Conference csc32::consolemanager

    If it's any help I have a copy of the core, on briris (DECnet 44.209)
    :-
    
    briris::/more-users/euan/microtel/cm-core
    
    Thanks again,
    
    		Euan

Euan,
     have you installed the patch kit foe ECO #2?

If not look on elbrig.reo.dec.com. in anon ftp account

       Console_Manager/CM_ECO_2 directory

      DECnet
 
             elbrig::/usr/users/ftp/Console_Manager/CM_ECO_2 directory.

   The core will be of little use as the images are stripped of the bebug symbol
tables.

Instal patches see if problem goes away. If not get back to us and we will build
some unstripped images to get a valid core dump.

Cheers simon....

    Simon,
    
    	Thanks, I'll put the patches in tomorrow and see how we go. 
    
    	Thanks again,
    
    		Euan

    Simon,
    
    	Apolgies for the delay. We do still have a problem, even with ECO 2
    installed. Can you let me have a pointer to the unstripped image so we
    can get core dumps that mean something ?
    
    	Also, do you want  me to log this officially through the CSC ?
    
    	Euan

Euan,
     copy the file:

                    elbrig::/usr/users/ftp/Unstripped_ensd

 
       or anon ftp form elbrig.reo.dec.com.


Checksum:

35790   526

Cheers SImon...

    Simon,
    
    	Thanks for the debug version. It turns out we didn;t need it as we
    found the cause of the problem. Somehow (not me honest !!) as scan
    event was put in with a context count of 25, which crashed cm whenever
    the event happened. Correcting this sorted the problem, and we have had
    no overnight crashes now.
    
    	Thanks again,
    
    		Euan

    Interesting...
    
    A context count of 25 should not cause that problem though. We have
    larger numbers than that which were ok.
    
    If you have any similar problems let me know.
    
    Cheers Simon....

    
    Hello,
    
    here's another DEC OSF/1 with PCM V1.1 doing core dumps. We notice this
    while the system is very ssssslllloooowwww. Via the 'Console Manager'
    'Startup' we solve the problem and everything works blazing. But now
    and then spontaniously ? the PCM stops running leaving a core. Also
    when I do a console -m it sometimes dumps core.
    
    Since I'm not too familiar with DEC OSF/1 can there be any system stuff
    set incorrectly ?
    
    PCM V1.1 has about 5 systems in the database where only 2 have a
    working console-line, so it is not heavily used (yet). It is a TELNET
    connection setup all the way covering (in production) a whole country
    with apporximately 25 sites.
    
    I have 2 core dumps at UTRUST::SWAS90:[RIJSBERGEN.PUBLIC] but as you
    said already you probably can not use it.
    
    Regards, Marcel.

    
    I have noticed, if telnet connection is lost (line or router is down
    for a short time) PCM sometimes dumps when next console -m command
    is given.
    
    -Olli
    

    RE .9
    Could be the case since we see sometimes 'link lost' messages. But the
    core dump also occurs when the C3 is on the screen. For a Console
    management product this is not appropriate !!! It should give messages
    but should not dump core. Until somebody starts PCM again there are no
    console scans !!! So no controle via ENS etc.
    
    Should I put a 'core-dumpable' images in place to supply dumps for debug ?
    
    Regards, Marcel

  We will be making a release of V1.5 available shortly (withing the next
  week or two), V1.5 is effectively a patch kit to V1.1, but we called it
  V1.5 as the OpenVMS VAX release had to be called that for users upgrading
  from VCS, this way we keep everything in sync. Anyway, we fixed a few
  daemon and monitor problems in this kit, can you wait until we release it,
  try it, if the problems you see persist, then we can arrange to get you
  images which will provide us with meaningful core dumps.

Cheers,
Phil

    OK, Phil
    
    will await new kit and announcement.
    
    Regards, Marcel.

    
    I finally had time to install V1.5 on ULTRIX. Very soon after
    installation telnet connection was lost to one DECserver90M
    with two system connected with PCM. PCM V1.5 reported nicely
    about connection lost (error 0) and recover connection.
    And PCM V1.5 worked just fine after this. 
    
    Thanks for fixing this!
    
    -Olli
    

    
    We started to get consoled dumps again.
    PCM V1.5
    ULTRIX V4.3A
    Telnet connections to DECserver300's, DECserver90M's and one
    system on Cisco terminal server. Again it seems when
    connection to one or more terminal servers breaks
    consoled dumps. But sometimes PCM is able to report about
    lost connections. 
    
    Could you provide again modified consoled image which provides
    more information  in core file. I would be happy to help again
    because we do couple of very critical 7*24 monitoring tasks with
    PCM in this environment.
    
    -Olli
    
    
    
    

    
    
    
    More core dumps on v1.5, v4.4 ultrix. 
    
    Dumps about once an hour (not regularily).  
    Customer can't reproduce it at will (tried disconnecting servers etc.
    as the last note implied). 
    
    NO other indications of problems by the way. 
    
    

	We have been using new consoled image which hasn't dumped.
	Modified consoled will in next patch kit.

	-Olli

    
    
    More core dumps with PATCHED v1.5 on OSF v3.0. 
    Core dumps about 3 times a week. Customer can't relate it to any
    particular event. 
    Nothing in the log files. 
    
    Core dumps are available if they would help.
    Customer states that this is very critical so I will probably need to
    escalate this as critical by tomorow.
    
    thanks

  Can I have a copy of the customers database, thats is an EXPORT of it.

Cheers,
Phil