[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference pamsrc::decmessageq

Title:	NAS Message Queuing Bus
Notice:	KITS/DOC, see 4.*; Entering QARs, see 9.1; Register in 10
Moderator:	PAMSRC::MARCUSEN

Created:	Wed Feb 27 1991
Last Modified:	Fri Jun 06 1997
Last Successful Update:	Fri Jun 06 1997
Number of topics:	2898
Total number of notes:	12363

2760.0. "can a process survive to DMQ shutdown ?" by ROM01::OLD_BOCCANER () Thu Feb 06 1997 08:55

    Hi,I'm developing an application for a customer in rome, on
    OpenVms platform.
    My server is receiving info from outside source and send this message
    into a brodacst system where PC client can receive them.
    Now on openvms platform, when DMQ servers go down, for a shutdown
    requested or for some other software problem, server dies.
    
    I always assume that is right if application is DMQ centered, but
    customer wants server does'nt die if DMQ bus is'nt available after a
    succesful attach.
    
    I suppose that in OPenVms implementation a rundown routine is
    associated to process program, when DMQ dies, every process on the bus
    die. Is this model the only one or  it is possible process survive to
    DMQ shutdown ? 
    
    
    regards 
    stefano

T.R	Title	User	Personal Name	Date	Lines
2760.1		XHOST::SJZ	Rocking the Messaging Desktop !	`Thu Feb 06 1997 10:30`	10
	On OpenVMS, if you are attached to the group when DECmessageQ goes down, then you go down. On all other platforms, you should receive a status of PAMS__FATAL for any call that is in progress and then an error of PAMS__PAMSDOWN for all subsequent calls. At that point you can do a pams_exit() (which always succeeds) and then put yourself into a loop trying to reattach to the group when it comes back up. _sjz.
2760.2	Workaround, New version Or ?....	ROM01::OLD_BOCCANER		`Tue Feb 11 1997 12:36`	3
	No workaround on openvms ? new version will remove this constarint ? regards stefano
2760.3		XHOST::SJZ	Rocking the Messaging Desktop !	`Tue Feb 11 1997 13:30`	7
	the VMS people will have to answer the workaround part. But there are no current plans to remove this constraint (which is non-trivial to say the least). _sjz.
2760.4	It's supposed to work...	KLOVIA::MICHELSEN	DECmessageQ Engineering	`Tue Feb 11 1997 14:25`	11
	...however it is not part of the standard regression tests. You need to define in a place where the COM Server can see it the logical DMQ$DISABLE_FORCEX to "YES". This causes the COM Server to not send the AST to all of the attached processes to exit. Therefore, the processes are not told that anything has happened to the group and you will need to manually shutdown the DmQ processes. The logical is also translated at the process level but has been reported not to work correctly. Marty