SUMMARY: VME dropped an INT-ACK cycle; is this serious?

From: Joe Ramey (ramey@ra.csc.ti.com)
Date: Thu Nov 05 1992 - 10:54:11 CST


Original question:

    Our 4/690 MP running SunOS 4.1.2 frequently prints these errors to
    /usr/adm/messages when the system is active:

    Oct 12 12:11:39 trdc000 vmunix: VME dropped an INT-ACK cycle
    Oct 12 12:11:39 trdc000 vmunix: MMU sfsr=b36: Bus Access Error on supv data fetch at level 3
    Oct 12 12:11:39 trdc000 vmunix: M-Bus Timeout Error

    What do these errors mean? Is it anything we should worry about? We
    asked our Sun office about them and they said that we should just
    ignore these messages. However sometimes it seems like the system
    runs slow around the same time that these messages come out (for
    example, some NFS clients of this server report `server not
    responding') at about the same time.

I had also mentioned that we were running with this patch (along with
several others):

    100542-03: SunOS 4.1.2: IPI - Galaxy jumbo patch

Thanks to peter@key.amdahl.com (Peter Sivo) for the right answer. We
needed to get a newer revision of the patch: 100542-06 says, among
other things that this fixes Bugid 1075751: Summary: "VME: Dropped
Int-Ack cycle" messages. Too bad our Sun distributor
couldn't/wouldn't tell us that. We have the patch installed now and
we're no longer getting these error messages.

Thanks also to jkays@msc.edu and mdl@cypress.com (J. Matt Landrum) for
their replies.

Joe Ramey
Texas Instruments
ramey@csc.ti.com

 



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:52 CDT