[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

3205.0. "TWA devices (only) hang" by VLNVAX::TSMITH (That rabbit's dynamite...) Mon Aug 13 1990 23:41

	***  This is in DECwindows, VMSTUNING & HACKERS  ***

I'm having a TWA device performance problem on one of my clusters.  

Users accessing the cluster (not just individual nodes on the cluster) using
RTA or V/LTA devices perform as expected.  If users access the cluster using
TWA devices, their process' hang for upwards of 20-30 seconds.  Also, this
may happen as often as every 2-5 minutes.

At first the network was suspected.  We tested the network, LAT & the 
DECwindows devices at the same time and used MONIT/PROC/TOPC/inter=1.  Only 
the TWA device hung.  All 3 processes were from the same user account.  Once 
the TWA process returned, the data was flashed on the screen in a second or 2
until it had cought up (as if we'd done ^S & ^Q).  We've tried SYSGEN 
parameters with no luck.  The systems were tuned again this weekend with no 
luck.

Where DECwindows uses the network I'm still leaning torwards something in the
network parameter area even though the RTA session wasn't affected.

Thanks,
/Tom
T.RTitleUserPersonal
Name		DateLines
3205.1RTA devices are also hangingVLNVAX::TSMITHThat rabbit's dynamite...Wed Aug 29 1990 00:5717
Hoping to awaken this note here's an update:

Someone sent mail recommending I check the counters/errors etc in NCP.
Good idea but no problems there.

Additionally, RTA processes DO hang so it's not just TWA devices that are
involved.  One easy test I was shown, which reproduces the error within a 
few minutes is to create a DECterm then slide the pointer along the menu bar 
causing the system to repeatedly pop up the different menus.  Eventually a 
popup menu will freeze for several seconds.

At this point all bets are off.  Any suggestions are welcome.  The Users are
really being impacted.  It appears to happen only in their section of the
building.  Other users on the cluster seem to work without experiencing the
problem at all.

/T
3205.2Could it be your segment? Maybe ncp service parameters?IO::MCCARTNEYJames T. McCartney III - DTN 381-2244 ZK02-2/N24Wed Aug 29 1990 15:3414

Is the system on which you are seeing the stall (I consider a hang to be
not recoverable) the boot node for something? Do you have a large number 
of boot requests on your ethernet segment? We have seen this type of problem
and almost always it has been tracable to either the network was incorrectly
set to with service enabled or there is a node on the segment which is flooding
the segment with "boot me" requests.

BTW, for some reason, LAT doesn't seem to be as sensitive to the problem of
an overloaded segment - so even while CTERM and X11 links were seeing the
problems here, the LAT connected terminals seemed fine. Don't ask me why!

James
3205.3I'll forward it to NetworksVLNVAX::TSMITHThat rabbit's dynamite...Thu Aug 30 1990 10:176
James,

Thanks for the input. I'll forward it to our network group.  I'm out of
the office right now but will leave a note whether that's it or not.

/T