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

Conference azur::mcc

Title:DECmcc user notes file. Does not replace IPMT.
Notice:Use IPMT for problems. Newsletter location in note 6187
Moderator:TAEC::BEROUD
Created:Mon Aug 21 1989
Last Modified:Wed Jun 04 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:6497
Total number of notes:27359

2524.0. "system crash after having used mcc_notification_fm" by TAEC::WEBER () Tue Mar 10 1992 15:02

    
    Just after re-installation solved their problem,
    my customer tried to process notifications and end up with a system crash.
    This happens in the following conditions:

    DECmcc Iconic map is started up.
    They test notifications by sending about 20 messages/during 4 seconds
    from an analyser, messages size is about 100 bytes.
    The messages are received by the event sink, which does not contain
    mcc_lock-XXX routines, but they use lex and yacc to analyse the messages.
    The events are put on the event manager (mcc_event_put), and treated in a
    AM (mcc_event_get).
    Notifications are handled properly throughout the sequence.
    After that, they stop DECmcc (by the mcc_kill command)
    and restart the iconic map. At this time, the kind of operations they
    use are: create entities in domains, delete them from domain.
    Then, they pull down menu "Application", item "FCL".
    The system hangs up and the following message is generated in the log 
	"Reboot after panic: MUNLOCK: dup page unlock"
    No core is generated.
    The crash is unpredicatible, i.e. this can happen 5 minutes after
    having tested the notification as well as 25 minutes later, but this
    happens anyway, even if DECmcc processes have been stopped meanwhile.
    
    They find it severe, since other users are blocked in their
    work on other applications.
    Any idea to investigate ? (code review, dump generation)

    Version : DECmcc T1.2.4.
    They use DNS and the DECstation runs other applications.
   
    Thanks to answer
    Florence

T.RTitleUserPersonal
Name
DateLines
2524.1Crashes are priority 1TOOK::MINTZErik Mintz, DECmcc Development, dtn 226-5033Tue Mar 10 1992 18:367
This is clearly a serious problem.

We will start to investigate, but could you please file a priority 1
QAR from the customer's QAR account (or let me know which customer).

-- Erik

2524.2Is this resolved?GIDDAY::DRANSFIELDMike Dransfield, Sydney RSSGFri Oct 02 1992 21:019
    Was there any resolution to this system crash?
    I have just seen exactly the same crash on T1.2-4.
    
    I will try and get the mcc up to V1.2, but would like to know if there
    was a problem with ULTRIX which need a patch or parameter change.
    Version of ULTRIX is V4.2A.
    
    Thanks,
    Mike
2524.3There is one patchTOOK::MINTZLKG2-2 near pole X3, cube 6072, dtn 226-5033Fri Oct 02 1992 22:576
I don't know the resolution of this particular note.
However, there is a patch to ULTRIX that is shipped with the V1.2
kit, to correct a possible system crash.  There were no other
known system crashes when V1.2 shipped.

-- Erik