locked
Server loses connection with drives being backed up RRS feed

  • Question

  • Getting a complete initial backup done is turning out to be a huge, and so far unsolveable problem.

    I have eleven hard drives attached to this console, two enternal,(an ata and a sata), and nine externals, (three firewire and six usb).  These include a 200 gb ide as the system drive, one 160, six 500 gb, one 750 gb, and two 1 tb.

    All of these drives show up in all situations, read, write, across the four computer network, etc, EXCEPT in WHS backups.  Backup gives the indication that some drive backups could not be done because "connection between the computer and server was lost", although the connection is Not lost, and the drives are behaving normally.

    I've turned off all other processes and dedicated the machine to feeding data to the server,  but the result is invariably 'incomplete'.  The drives that fail to backup have no obvious similarity.  Two of the firewire attached drives backup; the third does not.  The USB attached drives, without regard to capacity or whether the drive within the enclosure is pata or sata fail three of the total six.

    The only way I can accomplish a completed backup is to tick off the drives that have failed, which, although the server is satisfied that it has done a full backup, leaves four drives out of the picture.

    Since these backups fail because of what the server claims is lost connection to the console computer, and since these same drives remain connected to all other machines, including uploads and downloads, what is it about the server that makes it unable to make and maintain a connection, and is there some method of forcing it to do so?  I doubt my situation is unique, although I've combed the forums without seeing it addressed.

    The server is on a computer with an AMD Athlon 64 X2 4200 sitting on an ASUS mb, with a gig of ram.  It has a raid set of five 1 tb Samsung drives.  The console is an Intel Q6600 on a BFG iVidia 760 mb with 8 gb of ram.  The external drives are enclosed in BYTECC, Antec, and Ultra's, all of which are in good condition, with one tb drive in a CoolDrive caddy.

    Any help with this problem would be of great service, as the greatest attraction for building a server was its backup feature, and if it can't be made to operate correctly and fully, I may decide to eliminate it and opt for a pure raid array.

    Saturday, May 2, 2009 5:55 PM

All replies

  • Getting a complete initial backup done is turning out to be a huge, and so far unsolveable problem.

    I have eleven hard drives attached to this console, two enternal,(an ata and a sata), and nine externals, (three firewire and six usb).  These include a 200 gb ide as the system drive, one 160, six 500 gb, one 750 gb, and two 1 tb.

    All of these drives show up in all situations, read, write, across the four computer network, etc, EXCEPT in WHS backups.  Backup gives the indication that some drive backups could not be done because "connection between the computer and server was lost", although the connection is Not lost, and the drives are behaving normally.

    I've turned off all other processes and dedicated the machine to feeding data to the server,  but the result is invariably 'incomplete'.  The drives that fail to backup have no obvious similarity.  Two of the firewire attached drives backup; the third does not.  The USB attached drives, without regard to capacity or whether the drive within the enclosure is pata or sata fail three of the total six.

    Have you tried rotating them around?  Perhaps you're drawing too much power with all those drives as well.  In other words, if you disconnect the two FireWire drives that do normally get backed up, does the only FireWire drive left then get backed up?

     

    The only way I can accomplish a completed backup is to tick off the drives that have failed, which, although the server is satisfied that it has done a full backup, leaves four drives out of the picture.

    Since these backups fail because of what the server claims is lost connection to the console computer, and since these same drives remain connected to all other machines, including uploads and downloads, what is it about the server that makes it unable to make and maintain a connection, and is there some method of forcing it to do so?  I doubt my situation is unique, although I've combed the forums without seeing it addressed.

    The server is on a computer with an AMD Athlon 64 X2 4200 sitting on an ASUS mb, with a gig of ram.  It has a raid set of five 1 tb Samsung drives.
    FYI, RAID is unsupported on WHS.

    The console is an Intel Q6600 on a BFG iVidia 760 mb with 8 gb of ram.  The external drives are enclosed in BYTECC, Antec, and Ultra's, all of which are in good condition, with one tb drive in a CoolDrive caddy.

    Any help with this problem would be of great service, as the greatest attraction for building a server was its backup feature, and if it can't be made to operate correctly and fully, I may decide to eliminate it and opt for a pure raid array.


    Have you run chkdsk /r on each of the drives on your client (that you can't backup)?
    Sunday, May 3, 2009 5:50 PM
    Moderator
  • Your server and client architectures aren't laid out very clearly. It's not obvious to me where exactly all those drives are installed; in one place you say they're connected to the console and talk about them as though they're part of the storage pool, in another you talk about backing them up to the server, and in another you talk about a RAID array on the server.

    In any case, an error indicating that the connection between the client and the server was lost can have many causes. I would try to eliminate disk errors first, by running chkdsk /r on all drives that are involved, in both your server and your client PCs.

    I'm not on the WHS team, I just post a lot. :)
    Tuesday, May 5, 2009 3:21 PM
    Moderator