Jump to content

Scillonian

Members
  • Content count

    556
  • Joined

  • Last visited

  • Days Won

    23

Posts posted by Scillonian


  1. Was this file in OneDrive before it was deleted?

    When this problem with Files on Demand files first reared its ugly little head, any 'local'* copy of a OneDrive Files on Demand file that was deleted would still cause the -1103 error while it was in the Recycle Bin. I never bothered to investigate too far, I negated the -1103 error by not using Files on Demand, but using a reboot to Linux (most of my systems can native boot to Linux if necessary) to look at the files system it appeared that the deleted files only exists as a reparse point in the Recycle Bin folder to the actual location where the file is stored.

    Information on this is sparse, but it looks like the code that implements the Files on Demand is under certain circumstances creating invalid data structures in the NTFS MFT which most likely confuse Retrospect. When I've tried scanning NTFS volumes from Linux with Files on Demand enabled that give the -1103 error they always reported an unclean with an unknown error. NTFS volumes with Files on Demand enabled that don't give the -1103 error always reports as clean.

     

    * I have yet to find where the local copies of One Drive Files on Demand files are physically stored. Trying to resolve the target of the reparse point that is the OneDrive folder leads me nowhere. The local copies of Files on Demand are not much use as backups when Windows 10 won't boot and you need to perform data recovery from Linux.


  2. x509

    Do you have InstantScan running on the drives in question?

    I have encountered situations before where, although InstantScan is active and reports no detectable errors, the volume databases stops updating and so as far as Retrospect is concerned no new files have been added files have been added, changed, or deleted on the volume. When a file is added, changed or deleted the volume database should be updated fairly soon afterwards so if the file time stamp of the volume database is not updating then InstantScan is not updating.

    The InstantScan volume databases are in the C:\ProgramData\RetroISA\RetroISAScans folder. To reset stop the InstantScan services, delete the volume databases, the restart the InstantScan services.

    • Like 1

  3. If you look in the retroISA_log.utx file you will see entries similar to:

    	[*] walkUSNjournal: err 0xc0000022 adding path for $TxfLog.blf fileFRN 0x100000000039e with parentFRN 0x100000000039b
    	[!] walkUSNjournal: err 0xc0000022 for V:\
    	[*] ISAVol::Exit: error -1073741790 (unknown), deleting scan file for "V:\"
    	[*] ! ISAVolCon::IsaVolStopOne: stopped processing "V:\" on 2018-09-14 22:26:39
    

    and

    	[*] walkUSNjournal: err 0xc0000043 adding path for datastores.xml.tmp fileFRN 0x250000000005db with parentFRN 0x2a0000000015ed
    	[!] walkUSNjournal: err 0xc0000043 for C:\
    	[*] ISAVol::Exit: error -1073741757 (unknown), deleting scan file for "C:\"
    	[*] ! ISAVolCon::IsaVolStopOne: stopped processing "C:\" on 2018-09-16 10:13:18

    and

    	[*] walkUSNjournal: err 0xc0000043 adding path for amd64_microsoft-windows-directui_31bf3856ad364e35_10.0.16299.492_none_c449a34c59896544 fileFRN 0xce000000000f4e with parentFRN 0xc480000000000f5
    	[!] walkUSNjournal: err 0xc0000043 for C:\
    	[*] ISAVol::Exit: error -1073741757 (unknown), deleting scan file for "C:\"
    	[*] ! ISAVolCon::IsaVolStopOne: stopped processing "C:\" on 2018-10-16 21:46:11
    

    The file $TxfLog.blf is one of the NTFS special files.

    The [Windows] error codes translate to:

    0xC0000022  STATUS_ACCESS_DENIED  --  A process has requested access to an object but has not been granted those access rights

    0xC0000043  STATUS_SHARING_VIOLATION  --  A file cannot be opened because the share access flags are incompatible.

     


  4. I noticed the same problem when Instant Scan first became available. There are certain special NTFS management files (Journal, MFT, etc) that cause the Instant Scan process to fail, delete the scan database, and abort any further scanning until the system is restarted. I opened a support case on this and basically the problem is these special files don't like Instant Scan 'looking' at them and generate a sharing violation which aborts the scan.

    With each new release of Retrospect the Instant Scan scanning process has become more reliable and/or more tolerant of these special files. What I have noticed is some systems are more prone to this than others. The catch is you need to upgrade to the current version of Retrospect to get a more reliable Instant Scan.


  5. An update will show in Automatic Update and a manual check for updates from within Retrospect a few weeks after initial release. This presumably to allow the early adopters who go looking for the update to find any problems before the rollout to the masses via Automatic Update.

    As for safe to download and use, well I downloaded and installed it and so far the update has run without problems for me. (YMMV.)

    • Like 1

  6. Never experienced this situation (yet) but there could be Retrospect it trying to show you a dialog that requires your action but has become hidden behind the main Retrospect window*. Under the normal Windows UI tools there is no way to get this hidden dialog to the foreground with the result that the only way to regain control is to kill Retrospect.

    This hidden dialog problem is not unique to Retrospect and many Win32 apps experience it on Windows 10 in particular.

    There is a way to to bring these hidden dialogs to the foreground using a third party program. (Will detail when I'm back at the Windows PC.)

