Conference azur::mcc

Could you post the result of the "pstat -s" command please?

-- Erik

According to the person that did the work at my request, this system had just
been reconfigured to supply about 125MB of swap space.

Prior to doing this reconfiguration, we weren't even getting this far.

# pstat -s
122872k swap configured
75460k reserved virtual address space
 66324k used (23392k text, 544k smem)
 56548k free, 1776k wasted, 0k missing
avail: 1761*32k 196*1k
#


Al

Could you issue the shell command "ps -aux | grep mcc" after the notify
command but before the display command?  That will tell us which
module uses id=9 on your system.  Also, since I suspect this may require
some investigation, could you outline the contents of your configuration
(eg how many domains, what kinds of entities registered).

-- Erik

I hope this is what you want.

If it is not, I suspect I may have to issue that "kill" command in the
release notes to get rid of all MCC processes and try it again.  Please
let me know if ths is the case and I'll try it again tomorrow

The error comes back within a few seconds of issuing the NOTIFY command.
I suspect there is an existing process trying to satisfy the previous
notify.

Al

ddcds1.democtr.del.dec.com> ps -aux | grep mcc
decmcc     549  8.5  3.0  648  552 p0 R      0:00 ps -aux
root        35  0.0  0.0   64    0 ?  IW     0:00 /usr/etc/mcc_dna4_evl_monitor
decmcc     131  0.0  1.1 4496  196 ?  I      0:00 /usr/bin/dxsession :0
decmcc     159  0.0  2.4 2932  448 ?  S      3:35 mcc_dna4_am 5 N
decmcc     157  0.0  2.2 2056  400 ?  S      1:02 temip_osi_system_am 17 N
decmcc     156  0.0  1.9 2100  344 ?  S      1:17 mcc_collection_am 4 N
decmcc     155  0.0  1.9 2532  352 ?  S      0:59 mcc_domain_fm 6 N
decmcc     154  0.0  3.4 4608  640 ?  S      1:31 mcc_notification_fm 10 N
decmcc     142  0.0  0.4  376   64 p1 I      0:00 - (csh)
decmcc     140  0.0  2.7 3880  492 ?  S      0:04 /usr/bin/dxterm -ls
decmcc     139  0.0  3.0 3880  560 ?  S      0:00 /usr/bin/dxterm -ls
decmcc     138  0.0  1.4 1620  260 ?  I      0:00 /usr/lib/X11/xconsole
decmcc     141  0.0  0.6  376  108 p0 S      0:00 - (csh)
decmcc     137  0.0  0.0   56    0 ?  IW     0:00 sh -c /usr/lib/X11/xconsole
decmcc     136  0.0  2.7 3280  508 ?  S      0:06 /usr/bin/mwm
root       193  0.0  3.6 2936  664 ?  S      3:36 mcc_dna4_am 5 N
decmcc     361  0.0  1.4 2656  256 ?  S      0:43 mcc_testobj_am 19 N
root       190  0.0  2.8 2100  528 ?  S      1:13 mcc_collection_am 4 N
root       189  0.0  2.8 2092  528 ?  S      0:51 mcc_domain_fm 6 N
root       188  0.0  4.8 4108  900 ?  S      1:39 mcc_notification_fm 10 N
decmcc     515  0.0  0.6  248  104 p1 I      0:00 telnet ddcvs1
decmcc     177  0.0  0.0  376   36 p2 IW     0:00 - (csh)
decmcc     176  0.0  3.3 3896  608 ?  S      0:14 /usr/bin/dxterm -ls
decmcc     174  0.0  1.6 1708  296 ?  S      1:01 /usr/mcc/mmexe/mcc_tcpip_sink
decmcc     167  0.0  1.5 2340  284 ?  S      1:10 temip_event_log_fm 18 N
decmcc     166  0.0  1.5 2540  280 ?  S      1:26 mcc_alarms_fm 2 N
decmcc     165  0.0  1.6 2104  300 ?  S      0:58 mcc_historian_fm 9 N
decmcc     163  0.0  1.0 2152  172 ?  S      1:22 mcc_exporter_fm 8 N
decmcc     162  0.0  1.8 2220  336 ?  S      1:50 mcc_tcpip_am 14 N
decmcc     550  0.0  0.2   40   32 p0 S      0:00 grep mcc
decmcc     363  0.0  2.9 2108  544 ?  S      0:44 mcc_circuit_am 3 N
decmcc     364  0.0  1.5 2168  272 ?  S      0:41 mcc_registration_fm 12 N
root       199  0.0  1.3 2512  236 ?  S      1:52 mcc_alarms_fm 2 N
root       197  0.0  2.4 2108  444 ?  S      1:03 mcc_historian_fm 9 N
root       196  0.0  1.0 2152  176 ?  S      1:22 mcc_exporter_fm 8 N
root       195  0.0  2.9 2256  548 ?  S      1:31 mcc_tcpip_am 14 N
ddcds1.democtr.del.dec.com>

Yes, that is what I was looking for.

