The first application that we are going to migrate is a simple Product Inventory web-based application front-end, backed backed by Microsoft SQL Server. This application will deploy into the mssql-example namespace using a single PV backed by NFS for persistent storage.
First we'll need to deploy the mssql app to our source cluster using yaml from the mig-demo-apps repo.
$ oc create -f https://raw.githubusercontent.com/konveyor/mig-demo-apps/master/apps/mssql-app/manifest.yamlRequired Container Images - MSSQL Example
quay.io/ocpmigrate/mssql-sample-app:microsoft
quay.io/ocpmigrate/mssql-server:latest
After creation, we should see mssql-example come up.
$ oc get pods -n mssql-example
NAME READY STATUS RESTARTS AGE
mssql-app-deployment-6ffb46c5d6-n5fvv 1/1 Running 0 41m
mssql-deployment-1-xq4p4 1/1 Running 0 41mLet's get the route to the application, and bring up the webUI.
$ oc get route -n mssql-example
NAME HOST/PORT PATH SERVICES PORT
mssql-app-route mssql-app-route-mssql-example.apps.srccluster.com db-app-svc 5000
Let's go ahead and add a new product to the inventory. Click on the +Add button and enter some data.
You can see the application is functioning and state is being saved in the DB.
Let's also verify that the application is NOT installed on our destination cluster. First login to the destination cluster. You can see that no Pods are running; and in fact the mssql-example namespace does not exist.
The custom SCC yaml is available here. Download to your local machine, as we will apply it in the next step.
- Run the following to recreate the MSSQL
sccon the destination cluster:
$ oc create -f mssql-scc.yaml
securitycontextconstraints.security.openshift.io/mssql createdNext, let's open up the migration UI. Again, to get the route, run the following command on the destination cluster:
$ oc get routes migration -n openshift-migration -o jsonpath='{.spec.host}'
migration-mig.apps.cluster-a21d.a21d.sandbox67.opentlc.comThe screen should look something like:
First thing we want to do is add the source OCP cluster we wish to migrate the
application from. Click Add cluster:
Fill out the neccessary information. We will need an Service Account Token in order for destination cluster to talk to source cluster:
Run the following against your source cluster.
$ oc sa get-token -n openshift-migration migration-controller
eyJhbGciOifsfsds8ahmtpZCI6IiJ9fdsfdseyJpc3MiOiJrdWJlcm5ldGVzL3NlcnZpY2VhY2NvdW50Iiwia3ViZXJuZXRlcy5pby9zZXJ2aWNlYWNjb3VudC9uYW1lc3BhY2UiOiJtaWciLCJrdWJlcm5ldGVzLmlvL3NlcnZpY2VhY2NvdW50L3NlY3JldC5uYW1lIjoibWlnLXRva2VuLTdxMnhjIiwia3ViZXJuZXRlcy5pby9zZXJ2aWNlYWNjb3VudC9zZXJ2aWNlLWFjY291bnQubmFtZSI6Im1pZyIsImt1YmVybmss7gc2VydmljZWFjY291bnQvc2VydmljZS1hY2NvdW50LnVpZCI6IjQ5NjYyZjgxLWEzNDItMTFlOS05NGRjLTA2MDlkNjY4OTQyMCIsInN1YiI6InN5c3RlbTpzZXJ2aWNlYWNjb3VudDptaWc6bWlnIn0.Qhcv0cwP539nSxbhIHFNHen0PNXSfLgBiDMFqt6BvHZBLET_UK0FgwyDxnRYRnDAHdxAGHN3dHxVtwhu-idHKI-mKc7KnyNXDfWe5O0c1xWv63BbEvyXnTNvpJuW1ChUGCY04DBb6iuSVcUMi04Jy_sVez00FCQ56xMSFzy5nLW5QpLFiFOTj2k_4Krcjhs8dgf02dgfkkshshjfgfsdfdsfdsa8fdsgdsfd8fasfdaTScsu4lEDSbMY25rbpr-XqhGcGKwnU58qlmtJcBNT3uffKuxAdgbqa-4zt9cLFeyayTKmelc1MLswlOvu3vvJ2soFx9VzWdPbGRMsjZWWLvJ246oyzwykYlBunYJbX3D_uPfyqoKfzA
# We need to save the output of the 'get-token', that is the long string we will enter into the mig-ui when we create a new cluster entry.When done, click Add Cluster. You should see a Connection successful message. Click Close.
Now you should see the source and destination clusters populated.
Next we want to add a replication repository, which is an Object Storage endpoint. Currently we assume this is a S3 like endpoint.
If you do not have a Object Storage endpoint ready to use, refer to ObjectStorage.md
Click Add Repository:
Fill out the information for the S3 bucket you wish to use as S3 scratch space. Click Add repository and you should see a Connection successful message. Click Close.
You should now see the repository myrepo populated.
Now that we have a replication repository specified and both the source and
destination clusters defined, we can create a migration plan. Click Add Plan:
Fill out a plan name. Click Next.
Select the source and target cluster, the replication repository, and the mssql-example namespace (which we want to migrate over). Click Next.
Now we are displayed a list of persistent volumes associated with our
application workload. Select which type of action you would like to perform on the PV. For this example we've selected copy. Click Next.
Since we are copying to a new PV, we must select the storage class to be used for creation of new PVs. In this case we will be copying our data from NFS to AWS-EBS (gp2:kubernetes.io/aws-ebs). Click Next.
After validating the migration plan, you will see a Ready message and you can click Close.
Now we can select Migrate or Stage on the application. Since we don't care about downtime for this example, let's select Migrate:
Optionally choose to not terminate the application on the source cluster.
Leave it unchecked and select Migrate.
The migration will progress with a progress bar showing each step in the process.
Once done, you should see Migration Succeeded on the migration plan.
Let's check out Pods running in the mssql-example namespace on the destination cluster.
First we can check out the Pods running on the destination and verify they meet our expectation.
$ oc get pods -n mssql-example
NAME READY STATUS RESTARTS AGE
mssql-app-deployment-6ffb46c5d6-n5fvv 1/1 Running 0 5m
mssql-deployment-1-xq4p4 1/1 Running 0 6mNext, let's check for a newly created route on the destination cluster.
$ oc get route -n mssql-example
NAME HOST/PORT PATH SERVICES
mssql-app-route mssql-app-route-mssql-example.apps.destcluster.com db-app-svc 5000Open the route in a browser and verify the mssql demo app is functional.
Next Section: Section 4 - Migrate Sock Shop Application
Previous Section: Section 2 - CAM Overview
Home