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

Conference csc32::consolemanager

Title:POLYCENTER Console Manager
Notice:Kits, Scans, Docs on CSC32:: as PCM$KITS:,PCM$DOCS:, PCM$SCANS:
Moderator:CSC32::BUTTERWORTH
Created:Thu Aug 06 1992
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:1541
Total number of notes:6564

661.0. "Console link lost - again, more info" by KERNEL::ROOMEM () Mon Mar 27 1995 18:31

 Dear Dan/All

 I'd like some clarification on the operation of UCX and PCM. - 

Two questions.

 1) My customer sees intermittent problems were by he gets the message

 "console link for system x lost, error 0"

 He loses all TCP connections to all nodes, the nodes are using different 
terminal servers so the problem is local to the PCM node.

 Now, does PCM overide the PROBE and DROP timers within UCX when it makes the 
connection to the DECserver ? Are there hardcoded probe and drop timers within
PCM or can I change the probe and drop timers in UCX to retain the connection 
during this intermittent network problem ?

 2) What time period is used to make the reconnection to the nodes from the 
child process. Is this value hardcoded in PCM or is there a logical or variable 
that I can change that will force a reconection sooner.

 If the above doesn't make sense then let me know and I'll try and rephrase it.

 Thanks in advance for the answer.

 Martin Roome
 VMS Specialist
 CSC Basingstoke.   

T.RTitleUserPersonal
Name
DateLines
661.1OPG::PHILIPAnd through the square window...Mon Mar 27 1995 19:1237
Martin,

>>Now, does PCM overide the PROBE and DROP timers within UCX when it makes the 
>>connection to the DECserver ? Are there hardcoded probe and drop timers within
>>PCM or can I change the probe and drop timers in UCX to retain the connection 
>>during this intermittent network problem ?

  We do simple QIO's to connect to the remote TELNET server, we dont do ANYTHING
  with the UCX setup

  If we see that we have had no console data from the remote end withing X seconds,
  then in the case of TELNET, we send a TELNET GETSTATUS option and wait for a
  response. If no response is forthcoming within (I think 30 seconds) we declare
  the connection DEAD, close our end and put out an event. We will then attempt to
  reconnect after Y seconds (see the next comment).

  This all assumes that you have a terminal server which supports TELNET GETSTATUS,
  if you dont, then we detect this when we make the first connect and tell you that
  we will not be doing any further checks, we effectively will only notice the link
  has gone away when we attempt to do a QIO and it fails.

>>What time period is used to make the reconnection to the nodes from the 
>>child process. Is this value hardcoded in PCM or is there a logical or variable 
>>that I can change that will force a reconection sooner.

  If you start up the editor and type SET HIDDEN one of the parameters is used as
  the timer for reconnects, but beware, if you change it and everything breaks,
  you dont get any support from us.

  Having said all that, we have known TELNET connection ride out a network problem
  easier than say LAT, but, basically if you customer has a didgy network, get
  them to fix it!!!

Cheers,
Phil


661.2ThanksKERNEL::ROOMEMWed Mar 29 1995 11:196
    
     Phil.
    
     Thanks for the quick response, much appreciated.
    
     Martin