pung@ornl-ncc.ARPA.UUCP (03/30/87)
>>The fix for the 8600 memory bug that is tickled by DEUNA/LAT changes the >>cpu serial number. Field service usually neglects to tell you that your >>serial number is going to change. When you re-boot, some of your third >>party proprietary software may no longer work...... >VAXs have no software readable cpu serial number. The do have a system ID >number. This number is for DEC use in determining ECO and FCO status. Any >FCO will change the system id. That's why it's there. On some processor >type a lot of CPU have the same SID. And I don't believe all of the new CPUs >even have them. They are some times even settable by DIP switches. >I wish third party software vendors would get to understand this and NOT >use them as keys. It is my policy that is several systems are bid, I reject >any using this silly scheme unless the cost of any alternative is totally >prohibative. The time and effort involved in fixing this when you have >virtually every type ov VAX CPU DEC has ever made is absurd! (Have you ever >had to enter an ISSCO system password? It's several lines of random characters >which have to be exactly right!) You are correct that the CPU SID changes after FCO's are performed, however as I understand it, the rightmost 3 digits of the SID never change irregardless of what hardware changes are made. (CPU upgrades may be an exception) I am not aware of any processor that does not offer something in response to the f$getsyi("sid") command. We have a VAX capacity management utility which collects performance and capacity data from our machines and the data is identified by the 3 rightmost characters of the SID number. This project has only been operating for a short time but we have found no two serial numbers to be the same and have not as yet seen any of the numbers change. Are my assumptions correct or am I heading for failure?? Anybody (DEC ??) know how this works ?? |------------------------------------------------------------------| | Mike Pung | (615) 574-8082 | | Oak Ridge National Lab | Pung%ornl-ncc@ornl-msr.arpa | | Martin Marietta Energy Systems | | |------------------------------------------------------------------| ------
gwalker@SPCA.BBN.COM.UUCP (03/30/87)
In reply to the message from M. S. Pung at ornl-ncc.arpa (my mailer doesn't recognize that host): Every different VAX model has a slightly different format of the SID. All use bits 24-31 for a number identifying the VAX model. (31 is the highest order bit). After that, bits 0-23 are used differently for different models. For the 780, 782, 785, 8600 bits 0-11 are "The serial number assigned to the CPU" and each of those models has 3 or 4 bits assigned to "The number assigned to the Digital Equipment Corporation facility where the CPU was manufactured". The quotes are from my VAX Hardware Handbook which does not mention microvaxen or anything newer than the 8600. For the 750, 725, 730 processors there are no bits allocated to either of a machine serial number or a DEC facility. They have 3 fields each - one for the software revision level of the CPU microcode, one for the revision level of the CPU hardware, and one that isn't used. Not too much opportunity for uniqueness there. I haven't seen the handbook for microvaxen or the 8200, 8300, 8500 series so I don't know how the bits are assigned for those. I'm told that it is fairly likely to have 2 750's or 2 microvaxen with the same SID if they were ordered at the same time and configured the same. Probably what you are doing is safe enough if you are doing it for a limited number of systems. One can't rely on the SID's to be completely unique everywhere though as you can see from the above. -- Gail
Wahl.ES@XEROX.COM.UUCP (03/31/87)
I just had a software vendor ask for the SID on our MicroVax. Doing GETSYS, I got a hex SID of 08000000. Is this just my good fortune or are all MicroVaxen the same? --Lisa