References for Data Orchestration for Dell
Introduction
This page is a quick reference for Data Orchestration (Golden Copy): terminology, archive job types, commonly used searchctl CLI commands, folder definition parameters, and performance tuning settings. All commands below are run as the ecaadmin user from Golden Copy node 1 unless otherwise noted.
Terminology Glossary
| Term | Definition |
|---|---|
| Golden Copy | Superna's product name for the file-to-object copy and sync engine; branded as Data Orchestration on the Cyberstorage for Dell portal. |
| Virtual Accelerator Node (VAN) | An additional Golden Copy VM deployed to increase copy throughput. The first VM is the control VM; VAN nodes are used only for copying. Clusters can scale from 1 up to 99 nodes. |
| Folder Definition (archived folder) | A configured mapping of a PowerScale source path to an S3 target bucket/container, including credentials, cloud type, and schedules. Managed with searchctl archivedfolders. |
| Pipeline | The Golden Copy license feature enabling bi-directional data movement (file-to-object and object-to-file) and hybrid cloud workflows. |
| Recall | The process of restoring data from S3 object storage back to a PowerScale file system, with metadata (owner, group, ACLs, timestamps) re-applied where possible. |
| Full Archive | A job type that walks a folder path, copies new or modified files, and skips files that already exist unchanged on the target. Does not delete target files missing from the source. |
| Incremental Archive | A scheduled job type that uses the PowerScale change list API and snapshots to detect created, modified, and deleted files since the last run, and syncs those changes to the target. |
| Archive Data Audit | A job type that compares the full file system path to the target S3 bucket and reconciles differences, including removing target objects for files deleted from the source. |
| Delayed Deletes | A recycle mode where files deleted from the source during incremental sync are moved to a separate "trash" bucket instead of being deleted from the target, with an optional TTL for permanent purge. |
| ecactl | The CLI used for Golden Copy cluster-level operations (start, stop, reconfigure) — restricted to the ecaadmin appliance administrator. |
| searchctl | The CLI used for day-to-day Golden Copy configuration and operations (folders, jobs, licenses, users). |
Archive Job Types
| Job Type | Flag | Behavior |
|---|---|---|
| Full Archive | --full-archive-schedule | Walks the full folder path; copies new/modified files; skips unchanged files. Can be run repeatedly to catch missed data. |
| Incremental Archive | --incremental-schedule | Uses the PowerScale change list API and snapshots to sync created/modified files on a schedule. Does not delete target files by default (configurable). |
| Archive Data Audit | --archive-data-audit-schedule | Compares source and target, adds missing files, updates modified files, and removes target objects for files deleted from the source. Recommended to run weekly or monthly; can take hours to complete. |
Both --full-archive-schedule and --incremental-schedule accept a cron string (times are evaluated in UTC).
Commonly Used CLI Commands
Cluster and Licensing
| Command | Purpose |
|---|---|
searchctl isilons add --host <ip> --user eyeglassSR --applications {GC,SR} | Adds a PowerScale cluster to Golden Copy inventory. |
searchctl isilons license --name <cluster> --applications GC|GCA|GCP | Assigns a license (Golden Copy, Advanced, or Pipeline) to a cluster. Once assigned, a license is locked to that cluster. |
searchctl licenses add --path <path to zip> | Installs a license file. |
searchctl licenses list | Lists installed licenses. |
ecactl cluster down / ecactl cluster up | Shuts down / starts all Golden Copy cluster services. |
ecactl cluster rediscover | Wipes configuration while retaining clusters and folder definitions (1.1.9+). |
searchctl settings config export / importconfig | Exports/imports cluster configuration (clusters, folder definitions, encrypted credentials) — useful before a factory reset (1.1.9+). |
Folder Definitions
| Command | Purpose |
|---|---|
searchctl archivedfolders add ... | Adds a folder definition (source path + target bucket/container + credentials + cloud type). |
searchctl archivedfolders list / list --verbose | Lists configured folders and their IDs. |
searchctl archivedfolders modify --id <id> ... | Changes an existing folder's configuration. |
searchctl archivedfolders remove --id <id> | Removes a folder definition. |
searchctl archivedfolders test --id <id> | Validates S3 connectivity, upload, verification, and delete permissions before starting a job. |
Running and Monitoring Jobs
| Command | Purpose |
|---|---|
searchctl archivedfolders archive --id <id> | Starts an archive job on a folder (add --incremental for an on-demand incremental run, --follow to monitor immediately). |
searchctl jobs running | Lists currently running jobs. |
searchctl jobs view --id <job-id> --follow | Monitors a job's progress (MB/files queued, archived, % complete, error rate). |
searchctl jobs history [--folderid <id>] [--type <type>] [--tail <n>] | Lists job history, filterable by folder or job type (Incremental, Full Archive, Inventory). |
