You are correct Sandy, it doesn't, but it can cause a surge in
processes running. My way of thinking, why run the test if you don't
need to?
Absolutely. That's why -sw exists. Hey, maybe it'll be thread-based,
rather than process-based, someday -- that'd be so non-Unix-y, though.
Allow me to make a correction - under normal daytime loads, we are seeing an
increase in CPU - steady at 80-90%. Under 3.0.5.5 we saw an average of 40%
Sincerely,
Randy Armbrecht
Global Web Solutions, Inc.
804-346-5300 x112
877-800-GLOBAL (4562) x112
http://globalweb.net
-Original
Randy try changing
WAITFORMAIL 1500
To
WAITFORMAIL 500
See if that changes the delay.
David B
www.declude.com
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of GlobalWeb.net
Webmaster
Sent: Monday, October 24, 2005 7:16 PM
To:
David,
Do you have an exhaustive explanation of what exactly each of
the new commands does? I have seen a tremendous amount of emails to
alter this and alter that, however, what does it actually do and how
does one affect the other? Thanks for the aid.
Keith
-Original Message-
David
If one doesn't have the following in the declude.cfg, what are the default
values used by Declude?
WAITFORMAIL
WAITFORTHREADS
WAITBETWEENTHREADS
Thanks,
John
---
This E-mail came from the Declude.JunkMail mailing list. To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type
I changed the THREADS to 75 and CPU wnet down to about avg 50%. I'll try
WAITFORMAIL in just a bit.
I also noticed in my spool folder I have a new folder called s that
appeared with the time stamp of the new proc install. There is nothing in
it; just a blank folder. I deleted it.
Sincerely,
I've changed WAITFORMAIL to 500 and this seemed to help a bit more. I am
still seeing a collection of about 50 messages before proc engages and
processes them to the work folder...
Sincerely,
Randy Armbrecht
Global Web Solutions, Inc.
804-346-5300 x112
877-800-GLOBAL (4562) x112
Okay, I'm watching these discussions - as I'm still truly scared of
upgrading Declude.
I just can't figure out the logic behind all the CPU consumption and why
certain process parameters are suspected to have detrimental effects.
Please verify my thinking:
A machine that needs to process a
Hi,
I have a question for those of you that are using SmarterMail.
We are looking at the software to determine how we can link it to our user
creation process.
Currently, we have both classic asp (ASP not ASP.NET) and php scripts that
create users by POSTing to a form. The back-end of the form
My sentiments exactly. Well said.
Dan
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Andy Schmidt
Sent: Tuesday, October 25, 2005 12:33 PM
To: Declude.JunkMail@declude.com
Subject: RE: [Declude.JunkMail] 30511 problem
Okay, I'm watching these
It seems to me that something changed in the code
If all I do is change from V9 to V11 and I see a noticeable difference, then
it has to be caused by a change in the code from V9 to V11. No other
hardware or software was changed!!!
I have gotten direction to change settings, however that in no
Hi Erik,
Well, I'm actually all in favor of the service architecture - I just can't
comprehend how the service architecture would cause the same workload of
data to apparently cause significant increases in prolonged CPU usage?
What worries me is that the solution is tweaking of the process
Robert,
I have not programmed to SmarterMail API's, but I did look at them some time
ago. As I recall, the API is just a COM object, which you can call from ASP.
I don't believe ASP.NET is required for the management end.
Mike
- Original Message -
From: Robert E. Spivack [EMAIL
In every instance that we have observed so far the headers in the body are
caused by broken mail clients. This is not only an issue for Declude but for
mail servers as well. To illustrates the difficulty coming up with a single
algorithm that will detect all instances of these broken emails to
Randy,
Can you confirm for me the processes that are increasing the CPU ?
David B
www.declude.com
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of GlobalWeb.net
Webmaster
Sent: Tuesday, October 25, 2005 8:46 AM
To: Declude.JunkMail@declude.com
Subject:
Keith,
WAITFORMAIL3
Defined in milliseconds eg. 3 = 30 seconds this can be changed to set
the wait time that decludeproc will wait before checking the \proc directory
once empty for new messages.
WAITFORTHREADS 1500
Defined in milliseconds eg. 1500 = 1.5 seconds this can
John,
Default Values: (But my opinion is to use the WAITFORMAIL and change it to
5000 for single or 500 for dual proc)
WAITFORMAIL3
WAITFORTHREADS 1500
WAITBETWEENTHREADS 1
David B
www.declude.com
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL
This is correct behavior, remember once the \proc hits zero the WAITFORMAIL
kicks in, sometimes because this often happens quicker than what we can see.
You could even drop the WAITFORMAIL lower if you like.
David B
www.declude.com
-Original Message-
From: [EMAIL PROTECTED]
Decludeproc and queuemgr
Randy
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of David Barker
Sent: Tuesday, October 25, 2005 3:30 PM
To: Declude.JunkMail@declude.com
Subject: RE: [Declude.JunkMail] 30511 problem
Randy,
Can you confirm for me the
Thanks, David. I appreciate this information and your other contributions to
the list.
John
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of David Barker
Sent: Tuesday, October 25, 2005 2:34 PM
To: Declude.JunkMail@declude.com
Subject: RE:
Hmmm... if queuemanager CPU is increasing, then that would point to more
messages being processed, wouldn't it? So it would logically follow that
decludeproc should see more CPU usage as well.
Do you have reports to show the incoming and outgoing message load on the
server, both before and after
Something changed at the time the new architecture came out. Either folks
started using a lot of new / broken clients, or Declude became more
sensitive.
