24 Mar
2003
24 Mar
'03
12:02 p.m.
Hi,
I have hooked up 2 VAX6k and 1 HSC90 and one MTI StingRay CI-SCSI
adapter to my two star couplers all properly. Still I could never
get any of my CI stuff to work. I tried under ULTRIX 4.5 just to
ping through the SCS (?) IP physical/data-link layer but to no
avail. Don't get any errors configuring the interface and pinging,
just no host reachable. I may not have the CI nodes configured
right, I may need to figure out how to plug the jumpers on the VAXBI
backplane.
Two additional observations: my HSC90 reports K.ci status 111 and the
CI LINK module tells me it is not happy. Also, when I try to boot
any of the VAXen from media that I don't think is really bootable,
I get some funny error on the console: "Insufficient memory for CI"
(even though I have full boat of memory 0.5 GB in each VAX.) Could
everyone be unhappy about something strange in the CI system?
Otherwise:
- who knows how the CIBCA is configured with the jumpers?
- who knows how best to debug CIBCA under Ultrix or VMS?
- who knows how to set up (and debug!) IP over CI on Ultrix?
(I do not dare ask about IP over CI on VMS, but appreciate if
there are answers.)
I also have a VAX6000 Diagnostics tape (TK50) which I should be
using. I tried booting from it but I get this "insufficient memory
for CI" error soon after the first few blocks are read from tape.
I also didn't see much in the way of VMB.EXE on that tape. Anyone
know what to do with a VAX6k diagnostics tape? Do I need to use it
under VMS? (PS: if someone needs VAX6k diagnostics, I could help :-)
regards,
-Gunther
24 Mar
24 Mar
11:44 a.m.
could never get any of my CI stuff to work. I tried
under
ULTRIX 4.5 just to ping through the SCS (?) IP
physical/data-link layer but to no avail. Don't get any
errors configuring the interface and pinging, just no host
Is this supposed to work? I have no idea; I do know that it was
possible to use DECnet over CI under OpenVMS, but I don't recall
ever seeing that as a possibility for the VMS IP stack. Is
Ultrix supposed to be able to do this?
reachable. I may not have the CI nodes configured
right, I
may need to figure out how to plug the jumpers on the VAXBI backplane.
According to my notes, I have a CIBCA manual somewhere. It's not
scanned, so that means it may take me some time to find it
(otherwise it would have been scanned by now!). Hopefully someone
will dig one out soon, otherwise I'll have ahunt around at the
weekend.
Two additional observations: my HSC90 reports K.ci
status 111
and the CI LINK module tells me it is not happy. Also, when I
try to boot any of the VAXen from media that I
don't think is
really bootable, I get some funny error on the console:
"Insufficient memory for CI" (even though I have full boat of
memory 0.5 GB in each VAX.) Could everyone be unhappy about
something strange in the CI system?
Only two hits in google, one is you from a while ago, the other
is this:
http://mail-index.netbsd.org/port-vax/1999/08/20/0004.html
If your message looks like that (i.e. it starts %BOOT-F)
then it is either VMB or BOOT58 that is whining at you.
So it's not Ultrix, but the stage before that which
tries to get an OS started.
If I cannot find the CIBCA manual, I'll see if I can find
anything else that might have relevant info in. I do have
an HSC9x Installation Manual scan somewhere and maybe some
other HSC9x bits and bobs.
One thing I recall is that some of the later adapters were
capable of extra smarts, like 4K packets and being able to
use both CI paths simultaneously. The CIBCA is probably old
enough that it cannot do this. My recollection is that, for
most things, the newer stuff would "fall back" to the previous
mode and all would work (albeit suboptimally) but I'm sure there
were some settings that needed to be consistent throughout the
setup. Right now the details elude me. However, it might be worth
trying just the two VAXen in the cluster to start with (i.e. pull
out the StingRay and the HSC95) and see if the VAXen can see each
other at all. Naturally, I'd boot OpenVMS and cluster them to
prove it works, but I suspect you'd prefer a different approach :-)
Obviously, if your only attached drives are connected over the CI,
then you may have to think of something else!
Antonio
--
---------------
Antonio Carlini arcarlini(a)iee.org
12:36 p.m.
On Mon, 24 Mar 2003, Antonio Carlini wrote:
Yes, Ultrix can run IP over CI.
could never
get any of my CI stuff to work. I tried under
ULTRIX 4.5 just to ping through the SCS (?) IP
physical/data-link layer but to no avail. Don't get any
errors configuring the interface and pinging, just no host
Is this supposed to work? I have no idea; I do know that it was
possible to use DECnet over CI under OpenVMS, but I don't recall
ever seeing that as a possibility for the VMS IP stack. Is
Ultrix supposed to be able to do this? Right now the details elude me. However, it might be
worth
trying just the two VAXen in the cluster to start with (i.e. pull
out the StingRay and the HSC95) and see if the VAXen can see each
other at all. Naturally, I'd boot OpenVMS and cluster them to
prove it works, but I suspect you'd prefer a different approach :-)
Obviously, if your only attached drives are connected over the CI,
then you may have to think of something else!
Minimizing the configuration is always a good start. I'd start with one
VAX, and the HSC95. The HSC can give some diagnostics and information of
the CI which can be useful.
One suggestion is to make sure that all nodes have different addresses,
and that all are under 16, unless they all are configured for a 32-node
CI.
Other obvious things are to make sure the cables really are connected
correctly. TX to TX and RX to RX, channel A and B separate, and so on.
Also, while Gunther don't mention it, I hope (and assume) that there is a
star coupler in the middle.
Johnny
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt(a)update.uu.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
3:17 p.m.
There are a number of show commands in the crash (8) debugger which deal
with SCS, MSCP and the like. Might be worth looking at.
On Mon, 24 Mar 2003, Gunther Schadow wrote:
Hi,
I get some funny error on the console: "Insufficient memory for CI"
(even though I have full boat of memory 0.5 GB in each VAX.) Could
everyone be unhappy about something strange in the CI system?
--
7787
days inactive
7787
days old
3 comments
4 participants
participants (4)
-
arcarlini＠iee.org -
bqt＠update.uu.se -
gschadow＠regenstrief.org -
thompson＠new.rr.com