Couchbase Sync Gateway is a key component of the Couchbase Mobile stack. It is an Internet-facing synchronization mechanism that securely syncs data across devices as well as between devices and the cloud. Couchbase Mobile uses a websocket based replication protocol.
The core functions of the Sync Gateway include:
This tutorial will demonstrate how to -
We will be using a Android Java App as an example of a Couchbase Lite enabled client.
You can learn more about the Sync Gateway here in the Sync Gateway Documentation.
This tutorial assumes familiarity with building Android apps using Java and with the basics of Couchbase Lite.
If you are unfamiliar with the basics of Couchbase Lite, it is recommended that you walk through the following tutorials
Android SDK installed and setup (> v.28.0.0)
Android Build Tools (> v.28.0.0)
Android device or emulator running API level 21 or above
JDK 8 (now embedded into Android Studio 4+)
curl HTTP client
You could use any HTTP client of your choice. But we will use curl in our tutorial. Download latest version from curl website. MacOS Package manager users can use homebrew. Windows Package Manager users can use winget.
Docker
We will be working with a simple "User Profile" app which we introduced in the Quickstart in Couchbase Lite with Android and Java tutorial and extended in the Quickstart in Couchbase Lite Query with Android and Java tutorial.
In this tutorial, we will be extending that app to support data sync.
The app does the following
Allows users to log in and create or update his/her user profile information. The user profile view is automatically updated every time the profile information changes in the underlying database.
The user profile information is synced with a remote Sync Gateway which then syncs it to other devices (subject to access control and routing configurations specified in the sync function
).
To clone the project from GitHub, type the following command in your terminal:
git clone https://github.com/couchbase-examples/android-java-cblite-userprofile-sync
allprojects {
repositories {
...
maven {
url "https://mobile.maven.couchbase.com/maven2/dev/"
}
}
}
Then add the following to the app/build.gradle file.
dependencies {
...
implementation 'com.couchbase.lite:couchbase-lite-android-ee:3.0.0'
}
The sample app follows the MVP pattern, separating the internal data model, from a passive view through a presenter that handles the logic of our application and acts as the conduit between the model and the view.
In the Android Studio project, the code is structured by feature. You can select the Android option in the left navigator to view the files by package.
Each package contains 3 different files:
Activity: This is where all the view logic resides.
Presenter: This is where all the business logic resides to fetch and persist data to a web service or the embedded Couchbase Lite database.
Contract: An interface that the Presenter
and Activity
implement.
If you fhave followed along the tutorial on link:[Query Basics], you can skip this section and proceed to the Backend Installation section since we have not made any changes to the Data model for this tutorial.
Couchbase Lite is a JSON Document Store. A Document is a logical collection of named fields and values. The values are any valid JSON types. In addition to the standard JSON types, Couchbase Lite supports some special types like Date
and Blob
.
While it is not required or enforced, it is a recommended practice to include a "type" property that can serve as a namespace for related.
The app deals with a single Document with a "type" property of "user". The document ID is of the form "user::demo@example.com". An example of a document would be
{
"type":"user",
"name":"Jane Doe",
"email":"jame.doe@earth.org",
"address":"101 Main Street",
"image":CBLBlob (image/jpg),
"university":"Missouri State University"
}
For this tutorial the "user" Document
is first stored within an Object
of type Map<String, Object>
.
Map<String, Object> profile = new HashMap<>();
profile.put("name", nameInput.getText().toString());
profile.put("email", emailInput.getText().toString());
profile.put("address", addressInput.getText().toString());
profile.put("university", universityText.getText().toString());
profile.put("type", "user");
byte[] imageViewBytes = getImageViewBytes();
if (imageViewBytes != null) {
profile.put("imageData", new com.couchbase.lite.Blob("image/jpeg", imageViewBytes));
}
The app comes bundled with a collection of Documents of type "university". Each Document
represents a university.
{
"type":"university","web_pages": [
"http://www.missouristate.edu/"
],
"name": "Missouri State University",
"alpha_two_code": "US",
"state-province": MO,
"domains": [
"missouristate.edu"
],
"country": "United States"
}
When "university" Document
is retrieved from the database it is stored within an Object
of type Map<String, Object>
.
Map<String, Object> properties = new HashMap<>()
properties.put("name", row.getDictionary("universities").getString("name"));
properties.put("country", row.getDictionary("universities").getString("country"));
properties.put("web_pages", row.getDictionary("universities").getArray("web_pages"));]
We will install Couchbase Server and Sync Gateway using Docker.
Couchbase Server and Sync Gateway Server need to communicate with each other over the network. A network bridge in docker allows network traffic between servers. Create a docker network bridge named workshop.
docker network ls
docker network create -d bridge workshop
We have a custom docker image priyacouch/couchbase-server-userprofile:7.0.0-dev of Couchbase Server, which creates an empty bucket named userprofile and an RBAC user admin with sync gateway role.
