¶ RioCruise - Read Me
¶ Download Rio Cruise
Version 1.1.2 on Spectra Share
Releases on GitHub (requires credentials)
Configure Rio
Rio Cruise operates a client, submitting jobs to Rio (RioBroker). All data transfer is performed by Rio, and all source and destination brokers, and the endpoint used as a landing zone must all be configured in the RioBroker application before migration.
¶ Devices
Review device configuration and add new device if required from Rio UI → Settings → Devices: We support BP, VS3, Flashnet, and Diva migration devices.
¶ Diva Device
We support only Divarchive WS API (DIVAS) in Rio 3.1. Other API endpoints like Enterprise Connect or newer version of WS Api are not supported.
For all Diva versions 7.0 and up use: http://DIVAS_Host:9763/services/DIVArchiveWS_SOAP_2.1?wsdl
For Diva version 6.5 use: http://DIVAS_Host:9763/services/DIVArchiveWS_SOAP_1.0?wsdl
Flashnet Device
Brokers and Agents
RioCruise migrates content from Source Broker to Target Broker. We can migrate content from any supported device brokers (BP, VS3, Diva or Flashnet), but we can migrate only to Spectra device brokers (BP or VS3).
While adding read only agents brokers (Diva or Flashnet), we need to add one write agent, which is a constraint to add a broker. For diva, for agent name use same name as a category.
¶ Diva Agent
We need to add one diva agent for each category. In the above example agent is for category "Movies". Broker can have multiple diva agents provided all those categories doesn't have any duplicate files.
¶ Flashnet Agent
Endpoints
1. The endpoint name should be same in Rio and Diva or Flashnet (e.g., news_migration) because Rio Cruise uses it in restore/archive jobs.
2. We support only NAS type in Rio (not FTP or S3).
3. In Rio URI can be local drive: file:///C:/RioCruise/Migration or remote: file://ServerIP/RioCruise/Migration. Currently don't support FTP or S3 endpoints for migration.
4. In DIVA, name Source/Destination exactly same endpoint device name as migration endpoint in Rio. Configure CIFS source/destination type to point to the same folder:
Source Name: news_migration
IP Address: <Leave it empty>
Source Type: CIFS
Connection Options: -login <User> -pass <pwd>
Root Path: \\Server\RioCruise\Migration
Diva Configuration Utility → Source/Destination example for RioCruise:
Diva directly restores files to this location. Rio just submits restore requests to DIVA using object name, category and endpoint (Source/destination).
6. Test: Make sure Restore form legacy and Archive to BlackPearl shares the same endpoint.
¶ Migration Steps
1. Start Rio Cruise from command line like this:
rio_cruise.bat -h <host_of_rio_broker> -rp 5050 -p 5051 -bs 10 -bn 3 (use rio_cruise.bat --help)
2. Login to RioCruise UI: http://<host_of_rio_cruise>:5051. Ensure that source and all target brokers, and all endpoints are available in dropdowns
3. (Optional) From Gear menu (settings -> Special Handling -> Select source broker in left pane to view sample data. Update JavaScript method in right pane according to migration requirements - available parameters (RioCruise_TBPFR, RioCruise_Rename, RioCruise_Group, etc.) are shown in the sample.
Please get help from Spectra Logic for this purpose.
Test script (Clicking the Test button should draw a table of sample output at the bottom of the page. If this does not appear, there is an error in the JavaScript.
Run script so that RioCruise updates metadata.
4. Go to Migrate page.
5. Select Source Broker; Target Broker and Landing Endpoint.
6. Start Migration.
7. You can monitor progress, status or download report on same window.
8. To find objects migrated or failed go to Objects page.
¶ Best Practices
- In production use NAS endpoints with uri: file://ServerIP/RioCruise/Migration, so that all Rio Cluster Nodes, Rio Cruise and Devices can access same path for data.
- If you are using bn = number of drives (multiple tape) in the server execution command, it's good to run a single migration at a time. Use Special handling to utilize specified number of drives (tapes) effectively during migration. Rio Cruise will select bn number of tapes and submit migration requests on all tapes simultaneously. Once all bn batch of requests on these drives are done, it will submit next set of requests. Migration of selected tapes will continue until all files on those tapes are migrated.
- When you are adding Legacy Device agents (Diva, Flashnet) use dummy bucket for write agent (BP or Vail). So create a dummy bucket in BP/Vail before start adding any read only legacy agents. You can use same "dummy" bucket for all brokers. Make sure you don't archive any content to these legacy agent brokers (dummy bucket).
- For Diva migration, you can add as many categories as possible to a broker. If you don't have lot of categories, it's better use one category per broker and migrate to one target bucket. That way you can map category → bucket, but there is no restriction for target broker.
- Check if there is a possibility of duplicate files in different categories. If file content is same, then last one indexed in a Broker wins, we should be okay. If content can be different, then plan to add categories to different brokers and migrate to different buckets on BP.
- As soon as Migration is started with Source Broker, Target Broker and Endpoint selected, save migration or bookmark it, so that you can restart the same migration if RioCruise stopped for any reason.
- Work with PS team member to come up with Special Handling Scripts to customize migration for your needs.
¶ Logging
By default, RioCruise logs are 100MB in size and the last 10 logs are kept and rolled. If a client has a very large migration, some useful logging may be lost. To increase log size and/or the number of logs to keep modify these two lines in the log4j.properties file:
appender.rolling.policies.size.size=100MB
appender.rolling.strategy.max = 10
¶ Limitations
- If you are migrating from DIVA using Diva Agent, then you can add multiple read agents only if there is no duplication of objects or files across these categories. Only last one indexed wins which is same restriction we have on BP Agents indexing. Rio doesn't support versioning. If files are present in different folders, then it's okay as we include folders as part of object name.
- Multi-folder/files objects in Diva will be converted to single file (including path) objects in BP. Restoring of these objects from BP will retain original structure. Same with Diva Complex objects.