You realize that you have DECmcc running under 2 different user IDs? Multiple
users running from the same account will share processes, but different UIDs get
different processes. The result in your case is over 20 DECmcc processes.  It is
extremely unlikely that you will get that many processes running on the size
machine you have.  From the information you have provided, my guess
is that both the historian FM (id=9) and the notification FM ran out of
virtual memory.  I believe the installation guide calls for 100 Mb
swap space PER UID, and that assumes the basic BMS MMs, not including
the TeMIP MMs.  

We are aware of the fact that the messages resulting from insufficient
resources are less than helpfull, and in fact I'll check to make sure
that is really the problem that you are seeing.  But I think you'll be
better off trying one user at a time (use "mcc_kill ALL" from root to
get rid of all the processes if you want to re-start).

-- Erik

Just to add my 2 � : the NOTIF FM is only telling you that it encoutered
an exception, it does not tell you that your notification is over. What 
happens here is that the historian FM has aborted, and so you will not 
receive notification from it. Anyway, notification from other AM should
still be Ok.

Regards.

Pierre.

thanks very much....

I'll use the kill command and try to start with a single user for DECmcc,
however, I'm still a bit confused as to how the root account came into
the picture here.

I don't yet follow just how these various shell scripts are being invoked
and who is starting processes, but this is something that I will have to
learn from the local ULTRIX person.

I'll try to clear the system and start again and will report any further
problems with this issue in this note.

Al

Ok....I logged in as root and issued the mcc_kill ALL command afterwhich
I did another ps....

# ps -aux | grep mcc
root        35  0.0  0.0   64    0 ?  IW     0:00 /usr/etc/mcc_dna4_evl_monitor
root      1351  0.0  0.2   40   32 p0 S      0:00 grep mcc
decmcc     174  0.0  2.6 1708  476 ?  S      4:33 /usr/mcc/mmexe/mcc_tcpip_sink


I then went into line mode mcc and attempted to startup the notification
request again and w/o doing anything more received the error....

# manage
DECmcc (V1.2.1)

MCC> notify domain dwo event = (any configuration event)
%MCC-S-NOTIFSTART, Notify request 1 started
MCC> Exception in Mgt Module (id=9), CMA code = 177db005

%MCC-E-NOTIFEXCP, notify request 1 encountered an exception
Unexpected condition returned to Notification FM.
                      MCC Routine Error = %MCC-E-MMABORT, target management
                                          module thread has aborted


MCC>

Here is what is currently on the system immediately after the above command....

# ps -aux | grep mcc
root        35  0.0  0.0   64    0 ?  IW     0:00 /usr/etc/mcc_dna4_evl_monitor
root      1360  0.0  4.6 2100  872 ?  S      0:00 mcc_collection_am 4 N
root      1359  0.0  2.4 2092  448 ?  S      0:00 mcc_domain_fm 6 N
root      1358  0.0 12.4 3984 2340 ?  S      0:02 mcc_notification_fm 10 N
root      1372  0.0  0.2   40   32 p0 S      0:00 grep mcc
decmcc     174  0.0  2.6 1708  476 ?  S      4:35 /usr/mcc/mmexe/mcc_tcpip_sink
root      1369  0.0  7.7 2512 1452 ?  S      0:00 mcc_alarms_fm 2 N
root      1368  0.0  6.7 2104 1260 ?  S      0:00 mcc_historian_fm 9 N
root      1366  0.0  8.0 2220 1508 ?  S      0:00 mcc_tcpip_am 14 N
root      1367  0.0  6.5 2152 1228 ?  S      0:00 mcc_exporter_fm 8 N
root      1362  0.0  8.1 2920 1528 ?  S      0:01 mcc_dna4_am 5 N
#

Perhaps this is significant.....when we run the shell script to setup the
sink monitor, the first and last commands always fail with the same error.
I excuted the script by hand and listed the results below.....
Please note that I did substitute the real password for the "<>" in all cases.

