocs/hapi-fhir-base/ca/uhn/fhir/interceptor/api/Pointcut.html#STORAGE_PARTITION_IDENTIFY_CREATE) pointcut must be registered, and this hook method will be invoked every time a resource is created in order to determine the partition the resource is assigned to.
The criteria for determining the partition will depend on your use case. For example:
@ -77,7 +77,7 @@ The criteria for determining the partition will depend on your use case. For exa
## Identify Partition for Read (Optional)
A hook against the [`Pointcut.STORAGE_PARTITION_IDENTIFY_CREATE`](/apidocs/hapi-fhir-base/ca/uhn/fhir/interceptor/api/Pointcut.html#STORAGE_PARTITION_IDENTIFY_CREATE) pointcut must be registered, and this hook method will be invoked every time a resource is being created in order to determine the partition to create the resource in.
A hook against the [`Pointcut.STORAGE_PARTITION_IDENTIFY_CREATE`](/apidocs/hapi-fhir-base/ca/uhn/fhir/interceptor/api/Pointcut.html#STORAGE_PARTITION_IDENTIFY_CREATE) pointcut must be registered, and this hook method will be invoked every time a resource is created in order to determine the partition to assign the resource to.
## Example: Partitioning based on Tenant ID
@ -312,9 +312,9 @@ The following request body could be used:
# Limitations
Partitioning is a relatively new feature in HAPI FHIR (added in HAPI FHIR 5.0.0) and has a number of known limitations. If you are intending to use partitioning for achieving a multi-tenant architecture it is important to carefully consider these limitations.
Partitioning is a relatively new feature in HAPI FHIR (added in HAPI FHIR 5.0.0) and has a number of known limitations. If you are intending to use partitioning for achieving a multi-tenant architecture it is important to consider these limitations.
None of the limitations listed here are considered permanent. Over time the HAPI FHIR team are hoping to make all of these features partition aware.
None of the limitations listed here are considered permanent. Over time the HAPI FHIR team is hoping to make all of these features partition aware.
* **Server Capability Statement is not partition aware**: The server creates and exposes a single server capability statement, covering all partitions. This can be misleading when partitioning us used as a multitenancy strategy.
@ -330,4 +330,3 @@ None of the limitations listed here are considered permanent. Over time the HAPI
* **Search Parameters are not partitioned**: There is only one set of SearchParameter resources for the entire system, and any search parameters will apply to resources in all partitions. All SearchParameter resources must be stored in the default partition.
* **Bulk Operations are not partition aware**: Bulk export operations will export data across all partitions.

3
hapi-fhir-docs/src/main/resources/ca/uhn/hapi/fhir/docs/server_jpa/schema.md
View File

@ -240,7 +240,7 @@ The complete raw contents of the resource is stored in the `RES_TEXT` column, us
By default, the **HFJ_RESOURCE.RES_ID** column is used as the resource ID for all server-assigned IDs. For example, if a Patient resource is created in a completely empty database, it will be assigned the ID `Patient/1` by the server and RES_ID will have a value of 1.
However, when client-assigned IDs are used, these may contain text values in order to allow a client to create an ID such as `Patient/ABC`. When a client-assigned ID is given to a resource, a row is created in the **HFJ_RESOURCE** table. When an **HFJ_FORCED_ID** row exists corresponding to the equivalent **HFJ_RESOURCE** row, the RES_ID value is no longer visible or usable by FHIR clients and it becomes purely an internal ID to the JPA server.
However, when client-assigned IDs are used, these may contain text values to allow a client to create an ID such as `Patient/ABC`. When a client-assigned ID is given to a resource, a row is created in the **HFJ_RESOURCE** table. When an **HFJ_FORCED_ID** row exists corresponding to the equivalent **HFJ_RESOURCE** row, the RES_ID value is no longer visible or usable by FHIR clients and it becomes purely an internal ID to the JPA server.
If the server has been configured with a [Resource Server ID Strategy](/apidocs/hapi-fhir-jpaserver-api/undefined/ca/uhn/fhir/jpa/api/config/DaoConfig.html#setResourceServerIdStrategy(ca.uhn.fhir.jpa.api.config.DaoConfig.IdStrategyEnum)) of [UUID](/apidocs/hapi-fhir-jpaserver-api/undefined/ca/uhn/fhir/jpa/api/config/DaoConfig.IdStrategyEnum.html#UUID), or the server has been configured with a [Resource Client ID Strategy](/apidocs/hapi-fhir-jpaserver-api/undefined/ca/uhn/fhir/jpa/api/config/DaoConfig.html#setResourceClientIdStrategy(ca.uhn.fhir.jpa.api.config.DaoConfig.ClientIdStrategyEnum)) of [ANY](/apidocs/hapi-fhir-jpaserver-api/undefined/ca/uhn/fhir/jpa/api/config/DaoConfig.ClientIdStrategyEnum.html#ANY) the server will create a Forced ID for all resources (not only resources having textual IDs).
@ -522,4 +522,3 @@ The following columns are common to **all HFJ_SPIDX_xxx tables**.
</tbody>
</table>

2
hapi-fhir-jpaserver-base/src/main/java/ca/uhn/fhir/jpa/dao/index/IdHelperService.java
View File

@ -282,7 +282,7 @@ public class IdHelperService {
}
}
if (pid.isPresent() == false) {
if (pid.isEmpty()) {
throw new ResourceNotFoundException(new IdDt(theResourceType, theId));
}
return pid.get();