Will
I had the same problem a month or so ago. My issue turned out to be my firewall was preventing the mail server from going out. I have my DNS servers behind the firewall also. When this happens declude seems to hang on to some memory and it slowly consumes resourses. I allowed the mail server to get out and the problem went away. Check your firewall logs for your mailservers ip address and see if you're getting blocked. I tried the winsockcleanup and that did not solve the problem. This problem came up with the upgrade to declude  3.X , the mail server and firewall have been running with the same settings for years with no problems. Sadly my logs were deleted before I realized this had solved my problem and I did not get the port blocked nor the destination ip address.

John
 
 -----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]On Behalf Of Will
Sent: Tuesday, January 17, 2006 6:22 AM
To: [email protected]
Subject: [Declude.JunkMail] decludeproc causing dns queries to fail

For the second time in a month I have come to work to find over 500,000 messages in my proc folder.  The cause is declude’s inability to perform dns queries, thus mail backlogs.  The reason declude cannot perform DNS queries is that no outgoing traffic is being permitted on port 53.  The reason no outgoing traffic is being permitted on port 53 is that decludeproc has consumed the resources allocated to my network interface.  The fix for this is relatively simple.  I stop the decludeproc service and restart it and all is well.  

 

After the first occurrence, it was suggested that I enable winsockcleanup documented as the following at declude.com:

WINSOCKCLEANUP    ON
Located in Declude.cfg. Some customers had issues related to their network stack causing loss of functionality for basic network operations. The default for this directive is OFF

Now I’ve been fooled twice, so I guess shame on me.  Does anyone have a suggestion that would help resolve this?  In the meantime I will be writing a script to monitor the number of message in the proc folder and if it reaches a certain threshold, will restart the decludeproc service.  I really don’t care for workarounds such as this, but I’ve gained a bad image over these two occurrences and I cannot allow it to happen again.

 

Declude.cfg:

THREADS                   100

WAITFORMAIL                     5000

WINSOCKCLEANUP           ON

Reply via email to