|
|
Forum Tips
» This forum is for users to discuss their software problems and suggestions for improvement.
Remember that only a tiny percentage of our users experience such problems (To be precise less than 0.7% have ever posted on this forum)
» Include your software version number (as displayed in the About Box)
» Check the FAQ before posting as most common questions have already been answered
» You don't need to also e-mail us separately as our support team will respond to all questions on this forum
|
|
Author |
Topic |
|
jimbo69
USA
35 Posts |
Posted - Apr 21 2010 : 04:17:08
|
Revisited, I'm still on version 6.28, Windows XP service pack 2, disabled firewall Uninstalled antivirus. I've applied the tcp connection patch and upped open connections to 200. Don't know if it works or not though. Patch can be found here http://www.mydigitallife.info/2005/11/15/windows-xp-sp2-tcp-connection-limit-event-id-4226/
If I keep the simultaneous connections low failure rate is lower. but the timeout is 30 seconds minimum and this slows things way down. From reading this is an error generated internally by windows. smtp server Socket Error # 10004 Socket closed error (#10004): Interruption function call
Analyse:Windows Sockets code: WSAEINTR 10004
Description: Interrupted system call.
Solutions:A blocking operation was interrupted by a call to WSACancelBlockingCall.
WSACancelBlockingCall
The WSACancelBlockingCall function has been removed in compliance with the Windows Sockets 2 specification, revision 2.2.0.
The function is not exported directly by WS2_32.DLL and Windows Sockets 2 applications should not use this function. Windows Sockets 1.1 applications that call this function are still supported through the WINSOCK.DLL and WSOCK32.DLL.
Blocking hooks are generally used to keep a single-threaded GUI application responsive during calls to blocking functions. Instead of using blocking hooks, an applications should use a separate thread (separate from the main GUI thread) for network activity.[/b] Is it possible that my old version makes a call to a obsolete function? But happens more with increased threads. Strange. Thanks |
|
jimbo69
USA
35 Posts |
Posted - Apr 21 2010 : 04:56:35
|
I've been able to get the error messages down, Now I'm not so sure they are windows generated, although the article says they are, I'm also getting a lot of to many connection errors. Is there a way to randomize the send so it is not alphabetical or random by domain name? |
|
|
xequte
7504 Posts |
Posted - Apr 21 2010 : 13:16:50
|
Hi
MLK will generate the sending list using the current sort order, so you can "randomize" the sending list by sorting by "Member Date", for example.
Regarding the 10004 error are you sending via MLK's internal mail server or regular SMTP (via your ISP)?
With normal SMTP these are generally due to anti-virus software, such as Norton, intercepting your sending (where you need to disable the "worm blocking" option).
As you do not have an A/V product then your ISP could be doing something to prevent email overuse.
If you are using MLK's internal server then it is also quite possible that your sending is being rejected because your IP address is listed on a blacklist, such as Spamhaus.
Nigel Xequte Software www.xequte.com nigel@xequte.com
|
|
|
jimbo69
USA
35 Posts |
Posted - Apr 24 2010 : 10:38:06
|
Thanks no av software installed, I was using mlk's internal (to analyze the send information) server when you up the sending threads the error occurs more often, keeping them down below 40 helps a lot, but with a 30 second timeout built in needs threads about 100 to optimize sending time. But this also causes a bunch of errors talking to other mail servers when accessing the same domain (to many threads error). But will try your suggestion. Does the new version have the 30 second default send timeout restriction?
A question about mail servers in general, at the moment we are using merak lite older version but it is fast. I was looking at mercury32 and it has a endtoend smtp portion and a standard smtp portion. Is it possible to use the standard portion and disable end-to-end portion so I can always use mlk's internal engine. Or other mail servers that are fast that can co-exist with MLK. Probably will have to just test this. Hard on a production machine. But that's the only way to tell sometimes under heavy load.
Thanks for you help, I will be purchasing a upgrade, but not sure which one, I was looking at the database validation of the business and the domain validation would be nice, but think I can setup bounce information and email removal just as well. |
|
|
xequte
7504 Posts |
Posted - Apr 25 2010 : 02:12:35
|
Hi
If the timeout is too low then MLK may not have time to deliver the message. We can lower it if your requirements justify it.
MLK will work with any mail server that you like, both local and server based. The "standard" mode of Mercury/32 sends via an SMTP server, so IOW it is the same as using MLK's standard SMTP mode.
All MLK's editions offer the same validation options (we need to update the Comparison page).
Nigel Xequte Software www.xequte.com nigel@xequte.com
|
|
|
jimbo69
USA
35 Posts |
Posted - Apr 25 2010 : 09:32:05
|
Thanks for the info about validation, that's $100 less for us to upgrade, I will look into smtp, I thought MLK's engine would conflict. Yes 30 seconds is a little long for a single thread contacting to mail servers, they should respond within a second or 2, maybe 10 seconds at most.
Where still looking at a long time to validate via validation method, in production validation by processing bounces through MLK's internal engine would seem be be the most cost effecitve, and we can set bounce count at something like 5-10. |
|
|
xequte
7504 Posts |
Posted - Apr 26 2010 : 02:37:59
|
Hi
I absolutely agree validation is generally not a good way to use your time and resources.
The most reliable way to validate your addresses is just to send to your mailing list and then let MLK process the bounces and automatically remove the addresses that bounce too often.
If you do use the validation function make sure you use the Domain recommendation function, as regular validation is not very reliable.
Nigel Xequte Software www.xequte.com nigel@xequte.com
|
|
|
|
Topic |
|
|
|
| |