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

Conference hydra::axp-developer

Title:Alpha Developer Support
Notice:[email protected], 800-332-4786
Moderator:HYDRA::SYSTEM
Created:Mon Jun 06 1994
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:3722
Total number of notes:11359

3100.0. "Prime Associates Inc." by HYDRA::AXPDEVELOPER (Alpha Developer support) Wed Jan 29 1997 03:01

    Company Name :  Prime Associates Inc.
    Contact Name :  Peter Romano
    Phone        :  908-603-0956
    Fax          :  
    Email        :  [email protected]
    Date/Time in :  29-JAN-1997 03:00:48
    Entered by   :  Nick Hudson
    SPE center   :  REO

    Category     :  vms
    OS Version   :  
    System H/W   :  


    Brief Description of Problem:
    -----------------------------

From:	SMTP%"[email protected]" 28-JAN-1997 15:40:07.91
To:	[email protected]
CC:	[email protected]
Subj:	Forward request  

Return-Path: [email protected]
Received: by asimov.mro.dec.com (UCX V4.1-12, OpenVMS V6.2 VAX);
	Tue, 28 Jan 1997 15:40:02 -0500
Received: from pobox1.pa.dec.com by fluid.mro.dec.com; (5.65v3.2/1.1.8.2/19Nov96-0448PM)
	id AA07703; Tue, 28 Jan 1997 15:40:15 -0500
Received: by pobox1.pa.dec.com; id AA25105; Tue, 28 Jan 97 12:40:10 -0800
From: [email protected]
Received: from mail13.digital.com by mail2.digital.com (5.65 EXP 4/12/95 for V3.2/1.0/WV)
	id AA28679; Tue, 28 Jan 1997 12:34:25 -0800
Received: from sandpp.mro.dec.com by mail13.digital.com (8.7.5/UNX 1.5/1.0/WV)
	id PAA21049; Tue, 28 Jan 1997 15:28:45 -0500 (EST)
Received: from localhost by sandpp.mro.dec.com; (5.65v3.2/1.1.8.2/13Nov96-0415PM)
	id AA21702; Tue, 28 Jan 1997 15:28:46 -0500
Message-Id: <[email protected]>
To: [email protected]
Cc: [email protected]
Subject: Forward request  
Date: Tue, 28 Jan 97 15:28:46 -0500
X-Mts: smtp


Hi, 

I helped Prime Associates before in the war room on one issue.  Now,
they have a separate issue and need help.  Can someone in the war room
pick this up?

Thanks,
Miller

------- Forwarded Message

Return-Path: chin
Received: by sandpp.mro.dec.com; (5.65v3.2/1.1.8.2/13Nov96-0415PM)
	id AA21711; Tue, 28 Jan 1997 15:24:40 -0500
Date: Tue, 28 Jan 1997 15:24:40 -0500
From: Miller Chin <chin>
Message-Id: <[email protected]>
To: chin


>From daemon Sat Jan 25 19:55:42 1997
Received: from mail13.digital.com by sandpp.mro.dec.com; 
(5.65v3.2/1.1.8.2/13Nov96-0415PM)
        id AA17803; Sat, 25 Jan 1997 19:55:41 -0500
Received: from gatekeeper.mcimail.com by mail13.digital.com (8.7.5/UNX 
1.5/1.0/WV)
        id TAA26816; Sat, 25 Jan 1997 19:53:05 -0500 (EST)
Received: from mailgate.mcimail.com (mailgate.mcimail.com [166.40.135.3]) by 
gatekeeper.mcimail.com (8.6.12/8.6.10) with SMTP
id AAA29536; Sun, 26 Jan 1997 00:47:04 GMT
Received: from mcimail.com by mailgate.mcimail.com id ac08348;
          26 Jan 97 0:52 WET
Date: Sat, 25 Jan 97 19:51 EST
From: Peter Romano <[email protected]>
To: chin <[email protected]>
Message-Id: <83970126005138/[email protected]>
Status: RO

HELLO Miller Chin at Alpha Developer Support:

You're very welcome (for the compliment..)

