buffer overflow in access log

[lastgreatone]

We've been running apache with no problem but recently we've been getting the following message occasionally:

[error] (233)No buffer space available: accept: (client socket)

Can anyone tell me exactly what the problem is (and maybe how to fix it?)

Thanks.

[Pedro Sousa]

Hi Francoise.
check this threads:
http://forums.itrc.hp.com/cm/QuestionAnswer/1,1150,0xf5a5e7613948d5118fef0090279cd0f9,00.html
http://forums.itrc.hp.com/cm/QuestionAnswer/1,1150,0xe0f8a12d6d27d5118fef0090279cd0f9,00.html
good luck.

[MANOJ SRIVASTAVA]

Hi Francoise

Essentially form the previous link .


HP ERROR 233 .
When a new connect request arricves at the local tcp and it is immediately followed by a RESET from the remote system , the server application is awakened twice by the request to perform ( accept ( ) calls ) . Each is returned bu ENOBUFS or HP Error 233. The problem is the acceptin server application should not be awakened twice . This was caused by mishandling the connection id in the TPI messages . The problem is fixed by correctly tracking the connection id between the socket and TCP layers.

There is a PATCH also which HP recommends loading PHNE_23456 to resolve this problem


Manoj Srivastava