Alternatively, you can follow the instructions in our documentation — see: Get Started - Prepare, to install Couchbase Server and configure it with the relevant bucket.
docker stop cb-server && docker rm cb-server
docker run -d --name cb-server \
--network workshop \
-p 8091-8094:8091-8094 -p 11210:11210 \
priyacouch/couchbase-server-userprofile:7.0.0-dev
The server could take a few minutes to deploy and fully initialize; so be patient.
docker logs -f cb-server
When the setup is completed, you should see output similar to that shown in below:
a. Open up http://localhost:8091 in a browser. b. Sign in as Administrator and password in login page. c. Go to buckets menu and confirm userprofile bucket is created
Now we will install, configure, and run Sync Gateway.
When using Sync Gateway, we can opt to provide a bootstrap configuration -- see: Sync Gateway Configuration. We would then provision database, sync and other configuration using the Admin REST endpoints Alternatively, we can continue to run in legacy-mode, using the Pre-3.0 configuration.
In this tutorial - for the purposes of backward compatibility - we will run 3.x using its
legacy configuration option. That is, we will be running with the disable_persistent_config
option in the configuration file set to true
. You can, if you wish, run a 2.8 version of Sync Gateway instead.
The configuration files corresponding to this sample application are shown in Table 1.
They are available in the github repo hosting the app, which you cloned - look in:
/path/to/cloned/repo/android-java-cblite-userprofile-sync/src/
Table 1. Available configuration files
Release | Filename |
---|---|
3.x | sync-gateway-config-userprofile-demo-3-x-legacy.json |
2.x | sync-gateway-config-userprofile-demo-2-x.json |
Let us configure and launch Sync Gateway in a Docker container.
cd /path/to/cloned/repo/android-java-cblite-userprofile-sync/src
docker stop sync-gateway && docker rm sync-gateway
Launch Sync Gateway in a Docker container using directions below based on the version you are using.
Configuring and running Sync Gateway 3.x in Docker using the configuration in sync-gateway-config-userprofile-demo-3-x-legacy.json
.
Note the use of disable_persistent_config
in the configuration file to force legacy configuration mode.
docker run -p 4984-4986:4984-4986 \
--network workshop \
--name sync-gateway \
-d \
-v `pwd`/sync-gateway-config-userprofile-demo-3-x-legacy.json \
/etc/sync_gateway/sync_gateway.json \
couchbase/sync-gateway:3.0.0-enterprise \
/etc/sync_gateway/sync_gateway.json
Configuring and running Sync Gateway 2.8.
docker run -p 4984-4986:4984-4986 \
--network workshop \
--name sync-gateway \
-d \
-v `pwd`/sync-gateway-config-userprofile-demo-2-x.json:\
/etc/sync_gateway/sync_gateway.json \
couchbase/sync-gateway:2.8.4-enterprise \
/etc/sync_gateway/sync_gateway.json
Configure and run Sync Gateway 3.0 in legacy mode.
docker run -p 4984-4986:4984-4986 ^
--network workshop ^
--name sync-gateway ^
-d -v %cd%sync-gateway-config-userprofile-demo-3-x-legacy.json:^
/etc/sync_gateway/sync_gateway.json ^
couchbase/sync-gateway:3.0.0-enterprise ^
/etc/sync_gateway/sync_gateway.json
Configuring and running Sync Gateway 2.8.
docker run -p 4984-4986:4984-4986 ^
--network workshop ^
--name sync-gateway ^\
-d ^
-v %cd%/sync-gateway-config-userprofile-demo-2-x.json:^
etc/sync_gateway/sync_gateway.json ^
couchbase/sync-gateway:2.8.4-enterprise ^
/etc/sync_gateway/sync_gateway.json
Now we can confirm that the Sync Gateway is up and running.
docker logs -f sync-gateway
You will see a series of log messages. Make sure there are no errors.
{"couchdb":"Welcome","vendor": { "name":"Couchbase Sync Gateway", "version":"3.0" },
"version":"Couchbase Sync Gateway/3.0.0(460;26daced) EE"}
Now that we have the server and the sync gateway installed, we can verify data sync between Couchbase Lite enabled apps.
A key component of the sync process is the Sync Function and we will first look at how that can be set-up to control how data sync works.
The Sync Function is a Javascript function that is specified as part of the Sync Gateway Configuration. It handles Authorization, Data Validation, Data Routing, and Access Countrol.
Open the your configuration file using a text editor of your choice. It will be located in the repo at /path/to/cloned/repo/android-java-cblite-userprofile-sync/src
.
Locate the sync
setting in the file.
Now you can follow along with the rest of the sections below.