I say that based upon observation. We NEVER noticed the issue prior to
installing the beta of the latest version. Heck, it is possible that
There has been a rash of new spam/virus content with broken headers...
starting last month. We're on IMail 8.05 and Declude 1.82 and we've seen
them.
Darin.
- Original Message -
From: Robert Grosshandler [EMAIL PROTECTED]
To: Declude.JunkMail@declude.com
Sent: Tuesday, October 25,
No change in message load - I randomly picked approx 7pm yesterday to
upgrade.
Decludeproc service ran at a lower CPU at version 3.0.5.5
Sincerely,
Randy Armbrecht
Global Web Solutions, Inc.
804-346-5300 x112
877-800-GLOBAL (4562) x112
http://globalweb.net
-Original Message-
From:
Andy, don't be scared I can hold you hand :)
How do I explain the apparent inconsistency where users report that the
CPU consumption is considerable higher than in the prior version?
First we need to confirm that it is decludeproc that has increased in CPU
usage. So far I have not seen any
I have repeatedly switch back and forth between the two version (V9 and V11)
I assure that I have changed nothing else and that cpu usage is higher on
the V11
I watch V9 cpu usage
I go from V9 to V11 and cpu usage increases
Back to V9: it goes down
Back to V11 it goes up.
How else does one
Erik,
1. Because Imail 8.2x made changes to their product making Declude
incompatible with Imail.
2. There's been more and more incremental releases are you asking for less
incremental releases ?
3. Declude 3.0 has made many problems that were not noticble with prior
versions of Declude
Hmmm? You said queuemanager increased CPU usage as well. Were you just
meaning that only decludeproc and queuemanager were using CPU on the server?
Darin.
- Original Message -
From: GlobalWeb.net Webmaster [EMAIL PROTECTED]
To: Declude.JunkMail@declude.com
Sent: Tuesday, October 25,
Nothing except what I had posted earlier in the code has changed. And the
changes should not have effected the CPU in anyway.
Unless we can replicate the issue it becomes very difficult to resolve. If
you can send me any data that shows what you are seeing, we can then try
replicate the problem.
I recomend getting some stats/data on the topic.
Use performance monitor/log the data with v9 for a few hours, then with v11.
This would give them spikes and clear data as to what is going on.
I advise that you monitor cpu usage, along with declude proc, declude queue
process, imail
At 02:47 PM 10/25/2005 -0500, you wrote:
Something changed at the time the new architecture came out. Either folks
started using a lot of new / broken clients, or Declude became more
sensitive.
In my opinion a lot of broken clients started happening. Whereas I beta
tested the new version on my
Decludeproc runs at a higher CPU with v11 as compared to v5 - queuemgr also
runs a it higher. These 2 services are the ones that most consumes CPU with
v11
Randy
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Darin Cox
Sent: Tuesday, October 25, 2005
David,
Thanks for the explanation!
However, I have yet to see a message with SMTP headers in the body if I
remove Declude and send mail directly into Exchange.
Maybe Exchange will just not deliver?
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
Since I've tweaked settings, CPU is down to almost same levels as before; it
was the default settings that were causing higher CPU. At the default
settings declude was between 10-35% and queuemgr was 10-25%
POP3 and SMTP werer less than 10% combined. inVurl, sniffer and others made
up rest.
Hi Dave:
Thanks.G
My email was based on your responses, suggesting the tweaking the various
new process settings would address the problems that some people reported
with respect to CPU usage.
My logic was, that any tweaking could only effect the usage profile not
the overall consumption (if
But Queuemanager is IMail, not Declude. Why would it run higher unless
there was more load on the server (i.e. more messages being processed)?
Darin.
- Original Message -
From: GlobalWeb.net Webmaster [EMAIL PROTECTED]
To: Declude.JunkMail@declude.com
Sent: Tuesday, October 25, 2005
Hmm, thanks Sandy, that's interesting.
Let me make sure I understand this. I use the same command line options for
SPAMC32 as I already have, including the -e and -et? I then create 10
external SA tests for scores of 1 through 10. JM looks for exact scores
(which -e gives from rounding SA
I have seen the same thing.
Decludeproc crash while im on the server.
Running 3.05.05
Mark Smith wrote:
I'm not 100% sure if people aren't noticing Decludeproc crashing.
The installer sets the service recovery to restart in 0 seconds.
So, would a few of you mind checking your System event
Let me make sure I understand this. I use the same command line
options for SPAMC32 as I already have, including the -e and -et? I
then create 10 external SA tests for scores of 1 through 10. JM
looks for exact scores (which -e gives from rounding SA scores) from
1 to 10 from SA
You need not wonder how these guys can hit you with a new dictionary
attacking IP every 30 seconds from now until eternity...
'Bot herders' may have controlled 1.5 million PCs
http://news.com.com/Bot+herders+may+have+controlled+1.5+million+PCs/2100-7350_3-5906896.html?tag=cd.top
Matt
Nothing in my logs.
When these events occur, you need to do the following:
Check for DrWatson logs.
Check for files in the proc\review directory
Check the logs for that time period.
What do the above 3 indicate?
John T
eServices For You
-Original Message-
From: [EMAIL PROTECTED]
If I dump the files in the proc\review folder back into the proc folder they
process ok.
I've given up and went back to 2.06.x
Immediately I'm processing 3-4 times faster with less CPU time and no
crashing.
Can't continue to run 3.0 anymore so I'll wait until the bugs are worked
out
42 matches
Mail list logo
