Leaderboard

Excited to report that I got my Retrospect backups fixed now as well !! So, disabling the "Save space and download files as you use them" setting did not work for me. So in addition to that, I tried disabling OneDrive altogether (Unlink PC/account and then remove the entire OneDrive folder) and then linking my account again which then again downloaded all files. The next Retrospect scheduled (proactive actually) run then succeeded with 0 errors or warnings ! This worked now on 2 machines already, so will try kids laptop next, but pretty confident this will work as well. Thanks to the folks active in this thread !

Ok, I did some additional debugging and I found the issue: It's OneDrive. OneDrive has an option (don't know if it's new with 1803) that gives you the possibility to only actually download a file from OneDrive to your computer if you actually use it. All files in the OneDrive folder are shown as being present, but they are not actually physically present on your computer. So they are some sort of link. The explorer shows a cloud symbol with these files. As long as one such link is present on the drive (even if it is in the Recycle bin) it will cause the -1103 error. Scanning will be interrupted as soon as the scan engine hits the OneDrive folder. On both machines I turned the option off in OneDrive settings, which caused all OneDrive files to be downloaded to the computer, and both are backing up properly now, without errors. Also checked the machines that never showed the issue, they all have this OneDrive option set to off.

If you disable Open File Backup you will get the can't read, error -1020 (sharing violation) error because the files are in use by Windows. The two different data size errors in the compare will also be related because the file has been modified by Windows between the backup and the compare.

Hi All, Just another quick update. We are still investigating these errors, most of which appear to be tied to One Drive, but it seems not necessarily always tied to One Drive. The results from the internal investigation indicates that the problem is a bug introduced by Microsoft. Just to give a simplified overview, when using VSS we aren't ever actually backing up your files, we backup Shadowcopies of your files. In the case where we run into problems, after we access the bundle of Shadowcopies that Microsoft is preparing for us (after using standard VSS calls), we encounter write access violations. So VSS prepares everything, gives us this blob of data, and then says we aren't allowed to access it. What's also strange is we aren't actually writing during the scan, we are just reading the data. We confirmed that this isn't just bad error mapping on our part, we are actually receiving write access violations. We are implementing some changes that will be a part of the next release (v15.1) which should allow the backups to continue beyond this error. This is currently scheduled for release later this week. We will continue to work with Microsoft to address the issue further, whether that means changes on our end, or a patch on their end. We really appreciate the feedback from this forum thread, as well as the all of the information submitted to us via web tickets. Thanks, -jeff

In Windows 10 Home, I unlinked the account, turned off OneDrive and in Add / Remove programs, uninstalled it. That allowed me to then delete the OneDrive directories. True, the OneDrive folders returned later as visible in Windows Explorer ... but the issue with the directory is gone. Wonders never cease.

Are you certain that you disabled the OneDrive Files on Demand feature for all users? If the Files on Demand feature is enabled the OneDrive folder becomes a Reparse Point instead of a real folder. To check if the OneDriver folders is a real folder or a Reparse Point use one of the following two command line methods. If PowerShell is set as the default terminal: Right click on Start and select Windows PowerShell (Admin) Once the PowerShell window has opened type cd C:\Users\henkv and press enter Type ls one* and press enter The will return something like: Mode LastWriteTime Length Name ---- ------------- ------ ---- dar--l 2018-05-08 19:15 OneDrive If the last character of the mode string is a "l" (lowercase L) then the Onedrive folder is a Reparse Point If Command Prompt is set as the default terminal: Right click on Start and select Command Prompt (Admin) Once the Command Prompt window has opened type cd C:\Users\henkv and press enter Type ls dir /al and press enter If the result is: 2018-05-08 19:15 <DIR> OneDrive then the OneDrive folder is a Reparse Point If the result is: File Not Found then the OneDrive folder is a real folder Even if exclude folder(s) and/or file(s) in a selector they are still scanned.

