Time Saving Tips for problem reporting:
o The BSD/OS V2.0 and later release notes have a section on
"Problem Reporting Procedures", please refer to them for additional
information [everything in this message is in the release notes].
o Include your Customer ID (this is on the invoice and generally looks
like B9999, G9999, N9999) with the support request. If you don't know
customer ID then include other distinguishing information, e.g.,
company, postal address, and telephone information. If you don't
know it then just state that in the email and we will look it up and
give it to you for future use.
o Check the release notes for information relevant to your question.
o Include OS version information and what patches you have installed
(if you don't have all the patches installed please check the current
set carefully to see if something might be relevant to your problem).
Also, tell us if you are running a source or binary system. If multiple
systems are involved please include that information also (e.g., if
you are trying to dial into a SunOS 4.1.3 box from a BSD/OS V1.1 box
using uucico then it is helpful to know that). Basically, give us as
much context as possible.
Assuming you have support (if not, call 800-776-BSDI to sign up) and
that you are the support contact for your site, then you can get the
current patch information by sending email to patches@bsdi.com (``help''
will get you instructions). Remember, you must have registered your
email address with us before this will work. Contact us if you need
information about who is the support contact for your site.
To get the current list of patches use:
send index
If the system doesn't recognize you by your email address then you
can specify your customer ID like so:
cust G9999
send index
o If the item is really urgent then put `urgent: ' in the subject so we
can give it the appropriate priority.
o Send detailed problem reports and include the various configuration
files that might be related (see below for some tips on that). We
get many reports that don't have some of the basic facts in them and
that causes a lot of extra work for everyone.
If you can, run the software with debugging enabled to gather
extra information. The more information you give us, the more
likely we can solve your problem first try.
o Report the thing that actually failed first (sometimes failures
cascade and people report only a secondary failure, unless it is pretty
obvious then try to reproduce it a couple of times to make sure that
something else didn't fail first). Include the exact error messages
reported by the system, sometimes errors are very similar.
Use ``ktrace -i program args...; kdump'' to get additional
information if needed.
o Avoid making assumptions about the problem. The first part of the
problem report should be just the facts (``I typed this, the system
replied that, here are the configuration files, etc''). Then in the
second part you can explain what your thoughts, assumptions, and
conclusions are about the problem. Both parts are important but more
and more reports seem to tend towards containing only the latter.
o Explain in detail what you have done to try and solve the problem.
That way we can have a better idea of what it isn't.
o Include hardware configuration information (/var/db/dmesg.boot is
a good start but not always complete so check it over). Many
problems are nothing more than io port or interrupt conflicts.
Another source of frustration is the IO memory space. Each 128K
segment of IO memory must all be either 8- or 16-bits wide. 8-bit
cards will fail if you have a 16-bit card in that segment. The segments
are basically 0xa0000-0xbffff, 0xc0000-0xdffff, 0xe0000-0xfffff.
Another way this can bite you is say you have an 8-bit BIOS at
0xc0000 and a 16-bit ethernet card at 0xd0000. The 8-bit BIOS
will work during system boot because the ethernet card's memory
is disabled. Once you start the ethernet card however the BIOS will
no longer work, sometimes this is the cause of the "my system won't
reboot by itself" problem and sometimes more mysterious problems.
o TTY, Modem, login, dialout, uucp, etc, problems:
Include relevant portions of /etc/ttys, /etc/gettytab, /etc/remote,
UUCP config files. If processes are involved (e.g., hung or running
away) include the relevant portions of ``ps axlww'', ``pstat -t'',
``stty -a -f /dev/ttyxx'' (and please be sure to specify the
processes or ttys in question).
o SCSI problems:
These are almost always configuration problems. Hardware things to
check are cables and SCSI termination (only the *ends* should be
terminated, you should have exactly two terminators on the entire SCSI
bus). SCSI BIOS configuration options to try are: disable sync
negotiation, disable "fast SCSI", set the DMA transfer rate to 5.0MB/s,
make sure that if parity checking is enabled on your host adapter that
all devices have parity enabled.
Even if it's terminated "properly" you could still have termination
problems, try an active terminator. External devices often cause
problems and cables are very important for external devices, make sure
they are short and SCSI-II compliant. Total of all cables should be
less than 5 or 6 feet for best results, keep in mind that external
boxes have cables *inside* the box, sometimes 6 inches or more.
It's best if you can swap hardware around to try and isolate the
problem. Remove things to make the bus simpler.
Even if it works under DOS that doesn't mean there isn't a problem.
Different drivers do different things and multiuser, multitasking
systems make SCSI hardware work much harder.
o Network problems:
Include output from ``netstat -m'', ``netstat -rn'', ``netstat -ind'',
``netstat -s'', ``netstat -m'', ``vmstat -m'', and ``ifconfig xx0''.
And in many cases, tcpdump can be used to gather information about
failing network traffic. Also, ping and traceroute can be most helpful
[see also the release notes].
Send a description of the network load, e.g. how much PPP, SLIP, etc,
traffic, using which devices. Also, any ideas you might have on
correlation of system activity at the time the problem begins.
For SLIP or PPP problems send the configuration files
(/etc/ttys, /etc/ppp.sys, /etc/slip*).
o Routing Configuration Questions:
Be sure to included details about your entire network. What IP
addresses on which networks using what netmasks. Send an ``ifconfig
xx0'' for each interface plus a ``netstat -rn'' from each system
involved. Be sure to say which systems are running what software.
Make sure that ``options GATEWAY'' is enabled in the kernel
configuration (see release notes under "Gateways").
PROXY ARP
---------
If configuring a SLIP or PPP client using a network address inside
your ethernet's network you will need to configure Proxy ARP.
This can be done with the command:
arp -s 999.999.999.999 88:88:88:88:88:88:88 pub
Where 999.999.999.999 is the IP address of the connected machine
and 88:88:88:88:88:88:88 is the Ethernet MAC Addresss of the server.
[and when you have things working, put this command in /etc/netstart]
Get the ethernet address of the server with ``netstat -in''.
For example, given:
% netstat -ni
Name Mtu Network Address ...
tn0 1500 0.c0.f2.0.1e.c0 ...
^^^^^^^^^^^^^^^ ...
tn0 1500 240.124.135 240.124.135.1 ...
lo0 1536 ...
lo0 1536 127 127.0.0.1 ...
ppp0 1500 ...
ppp0 1500 240.124.135 240.124.135.2 ...
% pppconfig ppp0
ppp0: flags=9051
inet 240.124.135.2 --> 240.124.135.3 netmask 0xffffff00
^^^^^^^^^^^^^
You would run:
% arp -s 240.124.135.3 0:c0:f2:0:1e:c0 pub
o Kernel Hangs or Panics:
If you are reporting a kernel hang or panic then please read the
section "Kernel Debugging" in the release notes and submit that
information. In addition, a ``dmesg -M /var/crash/bsdcore.0''
is most helpful as is the /var/db/dmesg.boot file (which contains
your hardware configuration information).
--------------------------------------------------------------------------
US Customer Support Voice: +1 800 ITS BSD8 (487 2738)
Berkeley Software Design, Inc. Voice: +1 719 260 8114
7759 Delmonico Drive Voice: +1 512 219 5879
Colorado Springs, CO 80919 FAX: +1 719 598 4238
EMAIL: