locked
New Windows 7 install failing to backup - "The network connection between the computer and server was lost" RRS feed

  • Question

  • Hi Guys,

    I've just installed Windows 7 onto a machine. It has 3 drives: E: (SSD drive with the OS on it) C: D: (Older mechanical drives)

    I use WHS to backup this machine.

    Most nights I'm getting the error message "The network connection between the computer and server was lost. Connection marked dead by transport". C: and D: always backup successfully - E: sometimes fails.

    It's not consistent in that I had a successful full backup 2 nights ago.

    All of the drivers etc are up to date and I don't have any other problems with moving large amounts of data around?

    The LAN is all wired connections - there is no wireless setup.

    This exact machine was fine under Windows XP? It's only after installing Windows 7 that I now have a problem.

    Any ideas?
    • Changed type kariya21Moderator Thursday, January 14, 2010 2:28 AM technical question
    Wednesday, January 13, 2010 8:04 AM

Answers

  • I have to admit, that I cannot read these logs very well. But the message System.ArgumentException: Value does not fall within the expected range. seems to fall rarely in context with Windows Home Server, so it is not a common issue. Since you say, it happens during backup of the SSD only, I could imagine something playing a role here.

    If I remember correctly from stuff I read in the past, SSD controllers try some strange things to arrange data on different blocks to reduce wear and tear on single addresses, so they do behave a little bit different than normal disks. Creating a shadow copy on the drive will fill signifant parts of the drive temporary. Maybe this block usage maintenance process comes into a conflict with the VSS copy, and maybe this is causing a timeout.
    You see, it is all very speculative. I would recommend to register on Microsoft Connect for Windows Home Server and issue a bug report since such situations potentially caused by certain hardware are something, we cannot solve in the forums.
    Best greetings from Germany
    Olaf
    Tuesday, January 19, 2010 12:22 PM
    Moderator

All replies

  • Hi,
    is your client set to save power or hibernate after a certain time? If so, could this be the cause, that this time limit is hit? So the client goes away, while the home server still waits for data.
    To test this, you could change the power settings to increase the time the system stays up or disable this temporary.

    Best greetings from Germany
    Olaf
    Wednesday, January 13, 2010 8:17 PM
    Moderator
  • Hi Olaf,

    Thanks for your reply - No I've set my machine not to go to standby at all - not the machine or the hard drive.

    Thanks,

    Andy
    Thursday, January 14, 2010 12:53 AM

  • Thanks for your reply - No I've set my machine not to go to standby at all - not the machine or the hard drive.
     

    This answer is a bit hard to interprete - did you verify the settings or did you only not configure anything in this direction? A fresh installed Windows 7 has some power save settings enabled on its own, thats why I ask.

    If the power save settings are really not enabled, you could also check, what is the leasing time of your DHCP server (ipconfig /all on the client should show this information). If it is too short, bad stuff may happen.

    Would a backup always complete, if you unselect the normal disks from being backed up and only backup the SSD? If this fails as well, the issue could also be bound to the SSD. Can you see any related error messages or warnings in the event logs of the client?

    Best greetings from Germany
    Olaf
    Thursday, January 14, 2010 10:14 AM
    Moderator
  • Hi Olaf,

    I had another backup failure at 3:36am last night.

    Hmm, interesting - There is an application event on the client machine:

    taken from The client machine application event log --->
    Backup of volume E: to LONDON-WHS failed: System.ArgumentException: Value does not fall within the expected range.
       at System.Runtime.InteropServices.Marshal.ThrowExceptionForHRInternal(Int32 errorCode, IntPtr errorInfo)
       at Microsoft.HomeServer.Common.Client.Win32.ReadFile(Handle fileHandle, Byte[] buffer, Int32 bytesToRead)
       at Microsoft.HomeServer.Backup.BackupOp.BackupOperation.ReadMasterBootRecord(VolumeInfo volumeInfo)
       at Microsoft.HomeServer.Backup.BackupOp.BackupOperation.RunWithoutCatch()
    --->

    I turned off all the power saving in Windows 7 - so it's definitely not that.

    It's not the DHCP either - the lease on the IP address was valid all the way through yesterday:

    Taken from ipconfig -all --->
    Connection-specific DNS Suffix  . :
    Description . . . . . . . . . . . : Marvell Yukon 88E8001/8003/8010 PCI Gi
    Ethernet Controller
    Physical Address. . . . . . . . . : 00-18-F3-29-98-4A
    DHCP Enabled. . . . . . . . . . . : Yes
    Autoconfiguration Enabled . . . . : Yes
    Link-local IPv6 Address . . . . . : fe80::b474:435:8943:7502%11(Preferred)
    IPv4 Address. . . . . . . . . . . : 192.168.1.21(Preferred)
    Subnet Mask . . . . . . . . . . . : 255.255.255.0
    Lease Obtained. . . . . . . . . . : 18 January 2010 18:36:21
    Lease Expires . . . . . . . . . . : 21 January 2010 18:36:21
    -->

    Any more ideas?????

