The following issues are known to exist in this release: 



Release 5.2.5


If no network route is defined between the Delphix Engine and target host (using the addresses listed in "NFS Addresses" field of the environment), VDB operations will fail.

Fixed in 5.3


After removing the values in the NFS Addresses field for an Environment from the UI or by removing the nfsAddressList attribute of the host via the CLI, may during a VDB enable result in the VDB reporting a false success.

Performing a VDB start should restart the VDB.

Provisioning Masked VDBs: As of we have enabled Host Name Verification for Masked Provisioning over SSL/TLS. Any SSL certificates that do not specify the FDQN provided in the masking service config as either the CN or SAN, will fail with an unverified host exception.

Fixed in 5.3

Create a new certificate including the correct hostname and upload it.

Release 5.2.4


Release 5.2.3


Release 5.2

Management Server



Adding a duplicate JDBC connection string can cause a management stack error when validating database credentials.

Delete the duplicate JDBC connection string.


An attachsource job fails.

Fixed in

  1. Login to the Delphix Management application.
  2. Select Manage > Policies.
  3. Change the policy to another policy (or make a temporary one if required).
  4. Reset the policy back again to the original policy.


Cancelling a VDB provision or refresh job may leave behind Oracle background processes and cause errors for subsequent jobs for VDBs using the same Oracle SID.

Kill any Oracle background processes on the target system that are using the same Oracle SID.


Case mismatch between Oracle database SID and static listener SID list prevents automatic creation of database connection string.

  1. Manually add connection string, or modify the static listener SID to match case with database SID.
  2. Refresh the environment.


Hooks and other customer supplied scripts should avoid writing excessive amounts of output, as the output can cause later queries for job information to run out of memory.

Fixedin 5.3



Customer scripts should avoid starting operations on the same VDB that the script is running on.

Fixed in 5.3.8

DLPX-51648Provision a VDB to a target host with a bad listener results in successful provisioning, but with wrong online redo log size, if user specifies customized online redo log size for their VDB.None


SSH algorithm "diffie-hellman-group-exchange-sha256" is not currently supported.



Using SCAN names or IP addresses to discover an Oracle cluster causes discovery to fail with "The host .. is already managed by this Delphix Engine".

Supply a node name or node IP address for cluster discovery.


After initial server setup, Default Snapshot/Snapsync policies are always in "US/Pacific", regardless of the timezone selected during initial Engine Setup. An admin has to manually change the time zone.



Upgrade does not remove cluster resource dependencies for SQL Server Clustered VDBs and can result in the cluster service going down post-upgrade. This can be worked around by disabling clustered VDBs before the upgrade and enabling them after.   

Fixed in 5.3



Refreshing "All Environments" can remove nodes of a Windows failover cluster if individual nodes cannot be refreshed.

Fixed in 5.3.8

Re-run discovery when all nodes are available.


The Delphix Engine does not provide instructions to browsers to avoid caching HTTP responses (pages).

Fixed in



The session cookie (JSESSIONID=<VALUE>) is persistent before and after login.

Fixed in



For Oracle Grid upgrade scenarios, customers can modify the cluster home location but cannot change the cluster version.

With the management stack shut down, Delphix support personnel can change the cluster version in MDS.

DLPX-55122LiveSource APPLY RESYNC takes too long to 'dd' datafiles for large dSourceDiscard the LiveSource RESYNC. Convert the LiveSource to a dSource, then convert the dSource to a new LiveSource.


When performing a CLI or API provision/refresh, the Location parameter must be explicitly set when using TimeflowPointLocation.

In the CLI, either use the defaults command to set initial values when provisioning or refreshing, or explicitly set timeflowPointParameters.location to LATEST_POINT or LATEST_SNAPSHOT.

ln API calls, explicitly set timeflowPointParameters.location to LATEST_POINT or LATEST_SNAPSHOT.


Datafiles backups created by RMAN due to a SnapSync are marked as Unavailable, however the backup of the controlfile is not marked unavailable.  Avoid using the Delphix controlfile backups.