searchctl jobs summary --folders|--nodes|--jobs | Shows live summary statistics across folders, nodes, or jobs. |
searchctl jobs activecopies [--no-stream] | Shows files currently being copied on each node in real time. |
searchctl jobs cancel --id <job-id> | Cancels a running job. |
searchctl archivedfolders stats --id <id> or searchctl stats --folder <id> | Shows real-time throughput and error-rate stats for a folder. |
searchctl archivedfolders errors --jobId <id> --count <n> --tail | Returns the most recent errors for a job. |
searchctl job rerun --id <job-id> | Re-runs only the failed files from a completed job. |
searchctl archivedfolders rerun --id <id> --only-folders | Re-attempts a failed tree walk (folder listing) from a completed job (1.1.9+). |
searchctl archivedfolders modify --id <id> --paused true|false | Pauses/resumes all jobs on a folder. |
searchctl archivedfolders viewPaused | Lists all currently paused folders. |
Recall (Restore)
| Command | Purpose |
|---|---|
searchctl archivedfolders recall --id <id> [--subdir <path>] [--apply-metadata] | Recalls data from S3 back to the PowerScale recall staging path (/ifs/goldencopy/recall). --apply-metadata restores owner/group/mode bits and folder ACLs. |
--target-cluster, --start-time, --end-time, --timestamps-type | Advanced-license recall options: redirect recall to a different cluster, or select object versions by a created/modified date range. |
Folder Definition Parameters (archivedfolders add / modify)
| Parameter | Description |
|---|---|
--isilon | Name of the source PowerScale cluster (as shown in searchctl isilons list). |
--folder | Source path on the PowerScale cluster to copy or sync. |
--cloudtype {aws,ecs,azure,other,gcs,blackhole} | Target storage type. other uses AWS v4 signatures; otherv2 uses AWS v2 signatures. |
--bucket / --container | Target bucket name (all types except Azure, which uses --container). |
--endpoint | Target endpoint URL (required for ECS, AWS, Azure). |
--region | Required for Amazon S3. |
--accesskey / --secretkey | Target credentials (for Azure, --accesskey is the storage account name). |
--endpoint-ips | List or range of IPs for ECS/other targets that support multi-node load balancing. |
--tier | Storage tier for the target (Advanced/Backup Bundle license required); defaults to Standard (AWS) or Cool (Azure). |
--skip-s3-file-exists {true,false} | When true, skips checking whether a file already exists on the target before copying (avoids LIST/GET API costs on large full copies); defaults to false. |
--recyclebucket | Target bucket used to store deleted-file objects when Delayed Deletes is configured. |
--includes / --excludes | Glob patterns to include or restrict which files/folders are archived (1.1.4+). |
--incremental-schedule / --full-archive-schedule / --archive-data-audit-schedule | Cron strings to schedule the respective job types. |
--storeFileHistory | Retains a history of a file's movement/update lifecycle as a custom object property. |
--paused | Pauses all jobs on the folder. |
--customMetaData | Requires the Pipeline license; extracts and indexes file metadata from supported file types (not supported with Archive Engine). |
--cluster-name | Creates an alias for the root folder name used in the target storage path. |
--meta-prefix | Overrides the default x-amz-meta- prefix used for object metadata headers (some S3-compatible targets require a different prefix). |
Performance Tuning Parameters
Set in /opt/superna/eca/eca-env-common.conf on Golden Copy node 1; changes require ecactl cluster down / ecactl cluster up to take effect. Thread-count and memory changes should only be made with Support guidance, since increasing threads without addressing the actual bottleneck can reduce throughput.
| Variable | Purpose |
|---|---|
ARCHIVE_PARALLEL_THREAD_SDK | Number of threads used for multipart upload/download (default 10). |
ARCHIVE_PARALLEL_THREAD_COUNT | Number of parallel file-copy threads per VM (Golden Copy defaults to 100 threads/VM). |
AWS_HTTP_CONN_POOL_MAX_SIZE | Maximum HTTP connection pool size for AWS/generic S3 targets. |
RECALL_PARALLEL_THREAD_SDK | Parallel thread count for recall (restore) operations (requires 1.1.17+, used with AWS S3). |
INDEX_WORKER_RECALL_SKIP_EXISTENT_FILES | Skips files during recall tree-walking that already exist on disk. |
UNDERTOW_WORKER_THREADS | Only set on clusters of 12+ nodes; recommended value is node count × 4 × 2. |
ECA_GOLDENCOPY_ARCHIVEWORKER_PARTITIONS / ECA_INDEXWORKER_PARTITIONS | Partition counts used when building clusters larger than 6 nodes. |
ARCHIVE_SMART_INCREMENTAL | Set to true to enable fast incremental mode (skips an extra metadata API call) when daily incremental change volume exceeds roughly 3M files; owner/group/mode-bit metadata is not collected in this mode. |
ARCHIVE_FULL_TOPIC_REGEX / ARCHIVE_INCREMENTAL_TOPIC_REGEX | Used to change the default QOS prioritization between full and incremental jobs. |
ECA_KAFKA_LAG_CHECK_MINUTES | Large-file monitor timer; increase (for example, to 480 minutes/8 hours) for file systems dominated by large files. |
For full scale-out and scale-up guidance, including Virtual Accelerator Node deployment and memory/CPU sizing for large-file environments, see the Design Guide.