On another issue:

As I increase the number of users on my Alpha, more and more
processes including interactive Users are being "swapped" out.
Following is a SH SYS which indicates the concern I have. I note that
at that point response time was was degraded and x-window sessions
tended to act "sticky" and screens did not refresh. I am very concerned
and would like to know how I can improve performance given that I only have
32mb memory.  I am sure it must be possible. I need to get at least
10 users on without all this swapping:

$ sh sys
OpenVMS V6.2-1H3  on node PRIME1 25-JAN-1997 17:29:24.80  Uptime  0 00:56:32

  Pid    Process Name    State Pri      I/O       CPU       Page flts  Pages
00000041 SWAPPER         HIB   16        0   0 00:00:03.91         0      0
00000082 _FTA5:          LEFO   9       --  swapped  out  --             31
00000045 IPCACP          HIB   10        8   0 00:00:00.02        28     18
00000046 ERRFMT          HIBO   8       --  swapped  out  --             19
00000047 OPCOM           HIBO   7       --  swapped  out  --             15
00000048 AUDIT_SERVER    HIB    8      143   0 00:00:00.16       318     30
00000049 JOB_CONTROL     HIB    9      125   0 00:00:00.05       446     36
0000004A QUEUE_MANAGER   HIB   11      119   0 00:00:00.32      2332     34
0000004B SECURITY_SERVER HIB   10       38   0 00:00:00.26      2405     79
0000004C TP_SERVER       HIB    8      238   0 00:00:00.74      2483     43
0000004D NETACP          HIBO   8       --  swapped  out  --             32
0000004E EVL             HIBO  14       --  swapped  out  --             20
0000004F REMACP          HIBO   9       --  swapped  out  --             20
00000050 LATACP          HIB   12       12   0 00:00:00.11       379     30
00000051 TCPware_NETCP   LEF    9      375   0 00:00:00.35       783     49
00000052 TCPware_DNS     HIBO   5       --  swapped  out  --             20
00000053 TCPware_NTPD    HIB    6       27   0 00:00:00.05        83     92
00000055 DECW$SERVER_0   HIB    8      821   0 00:00:01.76      1829     30
00000055 DECW$SERVER_0   HIB    8      821   0 00:00:01.76      1829     30
00000056  DTLOGIN         LEFO   4       --  swapped  out  --             30
00000057  DTGREET         LEF    6      249   0 00:00:02.05      1180     19
0000005B IC_SAP          LEF   10      369   0 00:00:00.08       528     33
0000005C IC_LHSprinter   LEF   10      167   0 00:00:00.28       927     35
0000005D IC_QSG          HIB    6       60   0 00:00:00.16       821     32
0000005E IC_SPL          HIB    6       58   0 00:00:00.16       706     27

00000061 IC_FSS          HIB   10      280   0 00:00:00.33      1209     86
000000A5 _FTA6:          CUR    7     2127   0 00:00:02.54      3775     95
000000A7 PRIME_LOGGER    LEF    7      233   0 00:00:00.38      3063     77
000000A8 FILE_PROC       HIB    6      205   0 00:00:00.36       341     34
000000A9 IO_PROC8        LEF    4       30   0 00:00:00.22       707     34
000000AA IO_PROC10       LEFO   4       --  swapped  out  --             25
000000AB OUTGOING_INTF   HIB    4       73   0 00:00:00.26       429     34
000000AC IO_PROC32       LEF    4       38   0 00:00:00.17       462     34
000000AD IO_PROC40       LEFO   5       --  swapped  out  --             28
0000006E DECW$TE_006E    LEF    6     5830   0 00:00:03.69     33565     63
0000006F PRIMESTARTUP    HIBO   6       --  swapped  out  --             18
000000B0 IO_PROC41       LEFO   5       --  swapped  out  --             26
00000071 PRIMECOMM       LEFO   4       --  swapped  out  --             25
000000B2 IO_PROC42       LEFO   4       --  swapped  out  --             27
000000B3 TRANSMIT        LEFO   5       --  swapped  out  --             28
00000074 DECW$TE_0074    COM    4    12057   0 00:00:09.59     34691    257
00000075 _FTA3:          HIBO   6       --  swapped  out  --             22
00000076 _FTA4:          HIBO   5       --  swapped  out  --             21
000000B7 MSG_PROC        HIBO   5       --  swapped  out  --             28
00000078 _RTA2:          LEFO   4       --  swapped  out  --             28
00000079 _RTA3:          LEFO   4       --  swapped  out  --             27
$


Also, how do I shut/prevent from starting the DTLOGIN and DTGREET processes
 I do not know what they are..?
thanks again for yr help,

Peter Romano
Prime Associates Inc.
[email protected]
908-603-0956 voice






------- End of Forwarded Message

T.RTitleUserPersonal
Name
DateLines
3100.1KZIN::HUDSONThat&#039;s what I thinkWed Jan 29 1997 07:1878
From:	DEC:.REO.REOVTX::HUDSON       "[email protected] - UK Software
Partner Engineering 830-4121" 29-JAN-1997 12:17:22.50
To:	nm%vbormc::"[email protected]"
CC:	HUDSON
Subj:	RE:excessive swapping on VMS 32Mb system

Hello Peter,

Thanks for your question on excessive swap-outs.

