Fixed Issues

• HSR: Disk Geometry and Alignment: The Virtual Disks (VHD, VHDX, VMDK) generated using ImageManager Head Start Restore (HSR) are now optimized to provide improved read/write performance.
• Fixed an issue where the daily collapse process was starting but then didn't create a collapse file. Under rare circumstances, ImageManager failed to collapse files within a backup chain. This may have occurred in situations where ImageManager was collapsing files without any problems for a period of time. The issue appeared to have occurred more frequently when the backup chain was on a SMB/CIFS share. This is fixed in this release.
• Note: We have also seen similar issues when Linux-based NAS devices have hosted the SMB/CIFS share and where the NAS has been joined to an Active Directory Domain. The particular cause of this error is the result of permission problems on the shared folder.
• Fixed an issue where HSR Jobs stop after upgrade: In some rare situations, an upgrade of ImageManager may cause an existing HSR Job to stop if it was writing a VMDK file to a VMWare ESXi host Datastore. You may have seen a pop-up error message that said: "Target not found [datastore1]". You can see evidence of the problem in the ImageManager install log file:
• Copy

  RegRemoveValue(Name=VirtualFile,Value=[datastore1] ExSvr1 6-29/C.vmdk,)
  Executing op: RegAddValue(Name=VirtualFile,Value= ExSvr1 6-29/C.vmdk,)

• This happened because the HSR path to the VMDK file was getting changed in the registry and the "[datastore1]" component of the path was removed. Users were able to edit the registry key and add the missing part of the HSR path back in as a short term work-around. The underlying problem is fixed in this release.
• Fixed an issue where the Managed Folder state was incorrectly shown as Green (OK) in the user interface even when verification failed. Under rare situations, the ImageManager user interface was showing a managed folder in a GOOD state (green) in the top of the user interface and was showing a BAD state (red) in the lower part of the user interface. This occurred when a file verification failed. This has been fixed in this release.
• Fixed an issue where ShadowStream replication was stuck on Ready. In a rare situation, a replication job may have been stuck in a ready state. This would occur in the following scenario: (1) ImageManager was running at both the local/client side and the remote side. (2) The remote ImageManager deleted some -CM files as part of retention. (3) The local ImageManager had not yet removed the original -CM files as part of retention. (4) The local ImageManager service was restarted. After restarting the service the -CM files were queued to send again. This is fixed in this release.
• Fixed an issue where iFTP was writing files to the root directory after losing connection: Under rare conditions, an iFTP Replication job may have written files to the root folder on the destination FTP server rather than in the sub-folder specified in the replication job. This problem was more likely to occur if the network connection had been lost.
• Fixed an issue where stopping/restarting the IM service caused the 1053 error (service did not respond to the stop request) to be received. In some rare situations, the ImageManager service would hang or not respond to a restart when it was performing a Consolidation operation. When stopping the IM service, the service stop request failed with error 1053 (the service did not respond to the start or control request in a timely fashion). After the service was restarted, users were unable to connect to the ImageManager console unless the machine was rebooted. The ImageManager logs showed the following message when attempting to connect: "ImageManager.ThreadProc exception: Only one usage of each socket address (protocol/network address/port) is normally permitted." This issue has been fixed in this release.
• Note: It is also possible to see the 1053 Error temporarily in one other scenario where an attempt is made to Resume a ShadowStream Replication Job that is in a failed state, and then an attempt is made to stop or restart the ImageManger service. If this happens just wait 60 seconds and the service will restart without issues.
• Fixed an issue where ShadowStream replication was hanging after send failures and would not retry unless the IM service was restarted. In some rare situations, a new ShadowStream Replication Job might hang after send failures. Replication would not continue until the ImageManager service was restarted. This issue has been fixed.
• Fixed an issue where an HSR to VMDK failed because the destination volume was too small. ESXi HSR jobs where the source volume was resized no longer require a full re-managing of the backup folder. You now only need to delete the HSR job and re-create it to a new disk of the appropriate size.
• Fixed an issue where ShadowStream replication failed with error "unable to delete .tmp file". In some rare situations, ImageManager may have reported the following error in regard to a ShadowStream Replication Job: “unable to delete .tmp file”. This is fixed in this release.
• Fixed an issue where the HSR commit file process was failing with error -1326 (bad username or password). In some HSR jobs, replicating to Server 2012 R2 shares failed with authentication errors. This is fixed in this release.
• Fixed an issue where the bad.log file was not being created when the ReadHeader error was detected outside of normal verification procedures. ImageManager now creates a bad.log file, even when the ReadHeader error is detected outside of normal verification procedures.
• Fixed an issue where files being sent by ShadowStream show up corrupted on the receiving side. In some rare situations where there was a large number of corrupt packets on a network connection, there was a chance that ShadowStream could write a corrupt file. This is fixed in this release.
• Fixed an issue where an erroneous error message was logged on failure: "Log open exception". Under some rare circumstances, the following error message was logged in the ImageManager log file(s): "Log open exception". This could have occured where there was a failure with an operation such as replication. The message did not affect the running services and was benign. This release fixes the issue and the error is no longer generateed in the logs.
• Fixed an issue where changing the backup password in a local HSR job failed. The process could fail when attempting to correct a wrong backup password in an HSR job to a local drive if the backup had two volumes or more. This was not an issue when writing to an ESXi server. This issue is fixed in this release.
• Fixed an issue where if detailed logging was turned on, there was a chance the password associated with validating an HSR commit file would be included in the log. The password is now obscured in the log file if the user enables detailed logging.