a short time ago I posted the question below and I received the
responses also included below. I asked about sparc's halting when their dumb
terminal/term server consoles are power cycled or disconnected.
the suggestions range from a hardware solution to a kernel/driver patch.
I've chosen and tested the hardware solution.
A model 4273 "non-aborting serial console adapter" from NUData (32
Fairview Avenue, Little Silver, NJ 07739-1515 908.842.5757 fax: 908.842.1161)
for $89/ea. a little expensive, but they work.
Rob Aronson sent me his annex line config. useful and very straight
forward. I've included it below; hope he doesn't mind.
I asked my reseller about the kernel patches but they were, as usual,
useless. I plan to ask sun directly, when I have time.
I did not try any of the other suggestions which included tying pins
14 (ttyb's rxd, or was it txd) to 3 which worried me. and, putting a resistor
in...but, I don't know how many ohms or which pins and so I did not attempt
btw, I've noted this problem to exist on sparc 1+/2/LX/10 but not on
the 20. these are all that I've tested.
> From: Heas <firstname.lastname@example.org>
> To: email@example.com
> I want to use an annex (from xylogics, which is the same as the one
> which sun sells) for a remote console server for 4 sun sparc 2's, 10's and
> I can get it all set-up and working. the problem is that if the annex
> is rebooted or looses power, the sparcs halt just as if it were a dumb-terminal
> being used as a console.
> I recall someone mentioning what the cause was and a solution to the
> problem....but, I've forgotten. can anyone forward that note or information?
> From firstname.lastname@example.org Fri Jan 13 05:56:50 1995
> Hello Heas,
> Yes, there is a way you can solve this problem. Connect pin 14 to pin
> 3 in the RS232 connector on the Sun side. This will do the trick.
> -- Jos.
> Jos de Bruijne tel : +31 (0)3465 52888
> From email@example.com Fri Jan 13 07:59:21 1995
> Older machines require a pullup resister in the console cable. Sun fixed this problem at some point but I do not know which models are OK.
> Dave Russell
> From Rob.Aronson@fw.gs.com Fri Jan 13 09:36:30 1995
> I read your post with interest because we have about 10 Annex-3's installed,
> most are 32 port boxes but a few are 64 port servers. We have mostly Sparc 10's
> and 20's connected to them. I have on many occasions rebooted the Annex's without
> anything bad happening to any of the Suns connected to the particular terminal
> Powering off the Annex is another matter - I have seen that cause problems, but
> just rebooting the Annex shouldn't be a problem. There are two possibilities that
> you might try:
> 1 - perhaps your port configurations are causing the problem. I will email you
> a typical port config that we use. We set up our Annex ports based on the
> recommendations Xylogics tech support gave us after we encountered some
> problems after installing our first Annex-3.
> 2 - a company called NuData in Little Silver, NJ manufactures what are called
> "Sun non-aborting serial console adapters". I believe they're about $80
> a piece. We bought a bunch of them based on some testing I did with a sample.
> The testing I did was with a Wyse 50 terminal connected to a Sparc5. Without
> the adapters (we have christened them "dongles") a power-off of the terminal
> halts the Sun. With the dongle the Sun is unaffected by the power-off.
> However, when I tested these with the Annex they didn't work, i.e. we didn't
> seem to be able to talk to the Sun at all - the console seemed dead. I am
> currently working with my telecom people to find out what's wrong (it could
> be a problem with our premise wiring).
> Anyway, I am sure that Nudata will send you a sample to test with. If you want
> you can give Ron Seeley a call at (908) 842-5757 and I'm sure he can help you
> out, especially if you tell him I referred you.
> | Rob Aronson | Phone:(212) 902-2207 Fax:(212) 346-3729 |
> From Rob.Aronson@fw.gs.com Fri Jan 13 09:46:45 1995
> Here's a typical port config. I think the most important parms are in the
> first two sections. Most of the others are unimportant in this case.
> If there's anything else I can help you with let me know.
