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

Conference 7.286::fddi

Title:	FDDI - The Next Generation

Moderator:	NETCAD::STEFANI

Created:	Thu Apr 27 1989
Last Modified:	Thu Jun 05 1997
Last Successful Update:	Fri Jun 06 1997
Number of topics:	2259
Total number of notes:	8590

671.0. "DEMFA FAULTING?" by ICS::BARRY () Tue Aug 11 1992 11:50

We've recently replaced a number of older CPU's in our datacenter
with 6610's. We had them configured with DEMFA's for FDDI use down
the road. In the mean time we are using DEMNA's. However, we are 
seeing a problem on all the 6610's, 9 of them, that appears to be
related to the DEMFA's. Approximately every 5 minutes the systems
will hang and the DEMFA will re-init itself. The lights on the DEFMA
go off and the fault light on the console will come on. The error
count on the FXA0 device increments and as mentioned in an earlier
note nothing is written to the errorlog.
 
The DEFMA's are terminated and their state within NCP is OFF. Can
anyone help us understand what is causing the DEMFA's to re-init
every 5 minutes? Should we pull the boards until such time that
we hook up the FDDI ring? Should we be terminating the boards?
Could there be a cycling of the self test causing this?

We are running SSB V5.5-1. If there is any other info that would help
please let us know.

Thanks for you time,

Jim Crouch
Chris Barry
NEI-Maynard

T.R	Title	User	Personal Name	Date	Lines
671.1	In a VAXcluster?	CVG::TOMASWICK	Joe - VAXcluster Validation/aka CVG	`Tue Aug 11 1992 13:02`	11
	If you are running in a VAXcluster over the Ethernet SCA is also trying to use the DEMFA. The driver will keep trying to init the DEMFA every few minutes. You have a couple of choices. Put a turnaround in the DEMFA MIC connector, connect it to a concentrator, or pull the DEMFA out of the system. If the system is not in a cluster and no other protocols are running on it, I don't know what else it could be. Joe
671.2		ICS::CROUCH	Subterranean Dharma Bum	`Wed Aug 12 1992 10:05`	6
	After discussing this with Colorado we decided to have FS pull the boards until such time that they will actually be used. We did this last night and of course the hangs have disappeared. Jim C.
671.3		KONING::KONING	Paul Koning, A-13683	`Wed Aug 12 1992 11:36`	3
	Did you QAR this problem? paul
671.4	Life (as usual) with FDDI	STAR::GAGNE	David Gagne - VMS/LAN Development	`Fri Aug 14 1992 17:24`	12
	Because it causes a disruption to join a ring and because it takes time to join a ring, the VMS philosophy for FDDI is to bring the FDDI device up and leave it up. So even though there are no users of the FDDI, the driver is still trying to have it join the ring. The algorithm is to eventually get to a point where the attempts are made every 5 minutes. So what you are seeing every 5 minutes is the FDDI driver trying to make the device join the ring. You can pull the device, put a loopback on it, or plug it into a ring to have the temporary hangs go away.
671.5	Error log	STAR::GAGNE	David Gagne - VMS/LAN Development	`Fri Aug 14 1992 17:27`	4
	Also, there is no error log because the driver can tell that the wire is not plugged in and we feel that this is not an error - that is, it is not a problem for which you would replace the DEMFA. Remember that it's an ERROR log; not an EVENT log.
671.6		KONING::KONING	Paul Koning, A-13683	`Fri Aug 14 1992 17:47`	8
	Well, we've argued about whether it makes sense to restart every n minutes rather than just sitting there, but that wasn't the point. If I understood the original comment correctly, the system hangs for a noticeable amount of time when the restart is done. If that's actually what happens, that's a clear bug: the driver should do its restarting without disturbing the rest of the system... paul
671.7		MSBCS::KALKUNTE	Ram Kalkunte 293-5139	`Sat Aug 15 1992 00:36`	28
	Re .0 First of all, make sure that you have the DEMFA in the right slot. All advertised behavior of the DEMFA is restricted for slots on the XMI backplane that will support it. Read the installation manual to find out what slots are supported. Please read my comments below and supply more detail. Re .5 Dave, I see that the base note mentions that the DEMFA has been terminated. Assuming that the loopback connector is plugged in correctly and the DEMFA is in the right XMI slot, I would consider this a bug. Otherwise, all bets are off. Re .6 Paul, I don't agree with your interpretation of the base note. The author never mentions of any "noticeable ammount of time the system hangs" during restart. Also, the duration of the initialization sequence and hence the length of time the "system hangs" depends on the CPU. What is the duration of time which you call noticeable ? Of course you know that the fault indicator on the VAX 6000 is an effect of initialization. We need to know more before we can conclude if or not this is a bug. Ram
671.8		KONING::KONING	Paul Koning, A-13683	`Mon Aug 17 1992 12:08`	5
	"noticeable" to me means that the users can tell and complain about it. .0 said that "the system hangs every 5 minutes"; I assume he wasn't talking about milliseconds. paul
671.9	borads need reset time,.. is that a "bug" in VMS?	STAR::SALKEWICZ	It missed... therefore, I am	`Tue Sep 01 1992 15:48`	17
	The system should not hang at all when the DEMFA resets... Yes,.. the fault light comes on temporarily while the DEMFA is reinitializing,.. which takes 7-8 seconds,.. and during that short time there may be a noticable delay. As stated, the possible solutions are: 1) install a loopback (which must be different than whatever .0 is calling a "terminator" because a loopback fixes this) 2) Connect to a net 3) remove the boards until you are ready to do 1 or 2 4) live with it as is (How bad is it?) /Bill