The short answer to your question is that you need to add extra physical
memory to your machine (solution 1), or reduce the memory requirement of the
processes that you're running (solution 2).

You say you have 32Mb, and the 6.2 SPD for VMS says "The minimum amount of
memory required to install, boot, and log in to an OpenVMS Alpha system is 32
MB.  To ensure satisfactory performance for particular applications or numbers
of users, additional memory may be required."

From the SHOW SYSTEM you are running Motif, TCP/IP and DECnet, and a fair
number of other processes, so I don't think the system is behaving in an
unexpected way when it is swapping out so many processes.

Processes being swapped out is not necessarily a bad thing.  If a process isn't
doing something, then on a memory-constrained system such as yours, you want
that process to be swapped out and left there.  It is a problem if the demand
on memory is such that a set of processes are continually being swapped in and
out.  If that is happening, then this indicates that the collection of
processes you have has a net memory requirement which exceeds the physical
memory available.  The only thing you can do about this is to add memory
(solution 1), or reduce the memory requirement of the processes (solution 2).

If you have, say, 10 processes, which when they're all running need an
aggregate 40Mb of physical memory, then no matter what you do to your system
parameters, VMS is going to have to be moving things to and fro from the disk
in order to keep going.  If it wasn't swapping, it would have to page, or just
lock up altogether.

It is conceivable that by tweaking some of the SYSGEN parameters you may affect
the behaviour of the system a bit, but I this will likely be a bit of a
hit-and-miss affair, and I would be surprised if you can significantly improve
on whatever the settings are that AUTOGEN has derived.

For example, it may be possible that you could cut down on swapping by
constraining the physical memory that processes are allowed (i.e. reducing
working-set extents).  This might have the effect of making processes use less
physical memory, which might mean that you'd get more processes on before
swapping started.  But then the processes would just be paging more, which
could quite easily mean poorer performance than the original swapping was
causing.

Over a period of time, you should find that the processes swapped out are those
which aren't doing much (and so don't need to be swapped in).  For example on
your SHOW SYSTEM, you can see ERRFMT and OPCOM swapped out.  You can stop
processes from being swapped out by using "SET PROCESS/NOSWAP", but if you
apply that to an interactive process that belongs to someone who has gone to
lunch, you're wasting a lot of resources.  And that won't stop them being paged
out to disk.


Looking at your SHOW SYSTEM, the process with the most physical memory by far
is a DECW$TE... session.  This is what you get when you're running a DECterm. 
DECterms use a lot of memory, so maybe you could avoid having them running on
this machine (solution 2).  In fact if you could use this machine as a server
rather than a workstation (i.e. don't run DECwindows), I think that would free
up a fair amount of memory (solution 2).  Incidentally, the DTLOGIN and DTGREET
processes are DECwindows processes.

If you want to run without DECwindows, you can define the logical
"DECW$IGNORE_DECWINDOWS" to "TRUE" inside SYS$STARTUP:SYSTARTUP_VMS.COM.

I hope this is of some help to you.  I'm sorry that there isn't an easier
answer along the lines of "increase this SYSGEN parameter"...

Regards

Nick Hudson
Digital Software Partner Engineering
3100.2KZIN::HUDSONThat&#039;s what I thinkThu Jan 30 1997 05:5289
From:	HYDRA::AXPDEVELOPER "[email protected]" 30-JAN-1997 07:48:21.76
To:	rdgeng::hudson
CC:	AXPDEVELOPER
Subj:	FWD: Re: Xcursions error msg  

From:	SMTP%"[email protected]"   29-JAN-1997 17:27:27.95
To:	Peter Romano <[email protected]>
CC:	[email protected], [email protected]
Subj:	Re: Xcursions error msg  

Return-Path: [email protected]
Received: by asimov.mro.dec.com (UCX V4.1-12, OpenVMS V6.2 VAX);
	Wed, 29 Jan 1997 17:27:25 -0500
Received: from pobox1.pa.dec.com by fluid.mro.dec.com; 
(5.65v3.2/1.1.8.2/19Nov96-0448PM)
	id AA27525; Wed, 29 Jan 1997 17:25:38 -0500
Received: by pobox1.pa.dec.com; id AA06657; Wed, 29 Jan 97 14:23:15 -0800
From: [email protected]
Received: from mail13.digital.com by mail1.digital.com (5.65 EXP 4/12/95 for 
V3.2/1.0/WV)
	id AA15801; Wed, 29 Jan 1997 14:13:22 -0800
Received: from sandpp.mro.dec.com by mail13.digital.com (8.7.5/UNX 1.5/1.0/WV)
	id QAA23884; Wed, 29 Jan 1997 16:59:22 -0500 (EST)
Received: from localhost by sandpp.mro.dec.com; 
(5.65v3.2/1.1.8.2/13Nov96-0415PM)
	id AA23251; Wed, 29 Jan 1997 16:59:20 -0500
Message-Id: <[email protected]>
To: Peter Romano <[email protected]>
Cc: [email protected], [email protected]
Subject: Re: Xcursions error msg  
In-Reply-To: Your message of "Wed, 29 Jan 97 09:26:00 EST."
             <83970129142638/[email protected]> 
Date: Wed, 29 Jan 97 16:59:20 -0500
X-Mts: smtp


