Skip to main content

How to deploy a mirror Data Availability Server (DAS)


This document is currently in public preview and may change significantly as feedback is captured from readers like you. Click the Request an update button at the top of this document or join the Arbitrum Discord to share your feedback.

Running a regular DAS vs running a mirror DAS

The main use-case for running a mirror DAS is to complement your setup as a Data Availability Committee (DAC) member. That means that you should run your main DAS first, and then configure the mirror DAS. Refer to How to deploy a DAS if needed.

AnyTrust chains rely on an external Data Availability Committee (DAC) to store data and provide it on-demand instead of using its parent chain as the Data Availability (DA) layer. The members of the DAC run a Data Availability Server (DAS) to handle these operations.

In this how-to, you'll learn how to configure a mirror DAS that serves GET requests for stored batches of information through a REST HTTP interface. For a refresher on DACs, refer to the Introduction.

This how-to assumes that you're familiar with:

  • How a regular DAS works and what configuration options are available. Refer to How to deploy a DAS for a refresher.
  • Kubernetes. The examples in this guide use Kubernetes to containerize your DAS.

What is a mirror DAS?

To avoid exposing the REST interface of your main DAS to the public in order to prevent spamming attacks (as explained in How to deploy a DAS), you can choose to run a mirror DAS to complement your setup. The mirror DAS will handle all public REST requests, while reading information from the main DAS via its (now private) REST interface.

In general, mirror DA servers serve two main purposes:

  1. Prevent the main DAS from having to serve requests for data, allowing it to focus only on storing the data received.
  2. Provide resiliency to the network in the case of a DAS going down.

Configuration options

A mirror DAS will use the same tool and, thus, the same configuration options as your main DAS. You can find an explanation of those options in How to deploy a DAS.

How to deploy a mirror DAS

Step 0: Prerequisites

Gather the following information:

  • The latest Nitro docker image: offchainlabs/nitro-node:v2.3.4-b4cc111
  • An RPC endpoint for the parent chain. It is recommended to use a third-party provider RPC or run your own node to prevent being rate limited.
  • The SequencerInbox contract address in the parent chain.
  • URL of the list of REST endpoints of other DA servers to configure the REST aggregator.

Step 1: Set up a persistent volume

First, we'll set up a volume to store the DAS database. In k8s, we can use a configuration like this:

apiVersion: v1
kind: PersistentVolumeClaim
name: das-mirror
- ReadWriteOnce
storage: 200Gi
storageClassName: gp2

Step 2: Deploy the mirror DAS

To run the mirror DAS, we'll use the daserver tool and we'll configure the following parameters:

--data-availability.parent-chain-node-urlRPC endpoint of a parent chain node
--data-availability.sequencer-inbox-addressAddress of the SequencerInbox in the parent chain
--enable-restEnables the REST server listening on --rest-addr and --rest-port
--rest-addrREST server listening interface (default "localhost")
--rest-port(Optional) REST server listening port (default 9877)
--log-levelLog level: 1 - ERROR, 2 - WARN, 3 - INFO, 4 - DEBUG, 5 - TRACE (default 3) retrieval of sequencer batch data from a list of remote REST endpoints URL to a list of URLs of REST DAS endpoints that is checked at startup. This option is additive with the urls option of URLs including 'http://' or 'https://' prefixes and port numbers to REST DAS endpoints. This option is additive with the online-url-list option using a REST aggregator, checks if the data already exists in this DAS's storage. Must be disabled for fast sync with an IPFS backend (default true) using a REST aggregator, eagerly syncs batch data to this DAS's storage from the REST endpoints, using the parent chain as the index of batch data hashes; otherwise only syncs lazily using a REST aggregator that's eagerly syncing, starts indexing forward from this block from the parent chain. Only used if there is no sync state. using a REST aggregator, period to retain the synced data (defaults to forever) using a REST aggregator, directory to store the sync state in, i.e. the block number currently synced up to, so that it doesn't sync from scratch each time

To enable caching, you can use the following parameters:

--data-availability.local-cache.enableEnables local in-memory caching of sequencer batch data
--data-availability.local-cache.capacityMaximum number of entries (up to 64KB each) to store in the cache. (default 20000)

Finally, for the storage backends you wish to configure, use the following parameters. Toggle between the different options to see all available parameters.

--data-availability.s3-storage.enableEnables storage/retrieval of sequencer batch data from an AWS S3 bucket
--data-availability.s3-storage.access-keyS3 access key
--data-availability.s3-storage.bucketS3 bucket
--data-availability.s3-storage.regionS3 region
--data-availability.s3-storage.secret-keyS3 secret key
--data-availability.s3-storage.object-prefixPrefix to add to S3 objects
--data-availability.s3-storage.discard-after-timeoutWhether to discard data after its expiry timeout (setting it to false, activates the “archive” mode)

Here's an example daserver command for a mirror DAS that:

  • Enables local cache
  • Enables AWS S3 bucket storage that doesn't discard data after expiring (archive)
  • Enables local Badger database storage that doesn't discard data after expiring (archive)
  • Uses a local main DAS as part of the REST aggregator
--data-availability.parent-chain-node-url "<YOUR PARENT CHAIN RPC ENDPOINT>"
--data-availability.sequencer-inbox-address "<ADDRESS OF SEQUENCERINBOX ON PARENT CHAIN>"
--rest-addr ''
--log-level 3
--data-availability.local-cache.enable "http://your-main-das.svc.cluster.local:9877" "<URL TO LIST OF REST ENDPOINTS>" "BLOCK NUMBER" /home/user/data/syncState
--data-availability.s3-storage.access-key "<YOUR ACCESS KEY>"
--data-availability.s3-storage.bucket "<YOUR BUCKET>"
--data-availability.s3-storage.region "<YOUR REGION>"
--data-availability.s3-storage.secret-key "<YOUR SECRET KEY>"
--data-availability.s3-storage.object-prefix "<YOUR OBJECT KEY PREFIX>/"
--data-availability.s3-storage.discard-after-timeout false
--data-availability.local-db-storage.enable /home/user/data/badgerdb
--data-availability.local-db-storage.discard-after-timeout false