We use Basic Authentication in our application. The Id of the user making the request is specified in the Authorization
header.
Locate the /*Authorization*/
section of the Sync Function. You will see we are using the Sync functions requireUser()
API to verify that the email
property specified in the Document matches the Id of the user making the request.
function sync(doc, oldDoc) {
....
/* Authorization */
// Verify the user making the request is the same as the one in doc's email
requireUser(doc.email);
.....
}
In this case, we are doing some basic validation of the contents of the Document:
/* Data Validation */
// Validate the presence of email field.
// This is the "username"
validateNotEmpty("email", doc.email);
// Validate that the document Id _id is prefixed by owner
var expectedDocId = "user" + "::" + doc.email;
if (expectedDocId != doc._id) {
// reject document
throw({forbidden: "user doc Id must be of form user::email"});
}
email
property is not null. If it's null, we throw a JS exception (see validateNotEmpty()
function)Id
of the Document is of the required format (i.e. "user::<email>"). We throw an exception if that's not the case.email
property value has not changed. Again, we throw an exception if that's not the case.NOTE: You can learn more about the Sync Function in the documentation here: Sync Function API.
channels
are a mechanism to "tag" documents. They are typically used to route/seggregate documents based on the contents of those document.
Combined with access() and requireAccess()
API, the channel() API can be used to enforce Access Control.
As we shall see in a later section, clients can use channels to pull only a subset of documents.
/* Routing */
// Add doc to the user's channel.
var email = getEmail();
var channelId = "channel."+ username;
channel(channelId);
We can enforce access control to channels using the access () API. This approach ensures that only users with access to a specific channel will be able to retrieve documents in the channel.
// Give user read access to channel
access(username, channelId);
Two-way Replication between the app and the Sync Gateway is enabled when the user logs into the app.
startPushAndPullReplicationForCurrentUser()
function.public static void startPushAndPullReplicationForCurrentUser(String username, String password)
ReplicatorConfiguration
instance that specifies the source and target database and you can optionally, override the default configuration settings.ReplicatorConfiguration config = new ReplicatorConfiguration(userprofileDatabase, new URLEndpoint(url)); // <1>
config.setReplicatorType(ReplicatorConfiguration.ReplicatorType.PUSH_AND_PULL); // <2>
config.setContinuous(true); // <3>
config.setAuthenticator(new BasicAuthenticator(username, password)); // <4>
config.setChannels(Arrays.asList("channel." + username)); // <5>
source
as the local Couchbase Lite database and the remote
target as the Sync Gatewaytype
of PUSH_AND_PULL
indicates that we require two-way sync. A value of .PUSH
specifies that we only pull data from the Sync Gateway. A value of .PULL
specifies that we only push data.continuous
mode is specified to be true which means that changes are synced in real-time. A value of false implies that data is only pulled from the Sync Gateway.requireUser
API.channels
are used to specify the channels to pull from. Only documents belonging to the specified channels are synced. This is subject to Access Control rights enforced at the Sync Gateway. This means that if a client does not have access to documents in a channel, the documents will not be synched even if the client specifies it in the replicator configuration.Replicator
with the ReplicatorConfiguration
replicator = new Replicator(config)
In order to follow the replicator's progress, we can attach a callback listener to it.
Attaching a callback listener to the Replicator
means we will be asynchronously notified of state changes. This could be useful for instance, to inform the user of the progress of the replication. It is an optional step shown below.
replicatorListenerToken = replicator.addChangeListener(new ReplicatorChangeListener() {
@Override
public void changed(ReplicatorChange change) {
if (change.getReplicator().getStatus().getActivityLevel().equals(Replicator.ActivityLevel.IDLE)) {
Log.e("Replication Comp Log", "Scheduler Completed");
}
if (change.getReplicator().getStatus().getActivityLevel().equals(Replicator.ActivityLevel.STOPPED)
|| change.getReplicator().getStatus().getActivityLevel().equals(Replicator.ActivityLevel.OFFLINE)) {
Log.e("Rep Scheduler Log", "ReplicationTag Stopped");
}
}
});
Now, with all that done, we can start the replicator.
replicator.start()
When the user logs out of the app, the replication is stopped before the database is closed.
stopAllReplicationForCurrentUser()
function.public static void stopAllReplicationForCurrentUser()
replicator.removeChangeListener(replicatorListenerToken)
replicator.stop();
TIP: When you close a database, any active replicators, listeners and-or live queries are also be closed.
Couchbase Lite applications can set up live queries in order to be asynchronously notified of changes to the database that affect the results of the query. This can be very useful, for instance, in keeping a UI View up-to-date with the results of a query.