Hi Peter,

I will forward your question to our ASAP hotline
support.  One of our on-duty technical support
staff will response to you shortly.  If you
have further technical questions and issues,
please feel free to call us or send email 
to [email protected].

Regards,
Miller


ate: Wed, 29 Jan 97 09:26 EST
From: Peter Romano <[email protected]>
To: chin <[email protected]>
Subject: Xcursions error msg
Message-Id: <83970129142638/[email protected]>

ATTN:
Mr. Miller Chin, DEC ASAP Group

Hello Miller:

I was able to resolve the issue of the DTLOGIN and DTGREET
processes on my Alpha which I noted in my last e-mail by calling
CSC. These processes result by default from fact that Alphastation
is a graphics station (ie: with a graphics device) and so the CDE
environment is started. CSC told me how to avoid these processes
by setting a symbol in my CDESTARTUP.com and also how to avoid
the 2 DECW$server processes to avoid their overhead. That is done
by adding 2 symbol values to SYSTARTUP_VMS.COM.

I now have a question re: following excursions log message which I get
from time to time and thjat I would like to eliminate. Could you
please provide necessary information for resolution.

1/28/97    7:15:03 PM   X server Error  XDM:  too many retransmissions


Thanks,

Peter Romano
Prime Associates Inc.

6359777 at MCIMAIL.com
908-603-0956 voice

end




3100.3KZIN::HUDSONThat&#039;s what I thinkThu Jan 30 1997 11:0532
From:	DEC:.REO.REOVTX::HUDSON       "[email protected] - UK Software
Partner Engineering 830-4121" 30-JAN-1997 16:04:28.75
To:	nm%vbormc::"[email protected]"
CC:	HUDSON
Subj:	RE:  Xcursions error msg  

Hello Peter,

>I now have a question re: following excursions log message which I get
>from time to time and thjat I would like to eliminate. Could you
>please provide necessary information for resolution.
>
>1/28/97    7:15:03 PM   X server Error  XDM:  too many retransmissions


I can't find reference to this error anywhere on any of our internal databases.

Can you tell me

	- what's the exact format of the message (if it isn't what you've
	  written above)
	- where this error appears (on the console? in a file? which file?)
	- how often you're seeing the error
	- what other errors you're seeing around the same time (if any)
	- whether you're noticing anything fail on the system at around the
 	  time these errors occur
	- when you first noticed the errors

Regards

Nick Hudson
Digital Software Partner Engineering
3100.4KZIN::HUDSONThat&#039;s what I thinkThu Feb 20 1997 05:2846
From:	DEC:.REO.REOVTX::HUDSON       "[email protected] - UK Software
Partner Engineering 830-4121" 20-FEB-1997 10:28:11.15
To:	nm%vbormc::"[email protected]"
CC:	HUDSON
Subj:	re: xcursion error message

Hello Peter

A while ago you mailed in a problem description with an error you were getting
from X server relating to excursion.  I sent you the following mail on 30-Jan
but haven't heard from you since.  Is this still causing you a problem or have
you resolved it now?

Regards

Nick Hudson
Digital Software Partner Engineering


> Hello Peter,
> 
> >I now have a question re: following excursions log message which I get
> >from time to time and thjat I would like to eliminate. Could you
> >please provide necessary information for resolution.
> >
> >1/28/97    7:15:03 PM   X server Error  XDM:  too many retransmissions
> 
> 
> I can't find reference to this error anywhere on any of our internal databases.
> 
> Can you tell me
> 
> 	- what's the exact format of the message (if it isn't what you've
> 	  written above)
> 	- where this error appears (on the console? in a file? which file?)
> 	- how often you're seeing the error
> 	- what other errors you're seeing around the same time (if any)
> 	- whether you're noticing anything fail on the system at around the
>  	  time these errors occur
> 	- when you first noticed the errors
> 
> Regards
> 
> Nick Hudson
> Digital Software Partner Engineering
> 
3100.5KZIN::HUDSONThat&#039;s what I thinkThu Mar 13 1997 08:381
timed out