Conference smurf::ase

1917.0. "Prpblem of rpc.lockd and lockf() " by TKTV20::MATUZAKI () Wed Mar 05 1997 02:01

	I have a question about rpc.lockd on ASE. ( Digital UNIX ) 
	My customer uses ASE that the user's mail-spool ( /usr/spool/mail ) is 
	in the common disk. Client are nfs-mount of this disk. Client type is 
	HP, DEC, SUN, News(SONY)... But only HP host can not include the mail. 
        My customer created a simple program on HP, this program waited 
        at call lockf().

	It is OK if lockf() send local IP address from HP , but it is NG if 
        lockf() send common IP addres( alias address ).

	OK case.

	-------------      request (1.1.1.1)    ------------
	| DEC       |    <------------          | HP       |
	| rpc.lockd |      reply   (1.1.1.1)    | lockf()  | 
	-------------    ------------>          ------------
        local address  (1.1.1.1)
	common address (1.1.1.2)               
	
	NG case. 	

	-------------      request (1.1.1.2)    ------------
	| DEC       |    <------------          | HP       |
	| rpc.lockd |                           | lockf()  |  program is wait!! 
	-------------    ----X                  ------------
        local address  (1.1.1.1)
	common address (1.1.1.2) 

	I think that rpc.lockd replies packet of local IP address, no matter 
	rpc.lockd receives packet of common IP address form HP.
        If that is so, will this be the correct method for the rpc.lockd ?

	Thank in advance for your help.

							DEC-J/MCS/matsuzaki