Posted by Neil Badger on January 07, 2004 at 02:30:03:
Way back on the 6th of December,Saturday morning, one of my customer running w2k3 & comet 2002 with the fileserver, rang me to say that none of the pc's could find the server, not surprising as it was down, I had them reboot and all seemed fine until today when one of the accounts payable clerks came to us with a wad of invoices for which we had no purchase orders. On investigation I discovered they had all been transferred by Cometanywhere during the evening prior to the server going down, Cometanywhere calls the six branches and merrily uploads and downloads files throughout the evening. The branches work six and seven days a week so come Monday there was still some data transferred so it wasn't immediatley apparent that something had gone wrong. We'll blame the approach of Christmas for the lack of balancing controls. However , even if they had told me on the Monday that something was missing how and where did it go.
The server didn't go down physically till about 8am Saturday and the transfers start about 8 o'clock at night, rather a long delayed write. Do one of those meaningless logs in W2K3 give you clues to this ?, is there some file verification that should be run after the server crashes ? Is there some problem between Cometserv32 & w2k3 if it crashes ?
The other thing is that I can easily identify the remote stuff as it is processed in batch mode, but whatabout stuff done locally ? If 8 hours processing is missing why not 12?
It all sounds too difficult, is it worth chasing up or do we tighten our manual controls, and do even more backups and hope for the best?
Badge