TCP in CLOSE_WAIT with data in Send-Q (lots of them).

Hiya, I have a problem with one of my dedicated servers located at servermatrix.

Basically i keep getting "hangs" when accessing my websites, and others are experiencing problems too.

After some digging around there seems to be an issue with TCP connections going into a CLOSE_WAIT state without sending data back to the client.

Example netstat showing problem:
Code: 
tcp        1  13140 67.19.135.122:80        81.180.138.165:1154     CLOSE_WAIT
tcp        1  13140 67.19.135.122:80        81.180.138.165:1155     CLOSE_WAIT
tcp        1  12510 67.19.135.122:80        83.103.18.9:35230       CLOSE_WAIT
tcp        1  13140 67.19.135.122:80        81.180.138.165:1176     CLOSE_WAIT
tcp        1  13140 67.19.135.122:80        81.180.138.165:1172     CLOSE_WAIT
tcp        1  12312 67.19.135.122:80        64.80.224.162:7782      CLOSE_WAIT
tcp        1  13140 67.19.135.122:80        81.180.138.165:1150     CLOSE_WAIT
tcp        1  12312 67.19.135.122:80        192.132.24.81:62935     CLOSE_WAIT
About 10% - 15% of my netstat listing is like this with different IP addresses each time.

Apache is version 1.3.31 compiled with:
Code: 
./configure --enable-module=so --enable-module=rewrite --disable-module=userdir --disable-module=imap
Im running redhat kernel "2.4.21-15.0.3.EL", but thinking about an upgrade to 2.6.8.

Can anyone suggest what is causing this problem, and what can be done to resolve it?

 

 

 

 

Top