Hi x509, This will depend on the results of the investigation. If it looks like something is broken we'll submit a bug report. In fact, depending on what is broken we may have to submit a bug report in order to get it fixed. Cheers, -Jeff

Hi All, Thanks so much for all of your input. We have reproduced the issue in house, we are investigating an appropriate fix, indeed it does appear to be due to a change in Microsoft's One Drive. I'll keep you all updated on our progress. -Jeff

Just to throw out an idea after catching up with this thread, you might also want to investigate the settings in Control Panel | Sync Center in relation to One Drive, as well as anything else. Sync Center might have a bearing on this. I saw a bulletin regarding a completely unrelated application yesterday that encountered errors caused by the Sync Center. Then I read another article today discussing how One Drive syncs files and folders to the MS cloud. Now this thread to bring it full circle. I'm going to continue to watch this thread, and want to say that I really appreciate the extremely professional community discussions on this Forum. Bravo.

Lucky_Phil and others, I am now seeing similar problems. When the script runs from the scheduler on the host (local) computer, it fails during a scan of the C: drive. No problems with other drives on the host computer. Running the script via the Run pull-down will run successfully. I believe this may be a different problem than the client issues, although both are throwing the same -1103 error on the host. The open file backup on the client C: drive is failing and throwing VSS errors in the clients' event logs. The error propagates back to the host as a -1103 error. This problem only occurs on the C: drive, not on other drives on the client. For the problem backing up on the local machine, I can find no related problems in the windows event logs.

I am running Retrospect 15 and I'm having similar problems backing up a Windows 10 (Build 1803) laptop as a client. The client has 2 drives: C: (the OS) and Z: (a recovery partition). There are no issues backing up the host, which is a Windows 10 (Build 1803) desktop computer. No issues backing up 2 other clients that are Windows 7 computers. The first backup of the client laptop was working. It passed through the scan and was backing up files. However, I had to stop the backup midway for unrelated reasons. The next 2 attempts at a backup made it successfully through the scan for drive C: and some files were backed up. However, the -1103 error was thrown for the C: drive and the backup aborted. No issues at all backing up the Z: drive. All subsequent attempts for open file backup on the client resulted in the C: drive failing during the scan and no problems with the Z: drive. I confirmed VSS errors in the event log on the client at the same time as the -1103 errors. I can do a backup with "open files" disabled, although there are several warnings for sharing violations, which I assume would be expected. I have submitted a support ticket.

My testing so far suggests to me that during the upgrade from Windows 10 1709 (16299) to Windows 10 1803 (17134) some security settings related to Windows services are not being carried forward correctly. In the case of the Retrospect Client this is preventing it from correctly interacting with VSS for system volumes. The test machine I have 1803 installed on has two volumes — a system volume and a data volume. The system volume fails with the above error but the data volume completes without error. Initial comparison of some security permissions between a 1709 Client and the 1803 Client show some settings missing. However to be certain it is not something I changed in the past I'll need to do some more checking and do an 1803 clean install on something.

Stu, After a certain amount of playing around, I got Retrospect Mac 5.1 running on my ex-wife's old HDD and seeing my old tape drive attached to her old Digital Audio G4 tower (which she gave me for storage about 13 years ago after she moved to her own apartment). I did a search on one of the Storage Sets, and found 4 files ending in .tiff. However when I tried to Restore them, it turned out that no DAT/DDS tape I have with that Storage Set name matches the date that the Catalog File is expecting. Before I recreate a Catalog File from one of those tapes, please describe step by step what you did.

Oh well, #sigh I have a few Macs. All on WIFi All on DHCP They have to be this way, because they roam from office-to-office, client-to-client, home-to-work-to-home ... like laptops do. The error -530 and associated -519 has be around in almost every version of retrospect I have used for the past 20 years. I can never find the solution, apart from a full client reinstall, and I can not be doing that every week TBH! Is there ever going to be a stable answer? Each Mac, Client and Server has no firewall or anti-virus FYI. The fixed desktop Laptops that do not leave the office have no issues. The laptops returning to the office login automatically to the WIFI and have no network issues for everyday work; email, www, ftp, server shares, printing ... what so ever. Server: Mac OSX 10.11.6 Client: OSX 10.12.6 Retrospect 13.5 Please don't say 'upgrade to the latest version of Retrospect or I may have to slap you :-) That has never been an answer to this issue.