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

Conference bulova::decw_jan-89_to_nov-90

Title:DECWINDOWS 26-JAN-89 to 29-NOV-90
Notice:See 1639.0 for VMS V5.3 kit; 2043.0 for 5.4 IFT kit
Moderator:STAR::VATNE
Created:Mon Oct 30 1989
Last Modified:Mon Dec 31 1990
Last Successful Update:Fri Jun 06 1997
Number of topics:3726
Total number of notes:19516

2499.0. "xlib io error again!" by ISTG::DWILSON () Thu Mar 22 1990 13:04

    I understand from other notes that to reduce 
    	"xlib io error, connection aborted" I need
    to set some (?) values for DECW$SERVER_RETRY_WRITE)_MIN and MAX ...
    However, how do I decide what values I should use ??????
    
    I'm running V5.2 - can't upgrade to V5.3 yet.
    I have some users of decwrite and decdecision getting
    dumped out of their remote processes daily....
    
    Is there a rule-of-thumb as to what values we should change
    those logicals to????
    
T.RTitleUserPersonal
Name
DateLines
2499.1DECWIN::FISHERPrune Juice: A Warrior's Drink!Fri Mar 23 1990 15:0818
You might find that most of the people who should know (like me) don't remember
any more.  There are some suggested values elsewhere in this conference.  Try
searching for authors like TAN and FISHER.

Remember, these won't necessarily help if you really are in a deadlock.  They
will only give the system a better chance to keep trying to hopefully find out
that you really are not deadlocked.  What you want to do is decrease the
retry interval  (like to only a few milliseconds) and increase the number of
retries.  I think that the number of retries is actually specified in
total number of milliseconds (or was that V5.1?).  In any case, if you can't
find the note in question, you probably want to retry every 10 ms. or so and
let it repeat for 15  - 30 seconds.  If you want to make it repeat for longer,
you can; it's just that the workstation is locked up for all that time.

You really want to upgrade as soon as you can; the problem is essentially fixed
in 5.3.

Burns
2499.2rtrn'sGERUND::WOLFEThree dimensional programming on a 2-D screenFri Mar 23 1990 15:351
The gory details are in teh DECwrite release notes....
2499.3DECWIN::FISHERPrune Juice: A Warrior's Drink!Tue Mar 27 1990 13:326
I looked at the CD release notes and could not find a reference that looked like
it.  I can go search deeper when my curiosity gets the better of me, but if
there is a lesson to be learned for other layered products, I'd appreciate it
if you could publicize it.

Burns
2499.4from V1.0 rn'sGERUND::WOLFEThree dimensional programming on a 2-D screenTue Mar 27 1990 16:0982
    It's section 1.1.5 Installation issues related to VMS DECwindows....

    
    

          1.1.5  Installation Issues Related to VMS DECwindows

          1.1.5.1  Connection Aborted - XLIB-F-IOERROR

          PROBLEM: A DECwrite session is sometimes terminated with a
          message "xlib-f-ioerror, connection aborted."

                                        NOTE

             The problem usually occurs only when you run DECwrite
             remotely and display to your local workstation. Please see
             your system manager for assistance.


          When using the X Window System, events such as a mouse motion
          and keyboard input are queued by the X server and sent to the
          client application as it requests them. It is assumed that a
          client is continually processing these events. If the client is
          not processing events quickly enough, then events are queued by
          the server for later delivery. This might occur, for example,
          if the client is running on a machine that is compute-bound.
          Since buffers in the X server and the network are limited, the X
          server will eventually time-out the client application.

          WORKAROUND: The retry count and retry interval are set with some
          internal default values, but you can modify them by defining two
          logical names:

               DECW$SERVER_RETRY_WRITE_MIN
               DECW$SERVER_RETRY_WRITE_MAX

          DECW$SERVER_RETRY_WRITE_MIN defines the interval between retries
          in the unit of 1/10 of a ms. For example, the default is 15000,
          which is .15 second.

          DECW$SERVER_RETRY_WRITE_MAX defines the amount of time the
          server will spend retrying before forcing the client to be
          disconnected. The default value for this is currently set to
          300000, which is approximately 20 retries, in 3 seconds.

          For VMS 5.1 users only, we suggest that you define these two
          logical names with the following values:

               $ define/table=DECW$SERVER0_TABLE -
                      DECW$SERVER_RETRY_WRITE_MIN 150000
               $ define/table=DECW$SERVER0_TABLE -
                      DECW$SERVER_RETRY_WRITE_MAX 3000000

                                        NOTE

             These logicals are defined in a workstation's logical
             name table (that is, they are server logicals, not client
             logicals). In a VAXcluster, the logicals must be defined
             on each workstation in the cluster.


          By defining these logical names, you increase the length of
          retry interval and total retry time by a factor of 10, so the
          resulting retry period will be 30 seconds with a 1.5 second
          delay between each retry.

          The logical names are translated each time the server resets.
          The user can define the logical names, quit the current session,
          and log back in to use the new retry values (logging out resets
          the server).

          For VMS 5.2 users, the logical names are still valid but the
          units for the default values have changed to milliseconds. The
          VMS 5.2 defaults are:

          o  DECW$SERVER_RETRY_WRITE_MIN is 500 (0.5 seconds)

          o  DECW$SERVER_RETRY_WRITE_MAX is 30000 (30 seconds)

          Try adjusting these numbers if you encounter problems on VMS 5.2
          systems.

2499.5Oh dear...wrong noteDECWIN::FISHERPrune Juice: A Warrior's Drink!Wed Mar 28 1990 13:289
Oh gosh!  I'm sorry.  When I saw your first RTRN note, I thought this was the
topic about images that somehow got jumbled when displaying on a different-
endian machine.  That's the one which I did not understand why it was a
DECwrite problem.

Thanks for digging up this info though despite my lack of checking previous
replies!

Burns