# manage
DECmcc (V1.2.1)

MCC> DISABLE node4 DDCDS1 local sink MONITOR, by user root, by password <>

Node4 45.216 Local Sink Monitor
AT 1992-09-02-12:40:41.195Iinf

Internal error in DECnet Phase IV AM.
MCC> DELETE node4 DDCDS1 local sink MONITOR, by user root, by password <>

Node4 45.216 Local Sink
AT 1992-09-02-12:40:56.406Iinf

Deletion completed successfully.
MCC> CREATE node4 DDCDS1 local sink MONITOR, by user root, by password <>

Node4 45.216 Local Sink Monitor
AT 1992-09-02-12:41:04.898Iinf

Attempt to create duplicate entity rejected.
MCC> SET node4 DDCDS1 local sink MONITOR name /usr/etc/mcc_dna4_evl_monitor, -
_MCC>         by user root, by password <>
  
Node4 45.216 Local Sink
AT 1992-09-02-12:41:22.008Iinf Characteristics

Modifications completed successfully
                                   Name = "/usr/etc/mcc_dna4_evl_monitor"
MCC> DELETE node4 DDCDS1 outbound stream DDCDS1 remote sink MONITOR, -
_MCC>         by user root, by password <>

Node4 45.216 Outbound Stream DDCDS1 Remote Sink Monitor
AT 1992-09-02-12:41:44.465Iinf

Deletion completed successfully.
MCC> CREATE node4 DDCDS1 outbound stream DDCDS1 remote sink MONITOR class 0,-
_MCC>         event type = {0,6}, by user root, by password <>

Node4 45.216 Outbound Stream DDCDS1 Remote Sink Monitor
AT 1992-09-02-12:42:00.547Iinf

Creation completed successfully.
MCC> PASS node4 DDCDS1 outbound stream DDCDS1 remote sink MONITOR class = 2, -
_MCC>         event type = {0,1}, by user root, by password <>

Node4 45.216 Outbound Stream DDCDS1 Remote Sink Monitor
AT 1992-09-02-12:42:09.437Iinf

Pass completed successfully.
MCC> PASS node4 DDCDS1 outbound stream DDCDS1 remote sink MONITOR class = 3, -
_MCC>         event type = {2}, by user root, by password <>

Node4 45.216 Outbound Stream DDCDS1 Remote Sink Monitor
AT 1992-09-02-12:42:17.492Iinf

Pass completed successfully.
MCC> PASS node4 DDCDS1 outbound stream DDCDS1 remote sink MONITOR class = 4,-
_MCC>    event type = {3,4,6,8,9,10,11,12,13,14,15,18}, by user root,
 by password <>

Node4 45.216 Outbound Stream DDCDS1 Remote Sink Monitor
AT 1992-09-02-12:42:31.934Iinf

Pass completed successfully.
MCC> ENABLE node4 DDCDS1 local sink monitor, by user root, by password <>

Node4 45.216 Local Sink Monitor
AT 1992-09-02-12:42:46.020Iinf

Internal error in DECnet Phase IV AM.
MCC>  exit
#


An aside question.....I've noticed that in the original script, event classes
5,6 and 7 are commented out.  I believe we received the same error when we
tried to execute those lines, so I left them commented.  Does ULTRIX DECnet not
support these classes?


I've included some ncp information just incase.....

Question...under VMS several more MCC related objects show up in NCP....is
this not the case for DECnet ULTRIX?


# ncp
ncp>sho  known nodes

Known Node Volatile Summary as of Wed Sep  2 08:46:06 EDT 1992

     Node         State      Active  Delay   Circuit     Next Node
                             Links

45.216 (DDCDS1)   On
Identification                     = DECstation 5000/240
45.15 (DWODR1)    Reachable    0      1      SVA-0        45.15 (DWODR1)
45.185 (DDCVAX)
45.186 (DDCVS1)                1      1                   45.15 (DWODR1)
45.215 (DDCVS2)
45.224 (DDCDS2)
45.235 (DDCPC1)
45.245 (DDCPC2)

ncp>sho known obj

Known Object Volatile Summary as of Wed Sep  2 08:47:28 EDT 1992

Object          Number         File
  
TELL              0            /usr/bin/tell
DEFAULT           0
fal               17           /usr/etc/fal
nml               19           /usr/etc/nml
dterm             23           /usr/etc/dtermd
mir               25           /usr/etc/mir
mail11            27           /usr/etc/mail11dv3
dlogin            42           /usr/etc/dlogind
dtr               63           /usr/etc/dtr

ncp>

Hi Al,

	Reply to 3665.8.

