VCTC between first and 2nd level VM systems
Jean Snow
This has been discussed here before, but I don't think anyone ever
posted back how they got this to work. The systems in question
here are VM/ESA 1.2.2 (prod) and 2.1.0 (test), and TCP V2.3 at 9601 level
(so all the apars I've found on IBMLINK are on TCP).
I promise to report back on this, if I ever get it working...
prod TCP can only be seriously messed with in the wee hours, so
I'd *really* appreciate any help you can offer.
Here's my configuration:
Id of test system is V210TEST, ctc addresses on both sides are
620 and 621:
-- in prod TCPIP directory, add SPECIAL 620 CTCA V210TEST
SPECIAL 621 CTCA V210TEST
-- in first level directory for V210TEST, add the same specials
connecting to TCPIP
-- in 2nd level directory, DEDICATE 620 and 621 to TCPIP
-- in production PROFILE TCPIP, add
DEVICE CTCDPRD CTC 620
LINK CTCLPRD CTC 1 CTCDPRD
HOME 128.192.1.5 CTCLPRD <added after existing home stmt,
using ipaddr of production system>
GATEWAY 198.137.20 = CTCLPRD .... <added following existing
gateway stmts, including defaultnet>
-- in test PROFILE TCPIP, add
DEVICE CTCDTST CTC 620
LINK CTCLTST CTC 0 CTCDTST
HOME 198.137.20.5 CTCLTST <this is only home stmt>
GATEWAY 128.192. = CTCLTST.... <these are only gateway stmts>
DEFAULTNET 128.192.1.5 CTCLTST....
START CTCDTST
To start the CTC, from production TCP, I do:
CP COUPLE 620 V210TEST 621
CP COUPLE 621 V210TEST 620
<via OBEYFILE> START CTCDPRD
And this is what happens:
CTCA initializing: <=== on prod system when I start
Device CTCDPRD: the CTC
Type: CTC, Status: Not started
Envelope queue size: 0
Address: 0620
DASD 03EE DETACHED
CTCA device CTCDPRD: Unexpected CSW from SenseId command on device 0621
Keys: E1, CcwAddress: 001AFF30 (Deferred SIO CC 1)
Status bits: C0, SCFA: 4411
Unit Status: 90, Channel Status: 00
Byte Count: 1
CTCA shutting down:
Device CTCDPRD:
Type: CTC, Status: Sense ID on input
Envelope queue size: 0
Address: 0620
<=== prod netstat shows
netstat cp q v 620-621
VM TCP/IP Netstat V2R3
CP command output is:
CTCA 0620 COUPLED TO V210TEST 0621 SUBCHANNEL = 0002
CTCA 0621 COUPLED TO V210TEST 0620 SUBCHANNEL = 0003
CP return code = 0
Ready; T=0.05/0.07 08:32:24
netstat devlinks
VM TCP/IP Netstat V2R3
Device LCS1 Type: LCS Status: Ready
Queue size: 0 Address: 0920
Link ETH1 Type: ETHERNET Net number: 1
Device CTCDPRD Type: CTC Status: Inactive
Queue size: 0 Address: 0620
Link CTCLPRD Type: CTC Net number: 1
Ready; T=0.04/0.06 06:33:13
netstat gate
VM TCP/IP Netstat V2R3
Known gateways:
NetAddress FirstHop Link Pkt Sz Subnet Mask Subnet Value
---------- -------- ------ ------ ----------- ------------
Default 128.192.1.1 ETH1 1500 <none>
128.192.0.0 <direct> ETH1 1500 <none>
<other gates>
198.137.20.0 <direct> CTCLPRD 1500 <none>
And on test system, I get:
<=== test startup
TCPIP AT ESATEST VIA RSCS 08/08/96 06:27:53 EDT THURSDAY
<...>
CTCA initializing:
Device CTCDTST:
Type: CTC, Status: Not started
Envelope queue size: 0
Address: 0620
<...>
Sense id: FF 30 88 for device 0620
Sense id: FF 30 88 for device 0621
<=== test NETSTAT shows
netstat devlinks
VM TCP/IP Netstat V2R3
Device CTCDTST Type: CTC Status: Ready
Queue size: 0 Address: 0620
Link CTCLTST Type: CTC Net number: 0
Ready; T=0.06/0.09 08:30:54
netstat gate
VM TCP/IP Netstat V2R3
Known gateways:
NetAddress FirstHop Link Pkt Sz Subnet Mask Subnet Value
---------- -------- ------ ------ ----------- ------------
Default 128.192.1.5 CTCLTST 1500 <none>
128.192.0.0 <direct> CTCLTST 1500 <none>
Ready; T=0.06/0.08 08:30:59
Any ideas on what's wrong with this picture?
