HP3000-L Archives

August 1995, Week 4

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Jim Wowchuk <[log in to unmask]>
Reply To:
Jim Wowchuk <[log in to unmask]>
Date:
Mon, 28 Aug 1995 11:28:33 +1000
Content-Type:
text/plain
Parts/Attachments:
text/plain (39 lines)
At 10:39 AM 23/8/95 CDT, Craig Vespe wrote:
>   We've got the same problem here.  We worked around the problem by
>   reconfiguring the "keepalive" parameters in MPE.  In NMMGR, the TCP
>   Config screen has two parameters - the Connection Assurance Interval
>   (Secs), and the Maximum Connection Assurance Retransmissions.  The
>   default values are 600 seconds and 4 retrans.  These figures give
>   you around 40 minutes of idle connection.  We (Jeff Woods, actually)
>   now use 900 and 40, respectfully, giving around 10 hours - ample
>   time to get back from lunch. :-)
>
>   It is left as a non-trivial task of the user to actually find the
>   TCP Config screen (#94) in the depths of NMMGR.
 
There existed a while back a problem with F0. IMAGESQL, such that the daemon
process that was created for each user could become an orphan.  It was
waiting on activity from a closed TCP/IP socket (!).  The only means of
killing this process was to reboot.  But an alternative was to let it die
itself, which it would do when no "keepalive" packet was received.  Rather
than wait 40 minutes, our method was to reduce the keep alive down to 60
seconds by 4 - 4 minutes.  Reflection doesn't seem to have a problem with
this (I think it is sending a keepalive every 60 seconds), but now vt3k from
HPUX will die after 4 minutes of inactivity.
 
If anyone is still using this old version of IMAGESQL (my apologies for not
specify the version, but I've not tested this since moving to the G0
version), they may want to consider if changing the keepalive timing would
deterimentally affect them.
 
Then again, maybe its time to do some Spring cleaning of my mind to remove
the obsolete pieces of info collecting up there. :)  (I remember the orange
pocket guides too!)
----
Jim "seMPEr" Wowchuk           Internet:    [log in to unmask]
Vanguard Computer Services     Compu$erve:  100036,106
 _--_|\                        Post:        PO Box 18, North Ryde, NSW 2113
/      \                       Phone:       +61 (2) 888-9688
\.--.__/ <---Sydney NSW        Fax:         +61 (2) 888-3056
      v      Australia

ATOM RSS1 RSS2