Upgrade to release 3.7.0
Step 1: Prerequisites
- Inform users about the upgrade: stop creating new deliveries, transactional support on CRUD models will be disallowed.
- Stop the i-refactory server.
Step 2: Upgrading i-refactory metadata model
Follow the generic instructions for upgrading the metadata repository as described in Upgrading the i-refactory repository model in SQLServer.
This upgrade script will:
Create the following database columns:
acm_uam_datadel.s_delivery_properties.csv_loader_specification
Modify the following columns:
acm_uam_datalog.s_entity_update_status.message_textto nvarchar(max)acm_uam_datalog.s_interface_update_status.message_textto nvarchar(max)acm_uam_datalog.s_task_run_status.message_textto nvarchar(max)
Create the following index(es):
CREATE NONCLUSTERED INDEX [ix01] ON [acm_uam_datalog].[s_task_run_properties] ([l_delivery_id])
The upgrade is executed within a transaction. An error will result in a rollback. Contact i-Refact support in case of unexpected failures.
Step 3: Upgrade the i-refactory server to release 3.7.0
- Copy the complete folder i-refactory-engine to the appropriate location.
- Start the server, the upgrade will be executed automatically.
Changes applied during startup
During the startup of the i-refactory server the following changes will be applied automatically:
- The application version is updated to release 3.7.0
- For each temporal context table in the fact store (a table with transaction datetime in it's primary key) an additional unique nonclustered index will be created with a generated name
irefactory_generated_ak_idxif such an index does not exist. - The i-refactory generic data access interface objects are regenerated (views, procedures, triggers, ...).
Create the index irefactory_generated_ak_idx prior before upgrading
Creating the additional indexes upon restart of the i-refactory engine may take quite some time. You could opt for creating these addtional indexes prior before restarting the i-refactory server.
The following SQL script generates a create index statement. You can apply filters for which model or namespace you want to generate these index statements.
Copy/paste the content of the column create_index_statement to a new window and simply run the statements or wrap them in a stored procedure.
with context_entities as
(
select e.id
, i.code as interface_code
, i.database_code
, e.namespace_code
, e.code
, e.key_id_primary
from acm_datadef_oasi_bk.entity e
join acm_datadef_oasi_bk.interface i
on i.id = e.interface_id
join acm_datadef.[unique_key] k
on k.id = e.key_id_primary
where e.architecture_layer_code='CFPL'
and e.classification_code not in ('STABLE INDEPENDENT', 'STABLE DEPENDENT')
and exists
(
select 1
from acm_datadef.unique_key k
join acm_datadef.key_attribute ka
on ka.key_id = k.id
join acm_datadef_oasi_bk.attribute a
on a.id = ka.attribute_id
and a.classification_code='TRANSACTION TIMELINE START'
where k.id = e.key_id_primary
)
)
, index_columns as
(
select e.id
, string_agg( '[' + a.code + ']', ',') WITHIN GROUP (ORDER BY coalesce(ka.ordinal_nbr,999)) as index_columns
from context_entities e
join acm_datadef_oasi_bk.attribute a
on a.entity_id=e.id
left
join acm_datadef.key_attribute ka
on ka.attribute_id= a.id
and ka.key_id = e.key_id_primary
where a.classification_code != 'TRANSACTION TIMELINE START'
and (ka.id is not null or a.classification_code='TRANSACTION TIMELINE END')
group by e.id
)
select e.interface_code
, e.database_code
, e.namespace_code
, e.code as entity_code
, 'IF NOT EXISTS
(
SELECT *
FROM [' + e.database_code + '].[sys].[indexes]
WHERE name = ''irefactory_generated_ak_idx'' AND object_id = OBJECT_ID(''[' + e.database_code + '].[' + e.namespace_code + '].[' + e.code + ']'')
) AND EXISTS
(
SELECT 1 WHERE OBJECT_ID(''[' + e.database_code + '].[' + e.namespace_code + '].[' + e.code + ']'') IS NOT NULL
)
BEGIN
create unique nonclustered index irefactory_generated_ak_idx on [' + e.database_code + '].[' + e.namespace_code + '].[' + e.code + '](' + ic.index_columns +');
END;
' as create_index_statement
from context_entities e
join index_columns ic
on ic.id = e.id
where e.interface_code = 'TPC_H_CFPL'
;This is an example for a single entity for which an additional index should be created if it does not yet exist:
IF NOT EXISTS
(
SELECT *
FROM [IR_CFPL].[sys].[indexes]
WHERE name = 'irefactory_generated_ak_idx' AND object_id = OBJECT_ID('[IR_CFPL].[tpc_h].[h_brand_properties]')
)
AND EXISTS
(
SELECT OBJECT_ID('[IR_CFPL].[tpc_h].[h_brand_properties]')
)
BEGIN
create unique nonclustered index irefactory_generated_ak_idx on [IR_CFPL].[tpc_h].[h_brand_properties]([id],[acm_end_dt]);
END;