ECS: Cannot migrate clips from Centera to ECS - ERROR_RETENTION_INCORRECT

Clips are failing to migrate to the ECS from Centera with "ERROR_RETENTION_INCORRECT."

Example stack trace:

169.254.1.8 2019-11-21T19:30:53,212 [1215100270,sdk,ECSSYNCKDC01/3594002/WRITE_CLIP,-531c636:16e89995c11:-7a82,1] ERROR  AbstractTransaction.java (line 136) Failed to execute the transaction, aborting.
TransactionFailureException: com.emc.vipr.cashead.api.ObjectRetentionIncorrect: Failed to execute operation, error is ERROR_RETENTION_INCORRECT, object name is null
        at com.emc.vipr.cashead.storageserver.access.transaction.write.ObjectWriterImpl.finish(ObjectWriterImpl.java:80)
        at com.emc.vipr.cashead.storageserver.access.transaction.write.AwaitEOF.finish(AwaitEOF.java:173)
        at com.emc.vipr.cashead.storageserver.access.transaction.write.AwaitEOF.put(AwaitEOF.java:111)
        at com.emc.vipr.cashead.storageserver.access.msg.RichLastSegmentMessage.execute(RichLastSegmentMessage.java:61)
        at com.emc.vipr.cashead.storageserver.access.encoding.hpp.HPPResponder$EnqueableTransactionObject.execute(HPPResponder.java:311)
        at com.emc.vipr.cashead.storageserver.access.transaction.AccessTransaction.executeMessage(AccessTransaction.java:57)
        at com.emc.vipr.cashead.storageserver.access.transaction.AbstractTransaction.schedule(AbstractTransaction.java:131)
        at com.emc.vipr.cashead.storageserver.access.transaction.NullScheduler.registerSchedulable(NullScheduler.java:48)
        at com.emc.vipr.cashead.storageserver.access.transaction.AbstractTransaction.putGeneralMessage(AbstractTransaction.java:92)
        at com.emc.vipr.cashead.storageserver.access.transaction.AccessTransaction.put(AccessTransaction.java:61)
        at com.emc.vipr.cashead.storageserver.access.encoding.hpp.HPPResponder.processTransactionObject(HPPResponder.java:276)
        at com.emc.vipr.cashead.storageserver.access.encoding.hpp.HPPResponder.processPacket(HPPResponder.java:290)
        at com.emc.vipr.cashead.storageserver.access.encoding.hpp.HPPResponder.process(HPPResponder.java:181)
        at com.emc.vipr.cashead.storageserver.access.AccessNetworkServer.HandleConnection(AccessNetworkServer.java:51)
        at com.emc.vipr.cashead.hppserver.HPPServer.HandleConnection(HPPServer.java:84)
        at com.emc.vipr.cashead.net.NetworkServerWorker.Work(NetworkServerWorker.java:61)
        at com.emc.vipr.cashead.worker.Worker.call(Worker.java:227)
        at com.emc.vipr.cashead.worker.WorkerSchedulerThread.run(WorkerSchedulerThread.java:137)
Caused by: java.io.IOException: com.emc.vipr.cashead.api.ObjectRetentionIncorrect: Failed to execute operation, error is ERROR_RETENTION_INCORRECT, object name is null
        at com.emc.vipr.cashead.impl.vipr.ViPRObjectWriteHandler.checkResult(ViPRObjectWriteHandler.java:73)
        at com.emc.vipr.cashead.storageserver.access.transaction.write.ObjectWriterImpl.finish(ObjectWriterImpl.java:77)
        ... 17 more
Caused by: com.emc.vipr.cashead.api.ObjectRetentionIncorrect: Failed to execute operation, error is ERROR_RETENTION_INCORRECT, object name is null
        at com.emc.vipr.cashead.impl.vipr.ExceptionTranslator.translateError(ExceptionTranslator.java:38)
        at com.emc.vipr.cashead.impl.vipr.ExceptionTranslator.checkError(ExceptionTranslator.java:21)
        at com.emc.vipr.cashead.impl.vipr.ViPRObjectWriteHandler.checkResult(ViPRObjectWriteHandler.java:71)
        ... 18 more

Scenario 1:
If an object's retention period is outside of the bounds that are specified on the bucket, then an attempt to write the object fails.

When using retention policies, Selecting Infinite for Minimum Fixed Retention Period means that all retention values on all objects must be infinite.
Selecting infinite for Maximum Fixed Retention Period means that there is no maximum limit.

Min/max retention constrains are applied to any C-Clip written to a bucket. If a clip is migrated by any SDK-based third-party tool the retention should be within bounds, otherwise an error is thrown.

Scenario 2:
There are retention class policies defined on Centera. You can verify this by asking the customer to provide you with the below command from Centera:

Config# show retention all

Scenario 1:
So for example if you had CAS clips with retention of 7 years set on them, you would have to create a bucket with the following retention values:

Bucket Retention Period: 0 seconds
Minimum Fixed Retention Period: 0 seconds
Maximum Fixed Retention Period: 8 years
Minimum Variable Retention Period: 0 seconds
Maximum Variable Retention Period: 0 seconds

This way, the retention is within the min/max constraints.

Scenario 2:
Ask the customer to provide you with the below retention class policies.

1. Start the CLI for Centera.

2. Connect as a user with the compliance or monitor role.

3. Type the command: show retention all

   The output is similar to:

   Retention class name Period
   ------------------------------------------
   class 1       5 years
   class 2         3 months
   class 3         10 years

You then must add all these retention class policies to namespace retention policies on the ECS side. Ask the customer/PS to try the migration again after adding the namespace retention policies on the ECS side.