** Jean Snow UGA Systems JE...@UGA.CC.UGA.EDU **
** University of Georgia, Athens, GA **
Rick Barlow
On 8 Aug, Jean Snow wrote:
>Date: Thu, 8 Aug 1996 09:50:11 EDT
>From: Jean Snow <JE...@UGA.CC.UGA.EDU>
>Subject: VCTC between first and 2nd level VM systems
I have a working CTC between my first and second level VM systems for TCP/IP
testing. My experience has been that virtual CTCs don't always work if the
couple command is not issued before the second level VM system attaches the
CTC to a virtual machine. Therefore, I include the COUPLE commands in the
PROFILE EXEC of my VM guest (V210TEST for you). The alternative is to make
sure that the first level CP COUPLE commands are issued before the TCPIP
VSM is logged on to the second level guest.
*
Jean Snow
Rick, thanks for your reply: I'm really glad to hear that someone does
have this working.
On Fri, 9 Aug 1996 07:46:32 EDT Rick Barlow said:
> My experience has been that virtual CTCs don't always work if the
>couple command is not issued before the second level VM system attaches the
>CTC to a virtual machine.
OK, made sure the couples were done first by taking down and reloading
test TCPIP while CTCs were coupled. This did not seem to help, but
I noticed a couple of things I missed before:
-- the two ends of the CTCs do seem to be aware of each other: when
I started the prod end, the test end got
CTCA device CTCDTST: Unit Check on device 0620
CTCA device CTCDTST: Started sense on device 0620. SIOCC 0
CTCA device CTCDTST: Sense data for device 0620: 41
CTCA device CTCDTST: Received DEVICEend interrupt on device 0620 -- probable
CTCA restart
CTCA device CTCDTST: Received DEVICEend interrupt on device 0621 -- probable
CTCA restart
and when I took down the test end, prod got:
CTCA device CTCDPRD: Unit Check on device 0620
CTCA device CTCDPRD: Started sense on device 0620. SIOCC 0
CTCA device CTCDPRD: Sense data for device 0620: 41
And I also noticed these messages, which come up on both sides (not
at startup, but at time to time thereafter while the link is started):
Error: Possible configuration error in link statement, retrying
I don't think I left it up long enough to see these yesterday. I have
no clue: the link statements look ok to me:
Prod: Test:
DEVICE CTCDPRD CTC 620 DEVICE CTCTTST CTC 620
LINK CTCLPRD CTC 1 CTCDPRD LINK CTCLTST CTC 0 CTCDTST
<START CTCDPRD> <START CTCDTST>
Does this give anyone any ideas? Thanks.....
Jean
** Jean Snow UGA Systems/Postmaster JE...@UGA.CC.UGA.EDU **
** University of Georgia, Athens, GA JEAN@UGA (Bitnet) **
Shmuel Naparstek
> This has been discussed here before, but I don't think anyone ever
> posted back how they got this to work. The systems in question
> here are VM/ESA 1.2.2 (prod) and 2.1.0 (test), and TCP V2.3 at 9601 level
> (so all the apars I've found on IBMLINK are on TCP).
> I promise to report back on this, if I ever get it working...
> prod TCP can only be seriously messed with in the wee hours, so
> I'd *really* appreciate any help you can offer.
>
> Here's my configuration:
> Id of test system is V210TEST, ctc addresses on both sides are
> 620 and 621:
> -- in prod TCPIP directory, add SPECIAL 620 CTCA V210TEST
> SPECIAL 621 CTCA V210TEST
This your mistake: You must define the 620 and 621 as
SPECIAL 620 3088 V210TEST etc
If you define a virtual ctca it is not Good for TCPIP!!!!
> -- in first level directory for V210TEST, add the same specials
> connecting to TCPIP
> -- in 2nd level directory, DEDICATE 620 and 621 to TCPIP
>
> -- in production PROFILE TCPIP, add
> LINK CTCLPRD CTC 1 CTCDPRD
> using ipaddr of production system>
> GATEWAY 198.137.20 = CTCLPRD .... <added following existing
> gateway stmts, including defaultnet>
>
> -- in test PROFILE TCPIP, add
> DEVICE CTCDTST CTC 620
> ** University of Georgia, Athens, GA **
Marc9
I've had a similar problem with TCP/IP for VM and VCTC's and tried 3088's
(rather than CTCs) as a solution. Unfortunately, after it didn't work, I
traced VM source code and found that the code to define a virtual 3088 and
the code to define a virtual CTC is the same. Did I miss something? Is
there a technical difference between DEFINE 3088 and DEFINE CTC on a
VM/ESA 1.2.2 system?
Thanks!
-- Marc.