And here's an example of how to use a k8s deployment to run that command:

apiVersion: apps/v1
kind: Deployment
name: das-mirror
replicas: 1
app: das-mirror
maxSurge: 0
maxUnavailable: 50%
type: RollingUpdate
app: das-mirror
- command:
- bash
- -c
- |
mkdir -p /home/user/data/badgerdb
mkdir -p /home/user/data/syncState
/usr/local/bin/daserver --data-availability.parent-chain-node-url "<YOUR PARENT CHAIN RPC ENDPOINT>" --data-availability.sequencer-inbox-address "<ADDRESS OF SEQUENCERINBOX ON PARENT CHAIN>" --enable-rest --rest-addr '' --log-level 3 --data-availability.local-cache.enable "http://your-main-das.svc.cluster.local:9877" "<URL TO LIST OF REST ENDPOINTS>" "BLOCK NUMBER" /home/user/data/syncState --data-availability.s3-storage.enable --data-availability.s3-storage.access-key "<YOUR ACCESS KEY>" --data-availability.s3-storage.bucket "<YOUR BUCKET>" --data-availability.s3-storage.region "<YOUR REGION>" --data-availability.s3-storage.secret-key "<YOUR SECRET KEY>" --data-availability.s3-storage.object-prefix "<YOUR OBJECT KEY PREFIX>/" --data-availability.local-db-storage.enable /home/user/data/badgerdb
image: offchainlabs/nitro-node:v2.3.4-b4cc111
imagePullPolicy: Always
cpu: "4"
memory: 10Gi
cpu: "4"
memory: 10Gi
- containerPort: 9877
hostPort: 9877
protocol: TCP
- mountPath: /home/user/data/
name: data
failureThreshold: 3
path: /health/
port: 9877
scheme: HTTP
initialDelaySeconds: 5
periodSeconds: 5
successThreshold: 1
timeoutSeconds: 1
- name: data
claimName: das-mirror

Archive DA servers

Archive DA servers are servers that don't discard any data after expiring. Each DAC should have at the very least one archive DAS to ensure all historical data is available.

To activate the "archive mode" in your DAS, set the parameter discard-after-timeout to false in your storage method. For example:


Note that local-file-storage and ipfs-storage don't discard data after expiring, so the option discard-after-timeout is not available.

Archive servers should make use of the options described above to pull in any data that they don't have.

Helm charts

A helm chart is available at ArtifactHUB. It supports running a mirror DAS by providing the parameters for your server. Find more information in the OCL community Helm charts repository.

Testing the DAS

Once the DAS is running, we can test if everything is working correctly using the following methods.

Test 1: REST health check

The REST interface enabled in the mirror DAS has a health check on the path /health which will return 200 if the underlying storage is working, otherwise 503.


curl -I <YOUR REST ENDPOINT>/health

Security considerations

Keep in mind the following information when running the mirror DAS.

For a mirror DAS, using a load balancer is recommended to manage incoming traffic effectively. Additionally, as the REST interface is cacheable, consider deploying a Content Delivery Network (CDN) or caching proxy in front of your REST endpoint. The URL for the REST interface will be publicly known; ensure that it is sufficiently distinct from the RPC endpoint to prevent the latter from being easily discovered.

What to do next?

Once the DAS is deployed and tested, you'll have to communicate the following information to the chain owner, so they can update the chain parameters and configure the sequencer:

  • The https URL for the REST endpoint (e.g.

Optional parameters

Besides the parameters described in this guide, there are some more options that can be useful when running the DAS. For a comprehensive list of configuration parameters, you can run daserver --help.

--conf.dumpPrints out the current configuration
--conf.fileAbsolute path to the configuration file inside the volume to use instead of specifying all parameters in the command


The DAS comes with the option of producing Prometheus metrics. This option can be activated by using the following parameters:

--metricsEnables the metrics server
--metrics-server.addrMetrics server address (default "")
--metrics-server.portMetrics server port (default 6070)
--metrics-server.update-intervalMetrics server update interval (default 3s)

When metrics are enabled, several useful metrics are available at the configured port, at path debug/metrics or debug/metrics/prometheus.

RPC metrics

arb_das_rpc_store_requestsCount of RPC Store calls
arb_das_rpc_store_successSuccessful RPC Store calls
arb_das_rpc_store_failureFailed RPC Store calls
arb_das_rpc_store_bytesBytes retrieved with RPC Store calls
arb_das_rpc_store_duration (p50, p75, p95, p99, p999, p9999)Duration of RPC Store calls (ns)

REST metrics

arb_das_rest_getbyhash_requestsCount of REST GetByHash calls
arb_das_rest_getbyhash_successSuccessful REST GetByHash calls
arb_das_rest_getbyhash_failureFailed REST GetByHash calls
arb_das_rest_getbyhash_bytesBytes retrieved with REST GetByHash calls
arb_das_rest_getbyhash_duration (p50, p75, p95, p99, p999, p9999)Duration of REST GetByHash calls (ns)