In our app, the user profile view is kept up-to-date using a live query that fetches the user profile data used to populate the view. This means that, if the replicator pulls down changes to the user profile, they are automatically reflected in the view.
fetchProfile()
function.public void fetchProfile()
QueryBuilder
API. If you are unfamiliar with this API, please see: Quickstart in Couchbase Lite Query with Android and Java tutorial.Query query = QueryBuilder
.select(SelectResult.all())
.from(DataSource.database(database))
.where(Meta.id.equalTo(Expression.string(docId)));
Note: We query for documents based on document Id. In our app, there should be exactly one user profile document corresponding to this Id.
query.addChangeListener(new QueryChangeListener() {
@Override
public void changed(QueryChange change) { // <1>
ResultSet rows = change.getResults();
Result row = null;
Map<String, Object> profile = new HashMap<>(); // <2>
profile.put("email", DatabaseManager.getSharedInstance().currentUser);
while ((row = rows.next()) != null) {
Dictionary dictionary = row.getDictionary("userprofile"); // <3>
if (dictionary != null) {
profile.put("name", dictionary.getString("name")); // <4>
profile.put("address", dictionary.getString("address")); // <4>
profile.put("imageData", dictionary.getBlob("imageData")); // <4>
profile.put("university", dictionary.getString("university")); // <4>
profile.put("type", dictionary.getString("type")); // <4>
}
}
mUserProfileView.showProfile(profile);
}
});
SelectResult.all()
method is used to query all the properties of a document. In this case, the document in the result is embedded in a dictionary where the key is the database name, which is "userprofile". So we retrieve the Dictionary
at key "userprofile".In this exercise, we will observe how changes made on one app are synced across to the other app
In this exercise, we will observe changes made via Sync Gateway are synced over to the apps
Make sure you have completed Exercise 1. This is to ensure that you have the appropriate user profile document (with document Id of "user::demo@example.com") created through the app and synced over to the Sync Gateway.
Open the command terminal and issue the following command to get the user profile document via GET Document REST API. We will be using curl
to issue the request. If you haven't done so, please install curl as indicated in the Prerequisites section.
curl -X GET http://localhost:4984/userprofile/user::demo@example.com --user demo@example.com
NOTE: This GET retrieves the userprofile document with the id user::demo@example.com
{
"_attachments": { <2>
"blob_1": {
"content_type": "image/jpeg",
"digest": "sha1-S8asPSgzA+F+fp8/2DdIy4K+0U8=",
"length": 14989,
"revpos": 2,
"stub": true
}
},
"_id": "user::demo@example.com",
"_rev": "2-3a76cfa911e2c54d1e82b29dbffc7f4e5a9bc265", //<1>
"address": "",
"email": "demo@example.com",
"image": {
"@type": "blob",
"content_type": "image/jpeg",
"digest": "sha1-S8asPSgzA+F+fp8/2DdIy4K+0U8=",
"length": 14989
},
"name": "",
"type": "user",
"university": "Missouri State University"
}
If you had updated an image via the mobile app, you should see an "_attachments" property. This entry holds an array of attachments corresponding to each image blob entry added by the mobile app. This property is added by the Sync Gateway when it processes the document. You can learn more about how image Blob types are mapped to attachments here.
Record the revision Id of the document. You will need this when you update the document
NOTE: We chose to show how to update the address field via the REST API. You can choose to update any other profile information if you like. You will be prompted to enter the users password when you submit the curl command.
curl -X PUT \
'http://localhost:4985/userprofile/user::demo@example.com?rev=3-12d203d6024c8b844c5ed736c726ac63379e05dc' \
-H 'Accept: application/json' \
-H 'Cache-Control: no-cache' \
-H 'Content-Type: application/json' \
-d '{
"address": "101 Main Street", //<1>
"email": "demo@example.com",
"image": {
"@type": "blob",
"content_type": "image/jpeg",
"digest": "sha1-S8asPSgzA+F+fp8/2DdIy4K+0U8=",
"length": 14989
},
"name": "",
"type": "user",
"university": "Missouri State University"
}'
Here, in the
PUT
, we specify the:
- user id (
user::demo@example.com
)- revision Id (from the previous step
3-033fcbaf269d65a9247067be76d664f1111d033b
) to select the item we want to update
Confirm that you get a HTTP "201 Created" status code
As soon as you update the document via the Sync Gateway REST API, confirm that the changes show up in the mobile app on the simulator.
Data conflicts are inevitable in an environment where you can potentially have multiple writes updating the same data concurrently. Couchbase Mobile supports Automated Conflict Resolution.
You can learn more about automated conflict resolution in this blog Document Conflicts & Resolution.
Congratulations on completing this tutorial!
This tutorial walked you through an example of how to use a Sync Gateway to synchronize data between Couchbase Lite enabled clients. We discussed how to configure your Sync Gateway to enforce relevant access control, authorization and data routing between Couchbase Lite enabled clients.
Check out the following links for further details