Dir 5.1 on Solaris 8 simply crashes

Question

Our 5.1 DS crashed twice this day without putting any information in the access and error logfiles.
Also /var/adm/messages showed nothing.
When I run /distr/idir/5.1/bin/slapd/server/idsktune
I get the following output below.
=======output idstune===============
Sun ONE Directory Server system tuning analysis version 22-MAY-2002. Copyright 2002 Sun Microsystems, Inc.
NOTICE : System is usparc-SUNW,Ultra-60-solaris5.8_s28s_u7wos_08a (1 processor).
NOTICE : Patch 108919-14 (CDE 1.4: dtlogin patch) is not installed.
NOTICE : Patch 108949-07 (CDE 1.4: libDtHelp/libDtSvc patch) is not installed.
NOTICE : Patch 109470-02 (CDE 1.4: Actions Patch) is not installed.
NOTICE : Patch 110943-01 (SunOS 5.8: /usr/bin/tcsh patch) is not installed.
NOTICE : Patch 111548-01 (SunOS 5.8: catman, man, whatis, apropos and makewhatis patch) is not installed.
NOTICE : Patch 111596-02 (SunOS 5.8: /usr/lib/netsvc/yp/rpc.yppasswdd patch) is not installed.
NOTICE : Patch 111626-01 (OpenWindows 3.6.2: Xview Patch) is not installed.
NOTICE : Patch 111879-01 (SunOS 5.8: Solaris Product Registry patch SUNWwsr) is not installed.
NOTICE : Patch 108921-13 (CDE 1.4: dtwm patch (required for JRE 1.4)) is not installed.
NOTICE : Patch 108773-12 is present, but 108773-14 (SunOS 5.8: IIIM and X Input & Output Method patch required for JRE 1.4) is a more recent version.
NOTICE : Patch 112003-02 (SunOS 5.8: Unable to load fontset in 64-bit Solaris 8 iso-1 or iso-15) is not installed.
NOTICE : Solaris patches can be obtained from http://sunsolve.sun.com or your Solaris support representative. Solaris patches listed as required by the JRE are located at http://www.sun.com/software/solaris/jre/download.html or can be obtained from your Solaris support representative.
WARNING: There are only 1024 file descriptors available, which limits the number of simultaneous connections. Additional file descriptors, up to 65536, are available by adding to /etc/system lines like set rlim_fd_max=4096
set rlim_fd_cur=4096
NOTICE : The tcp_conn_req_max_q value is currently 128, which will limit the value of listen backlog which can be configured. It can be raised by adding to /etc/init.d/inetinit, after any adb command, a line similar to: ndd -set /dev/tcp tcp_conn_req_max_q 1024
NOTICE : The tcp_rexmit_interval_initial is currently set to 1500 milliseconds (1 seconds). This may cause excessive retransmission on the Internet. However, if clients are using Solaris 2.5.1 or earlier on the LAN, increasing this value can result in packet loss due to a bug in that version of Solaris. Upgrading any Solaris clients to 2.6 or later is recommended.
NOTICE : The NDD tcp_ip_abort_cinterval is currently set to 180000 milliseconds (180 seconds). This may cause long delays in establishing outgoing connections if the destination server is down.
NOTICE : If the directory service is intended only for LAN or private high-speed WAN environment, this interval can be reduced by adding to
/etc/init.d/inetinit:
ndd -set /dev/tcp tcp_ip_abort_cinterval 10000
NOTICE : The NDD tcp_ip_abort_interval is currently set to 180000 milliseconds (180 seconds). This may cause long delays in detecting connection failure if the destination server is down.
NOTICE : If the directory service is intended only for LAN or private high-speed WAN environment, this interval can be reduced by adding to
/etc/init.d/inetinit:
ndd -set /dev/tcp tcp_ip_abort_interval 60000
NOTICE : The NDD tcp_smallest_anon_port is currently 32768. This allows a maximum of 32768 simultaneous connections. More ports can be made available by adding a line to /etc/init.d/inetinit: ndd -set /dev/tcp tcp_smallest_anon_port 8192
WARNING: tcp_deferred_ack_interval is currently 100 milliseconds. This will cause Solaris to insert artificial delays in the LDAP protocol. It should be reduced during load testing. This line can be added to the /etc/init.d/inetinit file: ndd -set /dev/tcp tcp_deferred_ack_interval 5
======end output============
Could the missing patches or the low amount of file descriptors be the reason ?
The DS ran two weeks without any problems.
FYI on the server three DS instances are running.
The server is anE220R and it has 1 Gb memory, 2 Gb swap and 1 CPU.
Any help would be nice......

Answer 1

I checked the error and accesslogfile.
Nothing there only a gap in time !!!!
You only see something in the error logfile during the startup again :
[09/Apr/2003:13:00:21 +0200] - iPlanet-Directory/5.1 Service Pack 1 20020606 B2002.161.2234 starting up
[09/Apr/2003:13:00:21 +0200] - Detected Disorderly Shutdown last time Directory Server was running, recovering database.
[09/Apr/2003:13:00:43 +0200] - slapd started. Listening on all interfaces port 2389 for LDAP requests
[09/Apr/2003:13:03:31 +0200] - cos_cache_getref: no cos cache created