Supported Storage Target Feature Matrix
Golden Copy's core S3 protocol support is uniform, but feature-level support (versioning, checksum validation, multipart upload/download, and advanced/Cloud Browser-visible metadata) varies by target. Review the notes for your chosen target before relying on a specific feature, and consult the target vendor's own documentation for rate limiting, billing, and quota behavior. --cloudtype other uses AWS v4 signatures; some targets also require setting export ARCHIVE_S3_PREFIX="igls-" in /opt/superna/eca/eca-env-common.conf, noted below where required.
| Target | Cloud Type | Versioning | Checksum | Multipart Up/Down | Notes |
|---|---|---|---|---|---|
| Dell ECS | ecs | Yes (not combinable with bucket-level retention policy) | Yes | Load-balanced upload across ECS nodes with health checks | Compliance mode (Object Lock) supported. |
| Amazon S3 | aws | Yes | Yes | Yes (both directions) | Object Lock, Snowball Edge, and retention policies supported. Client-side KMS keys are not supported (TLS provides in-flight protection only). |
| Microsoft Azure Blob | azure | N/A (see Compliance Guide for retention/legal hold) | Yes | Upload only (multipart download not supported) | Object Lock-equivalent retention supported (1.1.6+). No content-type encoding — all files use binary/octet-stream. |
| Google Cloud Storage | gcs | N/A | Yes | Yes (both directions) | Native SDK API; all metadata options supported. |
| Cohesity | other | Per vendor documentation | Yes | Not supported (no multi-node load-balanced multipart) | Confirm versioning/retention support with Cohesity directly. |
| OpenIO | other | Not tested | Yes | — | Requires ARCHIVE_S3_PREFIX="igls-". |
| Ceph (v15/Octopus or later) | other | — | Yes | — | AWS v4 signature only. |
| MinIO | other | — | Yes | — | Requires ARCHIVE_S3_PREFIX="igls-"; must use --cloudtype other. |
| Scality (open source) | other | Not tested | Yes | — | Create a folder on the target that becomes the bucket name at the root of the device. |
| Wasabi | other | Tested | Yes | — | Requires full bucket permissions; special characters in metadata (for example, AD domain or file owner names) are not supported without base64 encoding — contact Support. No orphaned multipart cleanup lifecycle or abort-multipart API support on the target. Requires export AWS_USE_TRANSFER_MANAGER_V2=true. |
| IBM Cloud Object Storage | other | Tested | Yes | — | — |
| BackBlaze S3 | other | Tested | Yes | — | — |
| Cloudian | aws | Tested | Yes | Multipart upload | Bucket URL must resolve via DNS on the Golden Copy VMs using the URL access method. |
| Pure Storage FlashBlade | aws | Tested | Yes | Multipart upload | — |
| Weka S3 | other | Not supported | Yes | Multipart upload | Requires ARCHIVE_S3_PREFIX="igls-". Advanced metadata supported; Cloud Browser GUI is not supported. |
| Hitachi Vantara HCP CS | aws | Supported | Yes | Yes (both directions) | Advanced metadata and Cloud Browser GUI supported. |
| Dell ObjectScale | ecs | Supported | Yes | Yes (both directions) | Advanced metadata and Cloud Browser supported. Requires ObjectScale 1.2/1.3 with svc_param set -n com.emc.ecs.s3.auth.verify_signature_with_encoded_uri -v false -r test. |
| Quantum ActiveScale | other | Supported | Yes | Yes (both directions) | Advanced metadata and Cloud Browser supported. |
| Vast Data | aws | Supported | Yes | Yes (both directions) | — |
| NetApp ONTAP | aws | Supported | Yes | Yes (both directions) | Advanced metadata and Cloud Browser supported. |
note
Data copied into a Golden Copy target bucket using a tool other than Golden Copy will be recalled, but without metadata, unless it is placed under the same <cluster name>/ifs/... object path Golden Copy itself uses (1.1.6+).