1.  "DISABLE" sink monitor error:

>># ps -aux | grep mcc
>>root        35  0.0  0.0   64    0 ?  IW     0:00 /usr/etc/mcc_dna4_evl_monitor
>>root      1351  0.0  0.2   40   32 p0 S      0:00 grep mcc
>>decmcc     174  0.0  2.6 1708  476 ?  S      4:33 /usr/mcc/mmexe/mcc_tcpip_sink

	The reason you got an error is that /usr/etc/mcc_dna4_evl_process is
not up, while DISABLE command tries to abort it.

	On Ultrix, two processes need to be up before event logging can be 
activated:

	/usr/etc/mcc_dna4_evl_monitor
	
		This program is brought up by DECnet (by setting EXEC state 
		ON and OFF.) But before you do this, you should make sure 
		the system's EVL is switched to the required version.  The 
		details are described in release notes.

	/usr/etc/mcc_dna4_evl_process

		This process is actually created by ENABLE sink monitor
		command.  

	When a DISABLE command is issued, it does two things, abort the
latter process and set monitor state to OFF.  In your experience, the process 
can not be found, thus it returned an error.  


2. how to use script file

>>Perhaps this is significant.....when we run the shell script to setup the
>>sink monitor, the first and last commands always fail with the same error.
>>I excuted the script by hand and listed the results below.....
>>Please note that I did substitute the real password for the "<>" in all cases.

	Before using the script file to start the sink, make sure you
	switch the system EVL software as described in the release notes. 
	(Unfortunately, I noticed there is a sequence error in the release 
	notes that made things more confusing.  We are soon to release a 
	rekitted v1.2 that should clear up this confusion.)

	If you have already switched EVL software, can you start the
	mcc_dna4_evl_process by using ENABLE command at MCC>?   


3.  some event classes are not supported by DECnet-ULTRIX

>>An aside question.....I've noticed that in the original script, event classes
>>5,6 and 7 are commented out.  I believe we received the same error when we
>>tried to execute those lines, so I left them commented.  Does ULTRIX DECnet not
>>support these classes?

	Yes, please refer DECnet-ULTRIX Guide to Network Management Appendix 7
	for the supported events.  Class 1, 5, 6 and 7 are not supported at
	all.  Parts of other class 0, 2,..etc are also excluded from DEcnet-
	ULTRIX.  

	Thank you for your input.

	Jean

Thanks for the reply (.9), but it still appears that I'm having a problem.

As you pointed out, the mcc_dna4_evl_monitor is running, but the
mcc_dna4_evl_process is not.

When I try to issue the command MCC> ENABLE NODE4 0 LOCAL SINK MONITOR, the 
error "Internal Error in DECnet Phase IV AM" is generated.

With regard to the new image for the event logger, we did use the other
image as suggested in the release notes, the one that outputs its data in
NICE format.  This is the image we have been running all along.

Does the above error point to anything more definate on an ULTRIX system?

Al

Al,
	Please try the following:

	%ls -l /tmp/*evl*

	Do you see any of these files:

	    mcc_dna4_evl_pipe ..... is a pipe (you should see a 'p' in the
		begining of the line) that sends down NICE events to
		mcc_dna4_evl_process

	    evl_monitor_pid ..... is a file that keeps the PID of the 
		file mcc_dna4_evl_monitor

	    evl_pro_pid ......... is a file that keeps the PID of the
		file mcc_dna4_evl_process		

	The last two files are for validation.  If one of two processes is not
	up, ENABLE command will return an error message.  

	However, the catch is that if you switch software (for example: v1.1 
	to V1.2), system power down or someone manually killed evl processes, 
	the PID files will not be cleaned up automatically.  Consequently, when 
	the next ENABLE command is issued, the validation will fail because 
	the old PIDs do	not reflect the new PID anymore.  
	
	Let's assume one of above situations has occurred and start from
	scratch.  Delete all three files from /tmp.

	Then start the monitor by setting exec state OFF and on. (this should
	create pid file) 

	Make sure you have event sink and event filter set up 
	(via CREATE/PASS/BLOCK commands).

	Enable the sink process (when sink monitor state is OFF). (At this
	point, check the pipe and pid files again)

	Hope this will resolve your problem.  Thank you.

	Jean	

Many thanks to Jean Lee for helping find the answer.

As it turned out, my ignorance of DECnet on ULTRIX turned out to be the problem.

The default settings were in effect and therefor the only events that were
enabled for this systems DECnet logging were for console logging.

Once we enabled all events for MONITOR logging, I was then able to run the
script, get both processes running and am now receiving remote events.

Al