Conference csc32::consolemanager

    You can customize the event list to only show particular events. You
    first must get teh event list running and then, under the 'Options'
    pulldown menu, you will see a menu entry to 'Filter'. Follow this path
    to define th event syou want to see. You can save the filter. 
    Of course, every time you start the event list, it will use the 
    same filter. Optionally it would be nice to specify a filter name 
    when starting the Event LIst.
    
    Dave 

    To add to what Dave said, you *cannot* associate once of the filters
    in the config database with an *interactive* eventlist. 
    
    Why do you want to do this?
    
    Regs,
      Dan

.2
>>>    To add to what Dave said, you *cannot* associate once of the filters
>>>    in the config database with an *interactive* eventlist. 
>>>    
>>>    Why do you want to do this?
>>>    
>>>    Regs,
>>>      Dan

	A "PCM customer" explains to me his problem. 
He use some filters like this

PCM Edit> sho filter ping
Full or Brief listing? [f=full, b=brief] (b): f
Output file (SYS$OUTPUT):
Filter Name      = ping
    Information =
    Interval = 0 seconds
    Event Interval = 0 seconds
    Filter 1 events
      ping
    Filter 1 classes
      ping
    Filter 1 systems
      NODEA    SYSTEM
    Filter 1 SubSystems
      Vaxcluster console
    Filter all groups (0)
    Notify regardless of event priority
    Actions: (1)
      Action 1: multi-line window
        Information :
        Schedule:
            Run at any hour on any day
        User data passed to action:
            node1 0 0 decnet
            node2 0 0 decnet
            node3 0 0 decnet
            node4 0 0 decnet
	    node5 0 0 decnet
	     .		
	     .
	15 node are used as "dispatch Userdata".

When one of this node is "shutdown/reboot" or "decnet problem" , customer says
that no "multi-line window" are regenerated with the good event (ex: ping ) on
this node.
He says also that nothing appears in the logfiles (CONSOLE$ENS_DAEMON or
multilinelog).
For him , the only way to get this "multi-line" is to restart PCM.
Of course , this problem is intermittent and the customer doesn't want to 
restart PCM.He test some other ways like "$console eventlist "

Actually , PCM is running in debug mode ( "all") and i waiting from customer
the logfiles and the CONSOLE_CFG.DAT.
I will see the logfiles and contact you throught this notes.



Regards LAURENT 

    >When one of this node is "shutdown/reboot" or "decnet problem" ,
    >customer says that no "multi-line window" are regenerated with the good 
    >event (ex: ping ) on this node. He says also that nothing appears in the
    >logfiles (CONSOLE$ENS_DAEMON or multilinelog).
    >For him , the only way to get this "multi-line" is to restart PCM.
    >Of course , this problem is intermittent and the customer doesn't want
    >to restart PCM.He test some other ways like "$console eventlist "
    
    Then, of course, this is the real problem. he's already go it setup the
    correct way. Now if we are having a "decnet problem" we can't do
    anything in PCM to fix that. The debug logs will tell us a lot as there
    will be an audit trail when every event occurs that will tell us which
    actions the event is being passed too and if a new action process needs
    to be started.
    
    Regs,
      Dan
    

	Finally , i got informations about the customer' problem.

With a event "X" occurs, a "Multi_line" is created on node MTI001 (it 's work )
This node MTI001 is "shutdown/reboot".Normally the "Multi_Line" process is 
terminated.   Here a example with Multi-Line_006.log
-------------------------------------------------------------------------------
.....
$ Set NoOn
$ Define CONSOLE$ACTIONNAME "Multi-Line Manager"
$ Define CONSOLE$EVENTPORTNAME "Multi-Line__006"
$ CMACTION :== $CONSOLE$ACTIONS:manager$eventlist
$ CMACTION  mti001 0 0 decnet
POLYCENTER Console Manager
Eventlist Version V1.6-100
Copyright (c) 1995 Digital Equipment Corporation. All Rights Reserved

Display will be sent to :
    Node      : mti001
    Server    : 0
    Screen    : 0
    Transport : decnet
CM Eventlist: Getting SYSTEM preferences from DECW$SYSTEM_DEFAULTS:CONSOLE$EVENT
LIST.DAT
CM Eventlist: Getting USER preferences from CONSOLE$APP_DEFAULTS:Multi-Line_Mana
ger.mti001_0_0
CM Eventlist: Setting resource Indeterminate
              Value Light Sky Blue
.......
CMUserGetMainloopEvent: Enter/Leave
CM Eventlist: Event received <CMgr Disk Warning>
XIO:  fatal IO error 65535  on X server "_WSA4064:"
      after 5118 requests (5116 known processed) with 0 events remaining.
%XLIB-F-IOERROR, xlib io error
-SYSTEM-F-PATHLOST, path to network partner node lost
$ EXIT $STATUS
  SYSTEM       job terminated at 13-SEP-1995 15:40:57.39