Fixed in 5.3

Mark the backup of the controlfile taken by Delphix as unavailable:

RMAN> change backup of controlfile tag 
'<Delphix tag>' unavailable;

For example:

change backup of controlfile tag 
'DLPX2017111721493689966A19FE27' unavailable;

The 'Delphix tag' can be obtained from the command:

RMAN> list backup of controlfile summary;


If the toolkit directory for a staging environment is greater than ~43 characters, VDBs will not be able to be provisioned from the snapshots.            

Fixed in

Use a shorter toolkit directory path and re-sync'ing the dSource.


If this error occurs, one can recover by taking a snapshot on the affected VDB and attempting the replication again.

Fixed in 5.3

Take a snapshot on the affected VDB and attempt the replication again.


When configuring Serviceability from the Setup application, only a single email address may be used for testing SMTP validation.  Attempting a comma separated list of email addresses will fail.

Fixed in 5.3



Validated Sync can fail when monitoring ASE backup servers started by using the $DSLISTEN environment variable instead of the "-S" argument. This can be worked around by accessing $DSLISTEN in the RUN_xxxxx script and pass it down as -S.       

Fixed in

Access $DSLISTEN in the RUN_xxxxx script and pass it down as -S.


Environment discover does not update the version string for manually added ASE instances, which can cause sync failures due to a version mismatch with staging during a sync. This can be worked around by manually updating the version using the CLI.               

Fixed in                                                     


Unable to migrate SQL VDB between two Target Clusters from the UI.

Fixed in 5.3

Use the CLI for migration.


SQL Server dSource enable fails after a force disable.

Fixed in 5.3

Manually drop the staging database using SSMS from the staging SQL Server instance and retrying the enable. 


SQL Server VDB Refresh after a force disable can fail.    

Fixed in 5.3

Remove the database from the SQL Server instance using SSMS and re-running the refresh.

Provisioning to environments with no discovered or manually added listeners will fail.

Fixed in 5.3

If listeners are intended to exist on the Target Environment, start the listener and refresh the environment, or manually add the desired listener.

If no listeners should be on the target environment, go to the Environments page and manually edit and append a new placeholder listener.



The disable constraints feature disables all constraints, not just the ones for masked columns. 

DLPX-49184Custom Algorithms that log excessively may cause out of memory errors.None

Users need to purge their browser cache after upgrading the Delphix Masking Engine otherwise, the UI may not render correctly.

Fixed in

Manually clear the cache.
DLPX-50029Prescripts and postscripts to masking jobs may fail if the script includes a byte-order mark (BOM) character.On Windows, when using a text editor (like Notepad or Notepad++) Save and use an encoding without BOM (Byte Order Mark). The default encoding ANSI should do that - others are available and more advanced editors enable the exclusion of BOM. This applies to "Save as" as well. This avoids inserting a leading BOM character.

Masked provisioning jobs that require more than 240 minutes to execute should be called from a hook script that implements API session re-authentication.

Fixed in

Use a hook script.
DLPX-50877The "Drop Index" option on a Masking job does not support Oracle composite or functional indices.None

Auto-generated connectors remain after unsuccessful masking job. This is expected behavior and allows failed connections to be inspected. These auto-generated connectors may be deleted after an unsuccessful job.

Fixed in

Delete unwanted connectors through the GUI or the Delphix Masking Engine's V5 REST API (see API process below). The API can be used to automate the deletion.

List all connectors
For each connector
If the connector's name matches the pattern 
for an autogenerated connector name
delete the connector

Logical key specification, filters or custom SQL statements that include CR (Carriage Return,ASCII 13, \r) LF (Line Feed, ASCII 10, \n) characters may make it impossible to edit a masking rule set.

Fixed in

CR LF characters should not be used in these fields.

DLPX-56222The secure shuffle algorithm should not be used with on-the-fly (OTF) jobs.Evaluate using a different algorithm, such as secure lookup, segment mapping, or mapping.



SQL Server VDB on failover cluster may go into suspect mode during periods of high load.

Fixed in 5.3