[WT-support] Advanced setup network for M2

Matt IZ3EYZ iz3eyz at virgilio.it
Wed Nov 7 01:27:11 CET 2007


Hi Olivier & Larry.

Following your emails and Claudio ones, let me add a few things and go a bit
deeper on a few others, so maybe we can solve this issue ASAP.

1)These PCs have no internet access, some of them had virus infections
(trojans and worms) and were cleaned up in early Sept. The infection did hit
.EXE files, new software couldn't be installed, Windows wasn't running
anymore. One of these worms had the ability to broadcast thru LAN in
Win98/Y2K, thus whole local PCs were infected with the same viruses. Managed
to boot PCs in provvisory-modality and happened to install NOD32, cleaning
was successfully performed. Other anti-vir. didn't recognize any
virus as reported to me by Claudio. I remember Claudio should have installed
and checked with anti-spyware too and response was negative.

2)3 out of 8 PCs were affected by the slowdown/partial frozing and they are
the ones where people were typing stuff more frequently (RUN 1 & RUN 2 main
radio, MULT A stn). The dedicated CLX PC (UHF connection to CLX) has always
been vy slow although it's a P-IV. It was reported to me this PC wasn't
infected by the above viruses. NOD32 scanning response was negative. By that
time it had Win XP (so no virus attack), now it has Win Y2K. I didn't touch
this PC but on Sunday afternoon. I disconnected it from WT and LAN to run a
few tests I will describe at point 3).

3)
3.1 Moved CLX connection to another PC and the problem (frozen/slowdown)
kept going.
3.2 Disconnected the 3 PCs from all CLX connections (both in and outbound)
and they returned to easy life.
3.3 Brought back log sync EVEN to the 3 affected PCs. Still easy life.
3.4 Brought back CLX in/out bound one by one (test performed just on the
MULT A PC) and it was clearly due to: INBOUND PACKET SPOTS from INBOUND
PROTOCOL.
3.5 .PKT file (2.5MB by that time) was deleted. Nothing changed.
3.6 PC reboot to see if after days of non stop using it started to be memory
consuming. Nothing changed.

At this point I didn't do anything else but leaving CLX off on the 3
affected PCs, didn't run more tests as you can't either operating and
playing with PCs. The MULT A PC  was the one showing the worst
slowdown/partial frozing and it wasn't acceptable even for SSB at times. On
RUN 2 and RUN 1 the slowdown wasn't that bad and one could even live with
that. The main concern, considering we might even live without CLX spots on
3 PCs, is about what might happen in WW CW. Using PC CW generation I expect
erratic CW and chopping! This is the main issue and a PITA.

Yep, I do agree with you Olivier we should have checked Task Manager, I
thought that the day after the contest, after a good sleep and with a clear
mind.

Definitely this is the thing to check. About a "SH/DX/5000" request as far
as I know, at least DX-Spider clusters, can handle up to SH/DX/100 (maximum
spot value they can show you even if you type more), so I'm afraid it can't
be reproduced the CQWW scenario/spot rate. Anyhow it would be interesting to
see if even at a low spot rate the problem still exists and what about
tasks/CPU usage.

Anyhow these tests and more PCs hardware infos are up to Claudio.

FYI we had a Win Vista laptop running as well (might it be a problem??) and
it was solid like a rock!

As far as I remember it was reported to me NOD32 was quitted on Saturday
late evening when first slowdown signs started to appear. Yes, NOD32 has a
file system real-time scanning but it's vy light. I'm repeating this on my
PC with a WT log + WT DXtelnet running open:

Quittting NOD32 it keeps NOD32KRN.EXE (user name SYSTEM) which has 0% CPU
usage and 4.5MB memory usage. NOD32KUI.EXE also joins with NOD32 running (0%
CPU usage and 756 KB memory usage). I did that on a P IV here, 2ghz, 256MB
RAM. Both WT tools don't have more than 2-3% CPU usage. Memory usage is in
line with your qso nr parameters.
Even right now I have a lot of other tasks running like Skype, MSN,
internet, etc and NOD32 CPU usage keeps low as usually, mostly 0 and when
working it stays at 2/3% most of the time.



73

Matt.EYZ






----- Original Message ----- 
From: "Olivier F5MZN" <olecam at f5mzn.org>
To: <support at win-test.com>
Sent: Tuesday, November 06, 2007 6:57 PM
Subject: Re: [WT-support] Advanced setup network for M2


> HI -
>
> --- 8< --- [snip]
>> Check also that you don't have any "automatic and silent" Windows update
>> engaged. That might impact the responsivness and file system access, for
>> low-end machines.
>
> In addition to Larry, I suggest the next time it happens to open the
> Task Manager in order to see which processes are spending to much CPU
> time. Usually, WT should not use more than 2 or 3% of CPU, with a
> reasonable memory usage (less than 40 MB with all databases and an empty
> log, and abt 40/50 MB with 8000 Q loaded).
>
> Are you able to repeat the problem, sending some "SH/DX/5000" to the
> cluster?
>
> BTW, is the HD formatted in NTFS, FAT32? No compressed file system, or
> network shares around?
>
> As Larry, my guess is that hidden process is silently analyzing each
> file a software access to. This can be an anti-virus, anti-spyware ...
> or even a virus or spyware! One clue is that only some of your computers
> in the LAN were affected by this problem.
>
> HTH
>
> 73,
> -- 
> Olivier / f5mzn
> _______________________________________________
> Support mailing list
> Support at win-test.com
> http://www.f5mzn.org/cgi-bin/mailman/listinfo/support
>
> __________ NOD32 2639 (20071105) Information __________
>
> This message was checked by NOD32 antivirus system.
> http://www.eset.com
>
>



More information about the Support mailing list