Conference csc32::consolemanager

848.0. "Archive & Connect Problems" by OPCO::TSG_TAS (Mr Mass) Mon Jul 03 1995 18:25

    Folks
    
    VMS 6.1
    PCM 1.5-006
    
    A couple of problems...
    
    Firstly, a situation where an archive was stopped which left PCM aware
    that a particular managed system was being archived. Reconfiguring,
    stopping and restarting, deleting the system, restarting and readding
    the system using another name (S35A instead of S35) have all failed to
    get this system out from under the archive in progress flag. How do we?
    
    aside (I had thought that the configuration file must be ISAM and
    determined to try to solve this myself by seeing whether the archive
    information was held in this file. anal/rms revealed my naivity... this
    is a Unix hangover - a sequential stream file. Are we ever to be rid of
    this steam computing? Was it 25 years ago that IBM invented ISAM? My
    own bookshelves only revealed a tome � 1981 that suggested that IBM was
    dumping ISAM in favour of VSAM. We know the usefullness of ISAM in the
    VMS environment. Why a sequential stream file? Perhaps it has something
    to do with some early papers I read on Unix (which made up the Ultrix
    docset) which showed that the writers of Unix (2 man years, 10000 lines
    of code) reckoned that a locking mechanism was unnecessary as the
    operating system would, in the main, be used for editing documents etc.
    Here we are editing files and the problems caused by not being able to
    lock at the record level... soln in 1.6? lock at the file level. NT has
    better locking than VMS, but perhaps we shouldn't mention the war)
    aside over
    
    Secondly, we are experiencing a situation where PCM appears to be
    funtioning correctly except that no monitor, nor connect can be made.
    Every other aspect seems to work - showing events is up-to-date. But
    connect and monitor clear the screen and hang. Even the lock icon is
    not placed on the connected to system... I fear a resource limit, a
    reboot will fix (can only schedule one in 24 hours) is there a way to
    see which resources could be depleted? CONSOLE$TMP made available to
    the lads by mail.
    
    Tony Sherrard

Tony,

  We fixed a problem today where the archive flags in shared memory were
  not being cleared down correctly, this could be the problem (I'm almost
  certain that it is) you are seeing. This fix will be in the first ECO for
  V1.6 (No, dont ask me when, it all depends on resources).

  The archive flag is held in a shared memory section, dont whatever you
  do go playing with it, you can upset a whole bunch of other stuff if
  you do!!!

  The file is the format it is (stream) because we are a multi-platform
  product, we (Digital) dont offer a no cost ISAM files on Digital UNIX,
  so to make support easy we use the lowest common denominator. The same
  goes for the log files.

  Finally, upgrade to V1.6, it will fix a lot of your problems I believe
  the problem you have with the connect sessions hanging is fixed in 1.6.

Cheers,
Phil

    Folks
    
    A reboot has cleared both problems
    
    We are an FM or OMS or whatever it is now site, so have to be real
    careful about new versions of software. This is the first time we have
    seen either of these problems in well over six months. We are a site
    that has had bad experiences with the RWMBX problem, though this has
    been less of a problem recently.
    
    I have been holding off upgrading because of the processes going into
    LEF after an archive problem. If we have to restart PCM each archive
    it's pointless upgrading. How ubiquitous is this problem? In what sort
    of sites does it occur?
    
    Tony Sherrard

Tony,

  We believe we have just found the cause of the LEF after archive
  problem, after testing thoroughly over the next day or so, we will
  be releasing a IFT of the first ECO kit for V1.6.

Cheers,
Phil