  Accounting information:
  Buffered I/O count:            6920         Peak working set size:    6698
  Direct I/O count:               174         Peak page file size:     11395
  Page faults:                  12726         Mounted volumes:             0
  Charged CPU time:           0 00:00:09.93   Elapsed time:     0 00:13:30.83
-------------------------------------------------------------------------------
	Now this node MTI001 is up and the same event occurs.Normally a new 
"Multi-Line_xxx" process should be created for  MTI001 display.At this site
no new "Multi-Line_xxx" are created for "MTI001 display"
Here a extract of CONSOLE$ENS_DAEMON.LOG
-------------------------------------------------------------------------------
CONSOLE$ENS_DAEMON: Event received from Source - CMgr Disk Warning
      Details are:
      S.No.     : 12872, 2 (843579394)
CMShm: Searching for system : NETSYS
CMShm: Increment system event priority count
CMShm: Increment system event priority count
      Type      : 1
      System    : NETSYS
      SubSystem : CM Internal
      Source    : Console Manager
      Class     : CONSOLEmanager
      Info      : Log file path CONSOLE$LOGFILES: is short of disk space
>>>Sending to Action   : Multi-Line General, Userdata : mti001 0 0 decnet<<<
   Sending to Action   : Multi-Line General, Userdata : sys001 0 0 decnet
   Sending to Action   : Multi-Line General, Userdata : sys004 0 0 decnet
......

When CONSOLE$ENS_DAEMON says :
"Sending to Action   : Multi-Line General, Userdata : mti001 0 0 decnet"
it's because a "Multi-Line General" action exists on "mti001 0 0 decnet" , but
no action exists. The last action was "Multi-line-006" and this process is 
"logout"
----
%XLIB-F-IOERROR, xlib io error
-SYSTEM-F-PATHLOST, path to network partner node lost
$ EXIT $STATUS
  SYSTEM       job terminated at 13-SEP-1995 15:40:57.39
----

Is it a know problem ?
Is it IPMT time ?

PS: I don't post in this notefiles all the logfiles (with debug).It should be
to long .You can access to PRSPSU::euriware.bck  (47.109)


Laurent

    This still looks like a network problem. XWindows is loosing the
    connection to the remote system. What should happen is that the next
    time an event happens, ENS should see that it lost the connection and
    then try again to put the multiline window back. How often does this
    happen? If this happens frequently, con you try to run some other 
    DECwindows application between the 2 hosts and see what happens?
    
    
    Dave

.6
>>    This still looks like a network problem. XWindows is loosing the
>>    connection to the remote system. What should happen is that the next
>>    time an event happens, ENS should see that it lost the connection and
>>    then try again to put the multiline window back. How often does this
>>    happen? If this happens frequently, con you try to run some other 
>>    DECwindows application between the 2 hosts and see what happens?

	Dave ,
I quite agree with you , the problem 's source is probably a network problem.
Yes , XWindows is loosing the connection to the remote system because the
logfiles says :
-XIO:  fatal IO error 65535  on X server "_WSA4064:"
-      after 5118 requests (5116 known processed) with 0 events remaining.
-%XLIB-F-IOERROR, xlib io error
--SYSTEM-F-PATHLOST, path to network partner node lost
-$ EXIT $STATUS
-  SYSTEM       job terminated at 13-SEP-1995 15:40:57.39

>>ENS should see that it lost the connection and then try again to put the
>>multiline window back.
I agree with you but why ENS says 
>>>Sending to Action   : Multi-Line General, Userdata : mti001 0 0 decnet<<<
It ' look like an action existing ( there is not).

Yes , this happens frequently , each time a node (that receive a Multi-Line) is
"shutdown/reboot" or when there is some "saturation" on the Network.
Yes , other Decwindows applications work ( Decterm , xclock ...)

PS : You can find all the logfiles (multi-line*, console*) in debug mode and 
the Console_cfg.dat in PRSPUB::euriware.bck



Regards Laurent
 

    The only other alternatice I can think of to try would be this:
    Only have 1 dispatch per action definition and not the 15 like you have.
    Maybe there are too many X events coming to the one process.
    You would then have 15 filter definitions with 1 dispatch (multi-line
    window).
    
Filter Name      = ping_node1
    Information =
    Interval = 0 seconds
    Event Interval = 0 seconds
    Filter 1 events
      ping
    Filter 1 classes
      ping
    Filter 1 systems
      NODEA    SYSTEM
    Filter 1 SubSystems
      Vaxcluster console
    Filter all groups (0)
    Notify regardless of event priority
    Actions: (1)
      Action 1: multi-line window
        Information :
        Schedule:
            Run at any hour on any day
        User data passed to action:
            node1 0 0 decnet
    
Filter Name      = ping_node2
    Information =
    Interval = 0 seconds
    Event Interval = 0 seconds
    Filter 1 events
      ping
    Filter 1 classes
      ping
    Filter 1 systems
      NODEA    SYSTEM
    Filter 1 SubSystems
      Vaxcluster console
    Filter all groups (0)
    Notify regardless of event priority
    Actions: (1)
      Action 1: multi-line window
        Information :
        Schedule:
            Run at any hour on any day
        User data passed to action:
            node2 0 0 decnet
    

	Thanks for your help Dave, 
This customer will try as soon ( next week?) to use only one Userdata per filter.
See you next week .


Laurent

	Customer give me a feedback today .

After a new setting  ( Only have 1 dispatch per action definition ) , he get 
the same problem 
I can't reproduce this problem on my site.

Is there any ideas to continue the troubleshooting of this problem ?
Is it a know problem on other customer site ?


Thanks in advance 
Laurent