> But I still don't have an idea why you see that much      
> processes in keep-alive state...
> 
> Do you see any messages in your error log?          

No.  The server thinks that it is processing the request, it is just
not timed out.

> What are your settings in httpd.conf for KeepAliveTimeout
> MaxKeepAliveRequests etc.?

I've already sent these, here goes again:

Timeout 400
KeepAlive on
KeepAliveTimeout 15
MaxKeepAliveRequests 10
MinSpareServers 5
MaxSpareServers 10
StartServers 5
MaxClients 100
MaxRequestsPerChild 30

> Do you have any special settings for your tcp driver
> (e.g. do you use 'ndd' to tune /dev/tcp values)?

[EMAIL PROTECTED]:~/# grep ndd /etc/rc?.d/*
/etc/rc2.d/S69inet:ndd -set /dev/tcp tcp_old_urp_interpretation 1
/etc/rc2.d/S69inet:             ndd -set /dev/ip ip_forwarding 1
/etc/rc2.d/S69inet:             ndd -set /dev/ip ip_forwarding 0
/etc/rc2.d/S99httpd:# ndd -set /dev/tcp tcp_conn_req_max 32

> Your logfile directory (and thus the lockfile)           
> isn't located on an NFS mounted filesystem, isn't it?

No it is not.

> (Maybe you want to try our latest 1.3beta of Apache.)

I will when I can.  It is a production server (you remember, the problem
is only reproducible on a busy server).

I tend to think that the problem has something to do with signal processing,
maybe with a bug in signal processing in Solaris, and maybe it only reveals
on multiprocessor systems.  What makes me think this way is that if you
send SIGALRM to a sleeping process by hand, it does not notice it.  (But
if you sent SIGTERM it gracefully terminates).

Eugene

Reply via email to