Bulk sync relationships for global search (IBM Knowledge Catalog)

To start using the global search indexed data for relationships, you need to run the bulk sync utility after the upgrade.

If you are upgrading to 5.1.0 or later, you must complete both of the following tasks to be able to search for relationships using the global search bar and to sort by item name in the relationships table.

Prerequisites

Required roles

To complete this task, you must have one of the following roles:

Cluster administrator
Instance administrator

Clean up relationships metadata in global search

Run this procedure to clean up relationships metadata after you upgrade to 5.1.0 or later.

Do not run this procedure with any other payload than recommended by IBM. Once the metadata is cleaned there is no way to restore.

From your local environment, oc login as admin to the cluster.

Port forward elastic search pod to a port:

oc port-forward elasticsea-0ac3-ib-6fb9-es-server-esnodes-0 19200

Run the search call to verify:

curl  -X GET \
'http://localhost:19200/wkc-ALIAS/_search' \
--header 'Accept: */*' \
--header 'Content-Type: application/json' \
--data-raw '{"query":{"term": {"metadata.artifact_type": "relationship"}}}'

Delete the relationship documents:

curl  -X POST \
'http://localhost:19200/wkc-ALIAS/_delete_by_query' \
--header 'Accept: */*' \
--header 'Content-Type: application/json' \
--data-raw '{"query":{"term": {"metadata.artifact_type": "relationship"}}}'

Proceed to the bulk sync procedure for relationships.

Bulk sync relationships

Run a bulk sync for the global search index after upgrade. Ensure you are logged into the cluster as an administrator or through the command line.

Copy the following script into your local environment as cpd_gs_sync.sh:

 #!/bin/bash

 function prompt_for_namespace(){
     read -p "Please provide namespace in which WKC is installed: " NAMESPACE
     echo "Job will be deployed into \"$NAMESPACE\" namespace"
 }

 function all_container_confirmation(){
     read -p "ARE YOU SURE YOU WANT TO CONTINUE? THIS WILL AFFECT ALL CONTAINERS, AND MAY TAKE A LONG TIME. TYPE
     \"YES\" TO PROCEED: " PROCESS_ALL
     if [[ $PROCESS_ALL != "YES" ]]; then
         echo "Job run cancelled."
         exit 0
     fi
     echo "All containers will be processed"
 }

 function prompt_for_list_of_catalogs(){
     read -p "Please provide the list of containers to synchronize separated by commas, empty list indicates
     processing all containers(catalog/project/space): " CATALOGS_STRING
     IFS=', ' read -r -a CATALOGS <<< "$CATALOGS_STRING"
     if [[ $CATALOGS_STRING == "" ]]; then
         all_container_confirmation
     else
         echo "Following containers will be processed: "
         for index in "${!CATALOGS[@]}"
         do
             echo "$index - ${CATALOGS[index]}"
         done
     fi
 }

 function construct_db_list_string(){
     DB_STRING="["
     first=true
     for index in "${!CATALOGS[@]}"
     do
         if [ $first == false ]; then
             DB_STRING+=","
         fi
         first=false
         DB_STRING+="\"${CATALOGS[index]}\""
     done
     DB_STRING+="]"
 }

 function patch_cronjob(){
     oc set env -n $NAMESPACE cronjob/wkc-search-relationships-cronjob dbs_to_sync=$DB_STRING
 }

 function spawn_job(){
     oc delete job -n $NAMESPACE wkc-search-relationships-job
     oc create job -n $NAMESPACE --from=cronjob/wkc-search-relationships-cronjob wkc-search-relationships-job
 }

 function restore_cronjob(){
     oc set env -n $NAMESPACE cronjob/wkc-search-relationships-cronjob dbs_to_sync=[]
 }

 prompt_for_namespace
 prompt_for_list_of_catalogs
 construct_db_list_string
 patch_cronjob
 spawn_job
 restore_cronjob

Execute the code from step 1, using the following command:
```
chmod +x ./cpd_gs_sync.sh
```
Execute the job, using the following command:
```
./cpd_gs_sync.sh
```
Follow the prompts in the terminal to customize the job as required.

Note: If the list of containers is empty, all of the catalogs, projects, and workspaces will be resynched.

Run the following command to check the status of the job (wkc-search-relationships-job):

oc get jobs -n ${CPD_INSTANCE_PROJECT} -w | grep relationships

Once the job is complete, the status will display:

wkc-search-relationships-job                      1/1           3m52s      20m

When the job status is complete, run the following patch command to enable the use of global search relationships:

oc patch ccs ccs-cr -n ${CPD_INSTANCE_PROJECT} --type=merge --patch='{"spec":   {"portal_catalog_is_global_search_relationships_enabled": "true"}}'

To ensure that the portal-catalog pod has been restarted:
1. Get the pod name:
```
oc get pods -n ${CPD_INSTANCE_PROJECT} -w | grep portal-catalog
```
2. Delete the pod. Replace <pod-name> with the name of the pod from step a.
```
oc delete pod <pod-name> -n ${CPD_INSTANCE_PROJECT}
```

Parent topic: Administering IBM Knowledge Catalog