[pptp-server] High-Uptime with pptpd 1.0.1

[Joshua M. Schmidlkofer]

Howdy,

  I have been testing a  redhat 7.1 <-> Win2k sp2 PPTP vpn.  It has 
relatively low traffic [save for big db updates (50-100/meg at once)], 
and is actually hooked up across a 100Mbps lan currently.   I have been 
using it to test the robustness of MySQL's current 'replication' 
abilities.   Various tests have gone well both for pptp & MySQL.   
However, I have been testing the 'high-uptime' portion of things.   
After about 10 or 12 days of operating I came in to day, and had a load 
of the following messages in my syslog:

<snip>
May 26 23:29:42 widmers pptpd[980]: CTRL: Unexpected control message 0 
in disconnect sequence
May 26 23:29:42 widmers pptpd[980]: CTRL: EOF or bad error reading ctrl 
packet length.
May 26 23:29:42 widmers pptpd[980]: CTRL: couldn't read packet header (exit)
May 26 23:29:42 widmers pptpd[980]: CTRL: Unexpected control message 0 
in disconnect sequence
May 26 23:29:42 widmers pptpd[980]: CTRL: EOF or bad error reading ctrl 
packet length.
May 26 23:29:42 widmers pptpd[980]: CTRL: couldn't read packet header (exit)
<snip>

When I say a lot, I mean 779 Meg.  I rushed to clear out some junk 
before taking a serious look into where my space went.   It is just my 
desktop so I have 'var' on my root volume.  Anyway, the VPN was still 
running just fine, and everything was ok, but pptpctl was running at 
like 70% cpu time, and had accumulated about 2000+ minutes of cpu time. 
 ???  Anyway, I have no idea what would cause this, but anyone know?


Redhat 7.1
Vanilla Kernel 2.4.4 w/MPPE patches.
Downloaded pppd from ... whatever that site is. [I am running 2.4.1] 
w/ppp-2.4.1-openssl-0.9.6-mppe-patch
pptpd 1.0.1



thanks,
  Joshua