Lookout 4.0.1 behaviour vs Lookout 4.0

[Ken McWilliam]

More differences in behaviour between 4.0 (build 37) and 4.0.1 (build
51) have been noticed under NT4.0 (SP4):

Processes running under Lookout 4.0.1 build 51 have annoying flickering
of hypertrends seemingly above a certain size. It is particularly
noticeable on slower machines (e.g. Pentium 133). Shrinking them
sometimes helps, sometimes not; it seems to depend on what else you've
done in the mean time.

We have tried the same process file on the same machine, first running
4.0 then 4.0.l. The flicker only occurs on 4.0.1. We then reverted to
4.0 and the flicker went away. We reinstalled 4.0.1 and it reappeared.
We also notice a difference in CPU usage between the two versions.

In trying to come up with a degenerat
e process file to demonstrate the
problem I found what may be a clue:

Boot PC into NT: Screen resolution is 1024x780 65k colours 2MB video
memory.
Clean out the Citadel database files
Start Lookout 4.0.1.51
Create New Process
Create a Waveform object with default parameters
Create a Hypertrend
Add a group (default params)
Add an item: Waveform1.cosine_180
Edit database: Log Waveform1.cosine_180 at 0.1 deviation
Hypertrend looks fine so far
Add another item: Waveform1.sine_180
FORGET TO LOG IT TO CITADEL!
Hypertrend now flickers
Edit database and log Waveform1.sine_180 at 0.1 deviation
Hypertrend stops flickering
..... the same happened each time a hypertrend tries to display a
non-logged value...
.... unless you shrink it!

Fine, I thought, maybe we have a non-logged trace - but we don't. Maybe
the flickering happens when hypertrend can't get all the data it wants
from Citadel (for whatever reason) on 4.0.1.51

Hmmm... any ideas?

[John Dunlop]

Hi,

Heres a blinder... Found it while demonstrating a new system to a customer.
We were not amused!

Lookout 4.0.1.51 under NT4.0SP6

We have a Pot linked to the trendwidth of a hypertrend. In 5.0.1 if you
move the pot the CPU usage of Citadel goes through the roof and then eventually
crashes. Dr Watson pops up to tell me I need more memory. Hmm.

Doesn't happen in 4.0
Whats going on?

John

[Rajko Eichhorn]

I tried it here with my version of Lookout 4.0.1 (51) and NT SP 5 it didn't
happen

My advise: never trust Microsofts even number service packs ((:--

Regards

Rajko


"John Dunlop" schrieb im Newsbeitrag
news:38bd993a@newsgroups.ni.com...
>
> Hi,
>
> Heres a blinder... Found it while demonstrating a new system to a
customer.
> We were not amused!
>
> Lookout 4.0.1.51 under NT4.0SP6
>
> We have a Pot linked to the trendwidth of a hypertrend. In 5.0.1 if you
> move the pot the CPU usage of Citadel goes through the roof and then
eventually
> crashes. Dr Watson pops up to tell me I need more memory. Hmm.
>
> Doesn't happen in 4.0
> Whats going on?
>
> John

[Khalid]

Hello John,

I don't see this behavior on my NT4 Sp6a. A few things to check:

1> have you used the right Min, Max, and Resolution settings for the Pot?
They need to be in Time format. E.g., 0:10 for Min, 1:00:00 for Max, and
0:01 for Resolution.

2> Service Pack 6 was updated to 6a right after it was released. Do you
have 6a?

3> When you upgraded to 4.0.1.51, did you uninstall 4.0? And also made
sure there weren't any remnants from 4.0.. especially in the System32 folder.
Look for liles starting with 'lk'.

Regards,

Khalid


--------------------
"John Dunlop" wrote:

Hi, Heres a blinder... Found it while demonstrating a new system to a customer.
We were not amused! Lookout 4.0.1.51 under NT4.0SP6.