> --Boundary (ID 5qqebiosf8CdwcxWD5Ye+Q)
> admin : show port=1 all
> Port 1:
> Port Generic Parameters
> mode: slave location: ""
> type: hardwired term_var: ""
> prompt: "" speed: 9600
> data_bits: 8 stop_bits: 1
> parity: none max_session_count: 3
> allow_broadcast: Y broadcast_direction: port
> imask_7bits: Y cli_imask7: Y
> ps_history_buffer: 2000 banner: Y
> tcp_keepalive: 0 dedicated_address: 0.0.0.0
> dedicated_port: telnet type_of_modem: ""
> Flow Control and Signal Parameters
> control_lines: none input_flow_control: start/stop
> input_start_char: ^Q input_stop_char: ^S
> output_flow_control: start/stop output_start_char: ^Q
> output_stop_char: ^S input_buffer_size: 1
> bidirectional_modem: Y ixany_flow_control: N
> need_dsr: N
> Port Timers and Counters
> forwarding_timer: off forwarding_count: 0
> cli_inactivity: off inactivity_timer: 10
> input_is_activity: Y output_is_activity: N
> reset_idle_time_on: input long_break: Y
> short_break: Y
> Port Security Parameters
> user_name: "" cli_security: N
> connect_security: N port_server_security: N
> port_password: "<set>"
> CLI Line Editing Parameters
> attn_string: "" echo: Y
> telnet_escape: ^] telnet_crlf: N
> map_to_lower: N map_to_upper: N
> char_erase: Y line_erase: Y
> hardware_tabs: Y erase_char: ^?
> erase_word: ^W erase_line: ^U
> redisplay_line: ^R toggle_output: ^O
> newline_terminal: N
> Serial Networking Protocol Parameters
> local_address: 0.0.0.0 remote_address: 0.0.0.0
> dialup_addresses: N metric: 1
> slip_ppp_security: N do_compression: N
> allow_compression: N
> SLIP Parameters
> subnet_mask: 0.0.0.0 slip_load_dump_host: 0.0.0.0
> slip_allow_dump: Y slip_mtu_size: small
> slip_no_icmp: N slip_tos: N
> PPP Parameters
> ppp_mru: 1500 ppp_acm: 0x0
> ppp_security_protocol: none ppp_username_remote: ""
> ppp_password_remote: "<unset>" ppp_ncp: all
> --Boundary (ID 5qqebiosf8CdwcxWD5Ye+Q)--
> From firstname.lastname@example.org Fri Jan 13 10:21:55 1995
> When the sparcs see a break condition on that serial line you're using
> for a console, it's equivalent to the ol' L1-A keystroke.
> Sun Consulting has a zs driver (I think it's called zsbrk) which lets
> you redefine or eliminate this behavior. Somehow, I doubt it's free.
> If you're interested, drop me a note and I can help you get in touch.
> The other possibility is a cheap little circuit in the line which
> prevents the break condition from occurring. That one I don't have a
> copy of, but if you wind up with it, it'd be worth posting to
> Louis M. Brune 1 619 622 5119
> From email@example.com Fri Jan 13 13:37:49 1995
> Unfortunatley, I think the real fix belongs in the kernel, and for that
> you will need to speak to your Sun SE to see if you can get a modified tty
> driver - specifically zs_async.o if you're using 4.1.x. There is no
> official "patch" that I know of. If you discover another way to accomplish
> this, please let us know.
> Gary M. Blumenstein GREENWICH CAPITAL
> From firstname.lastname@example.org Fri Jan 13 13:55:32 1995
> There is a kernel patch that fixes this, according to the support people
> at Livingston Tech. Unfortunately, I can't seem to get them to tell me
> the patch number(s).
> I searched in SunSolve, but it brought up too much stuff to wade through.
> Please let me know (or just send a summary to the list) if you find it,
> since I need it, too.
> From email@example.com Mon Jan 16 09:45:58 1995
> NuData (908) 842-5757 sells an adapter device that will keep the signal
> up if the terminal server goes down.
> I did purchase one of these devices but have not had the chance to test it
> because I can't seem to get the correct settings on my terminal server.
> I was wondering if you could dump out the port setup screens for one of
> your console ports and send it to me. It would be much appreciated.
> Michael D. Reynolds /\ | \
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:15 CDT