how to test /configure a nic on alphastation255 vms6.21h3

[Christian Burschka]

Dear experts,

my running system (alpha with vms6.2) died.

I put the systemdisk dka0: into a similar hardware (onboard ewa0, second nic ewb0) formerly running with unix. The system boots, but ucx_startup crashes, last message "ucx-i-definte defining interfaces".
When I remove the second nic, ucx starts ok, network seems ok.
With second nic I can see no problem with "show device" or "lancp" for ewb0. how should I distinguish between a hardware defect and a configuration problem?

Thanks a lot for any hints.

[Karl Rohwedder]

Christian,

have you tried @SYS$MANAGER:UCX$CONFIG to configure the 2nd interface?
Do you really want to use it, else dismount it :-)

regards Kalle

[Christian Burschka]

Kalle,

thanks for answering.
I definitely need the second nic for communication to a diffraktometer. ;-)

The disk dka0: should have correct parameters, the (e)prom not certainly. Does Ucx$config write parameters to disk or to (e)prom ?

Christian

[Karl Rohwedder]

Don't know how a EPROM comes into the field?
Do you mean changing the MAC-address (like DECnet does)?

UCX$CONFIG just configures 2 ethernet devices, e.g. WE0 and WE1 to be used, it doesn't write into any eeroms's, as fas as I know. Should work without any problems.

LANCP SHO DEV EWB0/COU doesn't show any problems (the famous Fast/slow/halfduplex mismatch thing)?

regards Kalle

[Steven Schweda]

Assuming that "(e)prom" refers to the SRM
console variables, then no, UCX$CONFIG.COM
does not affect them (so far as I know).

You might look at them, however, to see if
you like the settings:

>>>show e*

Or perhaps:

>>>show e* | more

I have not run a VMS (or UCX?) version so
old on an Alpha, so I can't say much about
possible software problems.

"UCX SHOW VERSION" could be interesting, but
I too would vote for trying UCX$CONFIG.COM
to configure the interfaces.

More actual error messages might be helpful.

[Christian Burschka]

Kalle,

sorry, I meant "srom"s. No reason to change the mac adress.

According to Your hint I tried UCX$config also on ewb0 and accomplished options 1 - 4,

(there isnt really much to configure and the test 7 is said to run only with internet started)

5: shutdown digital tcpip
%system-w-nosuch device no such device available (??? something wrong ??)
ucx$ntpd server stopped
ucx$metric server stopped

but: I obviously cannot (option 6)
"startup Digital TCPIP/services for OpenVMS"
when I try , the whole system freezes, just the same as when I start ucx$startup.

Should I run any other configuration.com ?
like autoconfigure in sys$config etc?

Could it be that there are unsuitable parameters in the "ethernet ID ROM" on the motherboard or anywhere else (because it was run under unix before) ??

thanks and regards,

Christian.

[Steven Schweda]

> Could it be that there are unsuitable
> parameters in the "ethernet ID ROM" on the
> motherboard or anywhere else (because it
> was run under unix before) ??

Not because it was running Tru64 before. I
have run Tru64 and VMS on AlpSta 200 4/233
and XP1000 systems with no problems. (Well,
no _network_ problems. They do fight over
how the ToY clock is set.)

The on-board Ethernet on that system is only
10MHz, right? What are you using for the
second card? (Is it supported by VMS?)

[Christian Burschka]

Steven,

I have two (apparently?) identical cards, one from the old machine, originally purchased from DEC. (the chip is 21041-AA DEC 1994).

I cannot exclude a hardware defect on both.
I used one with 10M ethernet, which was quite normal. Right, LANCP says 10M for the working onboard ehternet, (the same for ewb0).

UCX show version "TCPIP services 4.0"

How can I delete ewb0 from the tcpip-database ?

How should one normally proceed, when first installing a new card ? (Ok, I know, read the manual....)

Thanks for answering, regards

Christian.

[Karl Rohwedder]

Not quite sure about UCX V4.0 (that is rather old), but it should be
UCX> SET CONF NOINTER WE1 ! permament
UCX> SET NOINTER WE1 ! volatile

regards Kalle

[Steven Schweda]

> I have two (apparently?) identical cards,
> one from the old machine, originally
> purchased from DEC. (the chip is 21041-AA
> DEC 1994).

Sounds like a DE450, which, if it is one,
should work (and may even be supported).

> How can I delete ewb0 from the
> tcpip-database ?

I'd expect TCPIP$CONFIG.COM do do that.

> How should one normally proceed, when first
> installing a new card ? (Ok, I know, read
> the manual....)

Yes, of course, but I believe that the manual
says to run TCPIP$CONFIG.COM, which has
always worked for me. (But I haven't
installed anything older than a DE500-BA
since I put an ISA card into an AlpSta 200
4/233.)

[David B Sneddon]

Christian,

I recall a restriction with the Alphastation 255
and similar boxes with regard to the placement
of ethernet and/or video cards. There was one
slot into which they should not be placed due
to some conflict... the manual/documentation
or a Google search may reveal the details (or
that my memory ist nicht so gut).

Dave

[David B Sneddon · Accepted solution]

Here is what I found...

The OpenVMS Alpha operating system does not support PCI option cards configured in PCI slot 0 on any AlphaStation 255 series systems.

PCI slot 0 is the lowest physical PCI option slot on AlphaStation 255 series systems. The interrupt signal for this slot is shared with the built-in Ethernet port. Because the OpenVMS Alpha operating system does not currently permit PCI devices to share an interrupt line, a PCI device installed in slot 0 will not function correctly or may cause errors to occur with the built-in Ethernet port. As a result of this restriction, AlphaStation 255 series systems support a maximum of two PCI option cards, configured in slot 1 and slot 2.

May or may not be relevant.

Dave

[Karl Rohwedder]

Christian,

check here:
http://h18002.www1.hp.com/alphaserver/workstations/retired/a255series.html

The SOC says, that only slot 6/7 (the upper ones) are for PCI cards, while the Combo slot is not supported under VMS.

regards Kalle

[Steven Schweda]

Yup.

http://h18002.www1.hp.com/alphaserver/download/ek-vllxa-ui-b01.pdf

The lowest PCI slot (the combination slot) shares its interrupts with the
embedded Ethernet controller. To use a PCI option in this slot, you must be
using an operating system that supports shared interrupts (that is, Windows NT
or Digital UNIX*). The OpenVMS operating system does not support shared
interrupts.
____________________________________________________________

[Christian Burschka]

Thats it!
*Jumpingaroundlaughingbuyyouallabeer*

David, Kalle , Steven,
thanks a lot for Your valuable help!!

changed only the slot, nothing else
@ucx$startup works.
ivp test works
telnet works
ftp works

(by the way, is there secure ftp or something
like ssh for vms? )

Best regards,

Christian

[Ian Miller.]

For the vintage of UCX you get with VMS V8.2 then to get ssh etc have a look at

http://www.process.com/tcpip/ssh.html

[Christian Burschka]

Dear experts,

thank You all for Your valuable comments.

We changed the system from unix to vms,
without Your information I wouldnt have cared about the slot for the second nic.

Regards,

Chris