    Andy
    Tuesday, January 19, 2010 7:53 AM
  • In fact - Here's the full capture from the WHS backupengine log:

    Created at 19/01/2010 08:02:06
    ***** Started at 19/01/2010 08:02:06 ***
    ***** Started at 19/01/2010 08:02:06 ***
    [1]100119.080206.3134: Status: Start
    [1]100119.080206.5966: Status: Connect to server LONDON-WHS, session ff08ccd7-45e3-4e93-be93-5b754e9f1ae8, protocol 20
    [1]100119.080206.6796: Status: Volume 292ff7bd-fbbd-11de-bf9b-806e6f6e6963 = E:
    [1]100119.080206.9082: Status: Acquire volume shadows
    [1]100119.080208.5869: Status: Volume 1 of 1 is E: size 128026450944 used 37598432256
    [1]100119.080208.6093: Status: Start volume E: using \\?\GLOBALROOT\Device\HarddiskVolumeShadowCopy18
    [1]100119.080209.0361: Status: Volume (shadow) size 128026447872 used 37923287040
    [1]100119.080210.9589: Status: File records: 62242 sent, 19178 changed, 18 metadata, 62149 received (bytes sent=1228800 received=1234751)
    [1]100119.080210.9677: Status: Prepare fixups: total restore size 75857575936
    [1]100119.080211.3271: Status: Determine changed: scan 2733888 size 11198005248
    [1]100119.080340.2910: Status: Determine changed: scanned 2733888 of 6690871 total with 4 fixups (bytes sent=57039399 received=0)
    [1]100119.080705.7089: Status: Server phase Reorganize1 complete
    [1]100119.080705.7109: Status: Send changed: 293736 requested of 2733888 total
    [1]100119.081055.1591: Status: Send changed: 293736 sent with 3 fixups (bytes sent=1205134707 received=83910)
    [1]100119.081055.2597: Status: Exception: ClientException
    [1]100119.081055.2597: Status:          : ArgumentException
    [1]100119.081055.2597: Status:          : Value does not fall within the expected range.
    [1]100119.081055.2597: Status:          :    at System.Runtime.InteropServices.Marshal.ThrowExceptionForHRInternal(Int32 errorCode, IntPtr errorInfo)
       at Microsoft.HomeServer.Common.Client.Win32.ReadFile(Handle fileHandle, Byte[] buffer, Int32 bytesToRead)
       at Microsoft.HomeServer.Backup.BackupOp.BackupOperation.ReadMasterBootRecord(VolumeInfo volumeInfo)
       at Microsoft.HomeServer.Backup.BackupOp.BackupOperation.RunWithoutCatch()
       at Microsoft.HomeServer.Backup.BackupOp.BackupOperation.Run()
       at Microsoft.HomeServer.Backup.BackupOp.BackupSetOperation.RunWithoutCatch()
       at Microsoft.HomeServer.Backup.BackupOp.BackupSetOperation.Run()
    [1]100119.081055.2607: Status: Bytes sent=1263413690, bytes received=1318923, 0 reconnects

    Tuesday, January 19, 2010 8:17 AM
  • I have to admit, that I cannot read these logs very well. But the message System.ArgumentException: Value does not fall within the expected range. seems to fall rarely in context with Windows Home Server, so it is not a common issue. Since you say, it happens during backup of the SSD only, I could imagine something playing a role here.

    If I remember correctly from stuff I read in the past, SSD controllers try some strange things to arrange data on different blocks to reduce wear and tear on single addresses, so they do behave a little bit different than normal disks. Creating a shadow copy on the drive will fill signifant parts of the drive temporary. Maybe this block usage maintenance process comes into a conflict with the VSS copy, and maybe this is causing a timeout.
    You see, it is all very speculative. I would recommend to register on Microsoft Connect for Windows Home Server and issue a bug report since such situations potentially caused by certain hardware are something, we cannot solve in the forums.
    Best greetings from Germany
    Olaf
    Tuesday, January 19, 2010 12:22 PM
    Moderator
  • Thanks Olaf, I'll go to Connect.

    Andy
    Tuesday, January 19, 2010 1:52 PM