Supported S3 Storage Classes (AWS)
| Storage Class |
|---|
| S3 Standard |
| S3 Standard-IA |
| S3 Intelligent-Tiering |
| S3 One Zone-IA |
| S3 Glacier Instant Retrieval |
| S3 Glacier Flexible Retrieval |
| S3 Glacier Deep Archive |
Storage Target Lifecycle and Tiering Operations
These are target-side (not Golden Copy-side) configuration steps for managing object lifecycle, retention, and orphaned upload cleanup on the storage bucket itself. Configure them directly in the target provider's console following the provider's own documentation.
- Retention and versioning use cases:
- Scheduled copy with optional versioning: set a bucket-level retention policy so new objects automatically inherit it, or use separate buckets for different retention needs. Enabling bucket versioning with a per-version expiry lets modified files upload while preserving prior versions; if a file changes multiple times between full copy jobs, only the version present at job time is captured.
- One-time archive for long-term/legal retention: add the folder definition with the
--paused(manual-only) option so it is excluded from scheduled copies, run the archive job once, verify success with an S3 browser tool, and only then use PowerScale's Tree Delete function to remove the source data (consult Dell documentation for Tree Delete). Set a bucket-level Time to Live (TTL) matching your retention requirement, and consider--prefixto copy the same source path to a second, non-overwriting object path if needed.
- Cleaning up orphaned multipart uploads: Golden Copy issues an Abort Multipart Upload API call to instruct the target to remove partial uploads left behind by a failed multipart transfer, but a target-side lifecycle rule should also be configured as a backstop in case the target does not honor that call. On AWS S3, add a bucket lifecycle rule (Management tab) that expires incomplete multipart uploads after a short number of days.
- AWS Glacier tiering: once an object is moved to a Glacier storage class (either manually via the object's Properties tab, or automatically via a bucket lifecycle rule based on object creation date), it is no longer accessible to an S3 browser or Golden Copy recall until it is restored. Use the bucket's Restore action (or the AWS API) to initiate a restore, specifying the retrieval speed and number of days the restored copy should remain accessible before it returns to Glacier.
- Azure Blob tier lifecycle policies: Azure's centralized Blob lifecycle management can move objects between storage tiers based on last-accessed or last-modified date; see Azure's lifecycle management documentation for policy configuration.
- Monitoring target storage against the source file system: compare PowerScale quota or InsightIQ reporting (source file count/volume) against the cloud provider's own billing and object-count reporting, and the Golden Copy GUI's total-data-backed-up figure. Expect some variance — Golden Copy also creates one ACL metadata object per folder (for example, 10 million source folders adds roughly 10 million ACL objects to the target), and file-size display differs between OneFS (logical/application size) and S3 (raw bytes); use an MD5 checksum to confirm content integrity rather than relying on size alone. As a best practice, run a scheduled Archive Data Audit (at minimum every 6 months) to reconcile source and target — review Cloud Provider Cost Considerations below first, since an audit job incurs API and egress costs proportional to the data set size.
Cloud Provider Cost Considerations
Cloud storage providers bill for API calls, uploads, downloads, and tier transitions. Understanding which Golden Copy operations drive which charges helps avoid surprise costs — reviewing and managing cloud storage costs is a customer responsibility.
| Operation | Cost Drivers |
|---|---|
| Archive job | API costs, HTTP PUT costs, per-GB storage costs. Re-walking a path to check for existing files also incurs API costs (can be skipped with --skip-s3-file-exists). |
| Recall job | API (GET) costs, plus egress data charges for the data retrieved. |
| Re-run archive job | API costs to re-check whether each file already exists on the target. |
| Archive Data Audit job | API costs plus GET/egress charges equal to the data volume audited. |
| Incremental job | API (POST) costs plus storage costs for new/changed data. |
| Lifecycle tier transitions | Per-object cost to transition between storage tiers, and potential recall costs to retrieve data out of an archive tier. |
General cost factors:
- Ingress (data copied into the cloud) typically has no bandwidth charge from the provider.
- API requests are typically billed per call; each file is at least one API call, and large files (over 10 MB) use multipart upload, adding roughly one API call per 10 MB chunk.
- Egress (data downloaded out of the cloud, including Golden Copy recall jobs) is billed at a materially higher rate than ingress — review the expected recall data volume before starting a large recall.
- Data storage cost estimation should sum the total volume to be archived/copied plus the expected monthly change rate; a GB of data on PowerScale becomes a GB of data in cloud storage (Golden Copy applies no compression).
Cost estimation tools:
- Superna provides an integrated AWS S3 cost calculator that estimates Standard, Glacier, and Glacier Deep Archive tier costs based on how Golden Copy copies and changes data (estimate only, not a guaranteed price).
- General cloud provider calculators: AWS Simple Monthly Calculator, Azure Pricing Calculator.
See Also
- Compliance Guide – Immutability, auditability, and RBAC posture.
- Disclaimers – Licensing dependencies and known product/version limitations.
- Troubleshooting – Diagnosing copy job failures, performance issues, and connectivity problems.