Title: Logging dies with Client Auth.

I'm running FW-1 4.1 SP1 on AIX.  We're trying to start using Partially Automatic Client Authentication for HTTP and HTTPS traffic.  After testing this for 15-20 minutes, logging on the firewall seemed to stop and some new processes were spawned:

# ps -ef
     UID   PID  PPID   C    STIME    TTY  TIME CMD
    root     1     0   0   Jun 07      -  1:38 /etc/init
    root  2682     1   0   Jun 07      -  0:00 /usr/lib/errdemon
    root  2874     1   0   Jun 07      -  0:00 /usr/sbin/srcmstr
    root  3208     1   0   Jun 07      -  0:54 /usr/sbin/cron
    root  3390     1   0   Jun 07      - 128:17 /usr/sbin/syncd 60
    root  3734     1   0   Aug 18      0  0:00 /usr/sbin/getty /dev/console
    root  3936     1   0   Aug 18      - 18:01 fwm
    root  4672  4954   0 22:39:09      -  0:00 funcchain au_auth 2 1 resolver_
    root  4954  9784   0 22:39:09      -  0:00 funcchain au_auth 1 0 resolver_
    root  5184  2874   0   Jun 07      -  0:00 /usr/sbin/portmap
    root  5420  2874   0   Jun 07      -  0:00 /usr/sbin/inetd
    root  5678  2874   0   Jun 07      -  3:08 /usr/sbin/named
    root  5936  2874   0   Jun 07      - 41:53 /usr/sbin/xntpd
    root  6196     1   0   Jun 07      -  0:00 /usr/sbin/uprintfd
    root  6452  2874   0   Jun 07      -  0:04 /usr/sbin/qdaemon
    root  6714  2874   0   Jun 07      -  0:00 /usr/sbin/writesrv
    root  6968     1   0   Jun 07      -  1:54 /usr/bin/AIXPowerMgtDaemon
    root  7766  7998   0   Jun 07      - 12:11 /usr/lpp/adsm/bin/dsmstat 39321
    root  7998     1   0   Jun 07      - 40:45 /usr/lpp/adsm/bin/dsmc sched -p
    root  8256     1   0   Jun 07      -  0:00 /usr/lpp/diagnostics/bin/diagd
    root  8456  9626   0 23:09:18  pts/0  0:00 -ksh
    root  8862  3936   0 23:15:38      -  0:00 /usr/lpp/CPfw1-41/bin/fwm 8 -lo
    root  9030  2874   0   Jun 12      -  0:14 /usr/sbin/syslogd
    root  9626  5420   0 23:09:18      -  0:00 telnetd -a
    root  9784     1   0   Aug 18      - 235:26 fwd
    root 10204  9784   0 22:34:42      -  0:08 in.ahttpd 80
    root 10830  8456   3 23:18:10  pts/0  0:00 ps -ef
    root 11108     1   0   Jun 20      - 1684:20 alertd -A -l
    root 11302  9784   0   Aug 18      - 58:44 alertd -A -l
    root 11990 11108  46                  0:00 <defunct>
    root 12170  9784   0   Aug 18      - 45:34 isakmpd
    root 12398  9784   0 20:08:49      -  0:04 in.atelnetd 23
    root 12668 11302  32                  0:00 <defunct>
    root 12850  9784   0   Aug 31      - 26:45 in.aftpd 21

The processes are the "funcchain" processes.  A lookup at checkpoint found a possible cause:

---------------------------------
Solution: funcchain process turns to Zombie and hangs the HTTP security server (10022.0.1908062.2483196)
Upgrade to FireWall-1 4.0 SP6

Problem Description
funcchain process turns to Zombie and hangs the HTTP security server
HTTP security server hangs
HTTP security server stops logging
 
Problem Environment:
FireWall-1 4.0 SP5
HP-UX 10.20
BG000517
 
Cause of this problem:
When a process dies, the system send a SIGCHLD signal to inform the father process about this event. The child process stays in a zombie state until the father intercepts this signal. In this case, the in.ahttpd process did not try to intercept this signal from its child - funcchain - which stays in zombie state and causes the HTTP security server hang.

 
---------------------------------

Well, all that is fine and dandy, but the solution is unacceptable because I don't want to upgrade FW-1 4.1 SP1 to FW-1 4.0 SP6.  I had enough trouble getting a 4.1 license!  ;-)

Anybody got any ideas?  I already checked PhoneBoy's site and found no answer.  Could this be an AIX problem?  Checkpoint's site didn't even mention AIX in the problem environment.

I really don't want this to fail when we roll it out but the pressure from above is beginning to squash me!

Thanks in advance,

Paul Mills
Data Security Analyst
CCSA, CCSE
//AMERICREDIT CORPORATION
[EMAIL PROTECTED]

#42:  It works the way the Wang did, what's the problem?

Reply via email to