     

    * When a Win32 program wants to put up a dialog for user interaction that dialog should appear above the host program in the 'Z' order. Especially Windows 10 seems to get the 'Z' order wrong and puts the dialog under the host program window with no way in the Windows UI to bring it to the foreground. (I've also had Windows put the dialog at coordinates that are outside the valid display area.)


  7. From from memory...

    If you set up a Restore with the option to replace missing files the click then button to view the files that will be restored a list of files should be shown. Files with a black diamond exist in both the backup and the destination. Files without a black diamond exist only in the backup. (I don't have access to Retrospect today so can't say the exact steps.)


  8. As you are using private IP address space anyway, would it be possible for your organisation to transition from the two Class C 192.168.45.m and 192.168.183.m networks to a single Class B 172.n.m.m. network? This would give you more than enough IP addresses for the 450 machines. (Would probably make network administration simpler too.)

    EDIT: Unless there is something really weird going on in your network the change from Class C to Class B can be achieved by changing the scope of the DHCP server.


  9. Try doing a thorough rebuild of the catalog for the cloud backup set. To effect a thorough rebuild you will first need to delete all the *.SESSION files that are in the folder(s) that hold the *.RDB files for the backup set. Once the *.SESSION files are deleted rebuild the catalog. A thorough rebuild of the source backup set may also be advisable just in case an error their cause the initial snapshot transfer failure.

    Because Retrospect is now reading the file information from the .RDB files instead of the .SESSION files, and depending on the connection speed to the cloud provider, the rebuild could take some time to complete.

    Hopefully after the thorough rebuilds a rerun of the transfer snapshot should pickup the missing files.


  10.  

    2 hours ago, cfieldgate said:

    For this backup it is both the source and the destination - one volume being the source (RAID 6, 5 disks) and another volume (1 disk) as the destination. It is this 1 disk it seems to have a problem with 'finding'.

    Am I correct in thinking that you are accessing the source RAID6 volume using the Retrospect Client and writing to the destination single disk volume over SMB?

    What is the CPU load on the NAS during this backup?

    I am wondering if the Retrospect Client is intermittently overloading the CPU.


  11. [In theory] Retrospect should be able to restore from read-only media.

    More of problem will be trying to mount a Linux file system under Windows. I've found it bad enough try to reliably read the older (and simpler) ext2, ext3, and ext4 file systems. BtrFS is still evolving with all that entails. Also with Microsoft pushing out a new version of Windows 10 every six months, at present, what may work now will be broken with the next version of Windows 10.

    Here are two more reliable solutions to consider:

    1. Mount the BtrFS drive and an NTFS on a Linux OS and copy the Retrospect files from the BtrFS drive to the NTFS drive. (If you only have one computer then boot that computer from a Linux live CD/DVD/USB to perform the copy.)
    2. If you have more than one computer and a network (preferably not wireless) then mount the BtrFS drive on a Linux OS and use SAMBA to share it over the network. A Linux live CD/DVD/USB can be used to do this if you do not have any native Linux computers.

    Ultimatly the only way to really know what will work is physical test the emergency restore procedure and keep testing it until it works reliably. The time spent now will be rewarded later when the restore becomes a real event not a theoretical one.

    "There are no successful backups, only successful restores."


  12. Is this SSD a replaceable one with a SATA, mSATA, or M.2 connection or is it soldered to the motherboard? The fact that Apple's Disk Utility says that SMART is not supported on the SSD would lead me to believe it is one that is soldered to the system board.

    Do you have easy access to an Apple Store? As you have warranty cover a thorough health check at the Apple Store would eleminate hardware as a problem.


  13. This could be a hardware problem. Personally I would check that the hard disk in the MacBook is healthy. From what you describe this would be indicative to me, if this was a Windows laptop, that the hard disk was failing.

    Firstly I would check the SMART data from the disk to see if there is anything untoward going on. (Not being a macOS user I don't know what utilities there are available for macOS to read the SMART data.) If you are not familiar with interpreting SMART data post the results here and I look over them.

    I've encountered enough failing hard disk to know that under normal day-to-day usage they appear fine but once something like Retrospect starts to use the disk intensively the problems start.


  14. 1 hour ago, jmosk said:

    I have no idea why a change in the Virtual Memory size setting would have resulted in an improvement in the backup speed.

    My suspicion is that with your original Virtual Memory setting, and Retrospect running, Windows was periodically deciding that some memory pages needed swapping between physical RAM and the pagefile more often than was desirable resulting in the erratic performance of Retrospect.


  15. 2 hours ago, jmosk said:

    I am not sure where to look to try to locate this bottleneck that is causing backup speeds to vary to such a high degree.

    To better try to find where the bottleneck may be you need need the extra functionality that is in the Task Manager in Windows 8.x and Windows 10. The feature available in Windows 8.x and Windows 10 that is of use here is the display of the percentage use of the system's individual resources. It is quite possible for a process to have a very low disk data throughput but be using 100% of the disk resources. This information is not visible in the Windows 7 Task Manager.

    A search online for use windows 8 task manager in windows 7 will reveal some third party Windows 8.x style Task Manager alternatives for Window 7. Being a Windows 10 user now I have no personal current experience of these solutions for Windows 7 so provide this search information as is for a starting point.

×