We have a Pot linked
to the trendwidth of a hypertrend. In 5.0.1 if you
move the pot the CPU usage of Citadel goes through the roof and then eventually
crashes. Dr Watson pops up to tell me I need more memory.

Hmm. Doesn't happen in 4.0
Whats going on?

John

[John Dunlop]

Hi Khalid,

Thanks for that. I didnt have 6a just 6.

A bit more information... I formatted the drive, reinstalled NT and sp6 and
THEN Lookout 4.0.1 and it works fine.

I think what may have happened is another person had been installing software
on the same machine and messing with ActiveX. Do you think this might have
messed up the registry?

Anyway, problem solved. I now love Lookout again 🙂

John


"Khalid Ansari" wrote:
>>Hello John,>>I don't see this behavior on my NT4 Sp6a. A few things to
check: >>1> have you used the right Min, Max, and Resolution settings for
the Pot?> They need to be in Time format. E.g., 0:10 for Min, 1:00:00 for
Max, and>0:01 for Resolution. >>2> Service Pack 6 was updated to 6a right
after it was
released. Do you>have 6a? >>3> When you upgraded to 4.0.1.51,
did you uninstall 4.0? And also made>sure there weren't any remnants from
4.0.. especially in the System32 folder.> Look for liles starting with 'lk'.
>>Regards,>>Khalid>>>-------------------->"John Dunlop"
wrote:>>Hi, Heres a blinder... Found it while demonstrating a new system
to a customer.>We were not amused! Lookout 4.0.1.51 under NT4.0SP6. >>We
have a Pot linked to the trendwidth of a hypertrend. In 5.0.1 if you>move
the pot the CPU usage of Citadel goes through the roof and then eventually>crashes.
Dr Watson pops up to tell me I need more memory. >>Hmm. Doesn't happen
in 4.0>Whats going on?>>John

[Ken McWilliam]

We have resolved this problem of flashing hypertrends - we had a hard to
spot **un-logged** trace in our hypertrend defintions.

Our hypertrends no longer flash no matter what size they are.

The guys at NI were very helpful during the investigations - many thanks.

Regards

(a much happier) Ken

Ken McWilliam wrote:

> More differences in behaviour between 4.0 (build 37) and 4.0.1 (build
> 51) have been noticed under NT4.0 (SP4):
>
> Processes running under Lookout 4.0.1 build 51 have annoying flickering
> of hypertrends seemingly above a certain size. It is particularly
> noticeable on slower machines (e.g. Pentium 133). Shrinking them
> sometimes helps, sometimes not; it seems to depend on what else you've
> done in the mean time.
>
> We have tried the same process file on the same machine, first running
> 4.0 then 4.0.l. The flicker only occurs on 4.0.1. We then reverted to
> 4.0 and the flicker went away. We reinstalled 4.0.1 and it reappeared.
> We also notice a difference in CPU usage between the two versions.
>
> In trying to come up with a degenerate process file to demonstrate the
> problem I found what may be a clue:
>
> Boot PC into NT: Screen resolution is 1024x780 65k colours 2MB video
> memory.
> Clean out the Citadel database files
> Start Lookout 4.0.1.51
> Create New Process
> Create a Waveform object with default parameters
> Create a Hypertrend
> Add a group (default params)
> Add an item: Waveform1.cosine_180
> Edit database: Log Waveform1.cosine_180 at 0.1 deviation
> Hypertrend looks fine so far
> Add another item: Waveform1.sine_180
> FORGET TO LOG IT TO CITADEL!
> Hypertrend now flickers
> Edit database and log Waveform1.sine_180 at 0.1 deviation
> Hypertrend stops flickering
> .... the same happened each time a hypertrend tries to display a
> non-logged value...
> ... unless you shrink it!
>
> Fine, I thought, maybe we have a non-logged trace - but we don't. Maybe
> the flickering happens when hypertrend can't get all the data it wants
> from Citadel (for whatever reason) on 4.0.1.51
>
> Hmmm... any ideas?