TCPIP-databases

[Kjartan Konradsson]

Hello,
I'm have problem with cluster with 4 nodes, when they start up, they seam to sometimes get wrong tcpip$hostname. So I'm wondering about the tcpip-databases. Should they not be located on SYS$SPECIFIC for each node not on SYS$COMMON?
Hope to hear from some one.
Thanks
Kjartan

[John Gillings]

Kjartan,

I'm not sure what you mean by "tcpip$hostname" that's not a standard TCPIP logical name. Perhaps you mean TCPIP$INET_HOST?

Whatever, the TCPIP data base files are cluster aware and intended to be physically shared between nodes (ie: they should be in SYS$COMMON). The choice of host configuration should be determined by the SCS* SYSGEN parameters. If this isn't working correctly, you should log a case with your local customer support centre for proper diagnosis.

[Martin P.J. Zinser]

Hi Kjartan,

if we are talking about files that determine the identity of the host, modparams.dat should be in the specific tree, not in sys$common.
Also make sure you are not booting by chance into the wrong root of your system disk.

Greetings, Martin

[Ian Miller.]

the TCPIP datafiles are in sys$common:[sysexe] and are indexed by SCSNODENAME.
Whats the exact logical that is wrong?

[Kjartan Konradsson]

Hi!
The logicals are TCPIP$INET_HOST and TCPIP$INET_HOSTADDR. They seem to be ok at startup but after some time they have changed and are the same as another node in the cluster. I have 4 node cluster (GS140,GS60E and 2 ES80) and this seems to happen only between the ES80 nodes. I have checked the SCSNODE and SCSSYSTEMID and they differ on the nodes.
KK

[Wim Van den Wyngaert]

Kjartan,

Is the value of the nodename as seen in "ucx sho comm" also changed ? If so try a
search sys$startup:*:com new-host

Did you try a
search sys$startup:*.com TCPIP$INET_HOST
to see if nobody defined the logical e.g. during applic startup ?

Wim

[Lokesh_2]

Hi,

If they were OK when startup, then why and how they got changed afterwards?
Was someone doing any reconfiguration?
Are you using common system disk for all the nodes ?

Check the Audit file and operator.log if any clue is there .

Thanks & regards,
Lokesh

[Wim Van den Wyngaert]

http://64.233.161.104/search?q=cache:uGjPYXs5r4sJ:starlet.deltatel.ru/sys%24common/syshlp/tcpip53eco02.release_notes+TCPIP%24INET_HOST+release&hl=nl
and find TCPIP$INET_HOST

Do you use DHCP ? If not, what version of tcpip ?

Wim

[Kjartan Konradsson]

I'm using 5.3 ECO 2 of tcpip services. And under those conditions the local host name in UCX SHOW COMM is rigth.
I'm not using DHCP. As I saw in the notes there is a correction on this matter in ECO 2. But what I find strange is that it began after we installed our ES80's, we were replacing for old GS80 machines.

Kjartan

[Wim Van den Wyngaert]

On the es80 nodes :

dir sys$system:*tcp*.dat
Are all files in sys$common ?

Or check
dir sys$specific:[*]*tcp*
Are important files shown ?

Wim

[Kjartan Konradsson]

Hi

All the tcpip files are on sys$common and only TCPIP$SERVICES_LIST.DAT;1
TCPIP$SHUTDOWN_LIST.DAT;1
TCPIP$STARTUP_LIST.DAT;1
TCPIP$STARTUP_ORDER.DAT;1
are in sys$specific:

kjartan

[Wim Van den Wyngaert]

The files seem ok to me. The ones in specific are work files of startup/shutdown.

But the logicals are wrong.

Is ucx show interface giving the correct IP address ?

If yes, how was the gs replaced by the es ? Same name ? Same system root in cluster ?

Wim

[Kjartan Konradsson]

Hi,
ucx show inferf shows the rigth address when this happens. The swapping of the machines was done by just replacing the hardware. Same name same address.

Is it not any tool one can use to see what is in those tcpip databases?
I was rebooting one of my nodes yesterday and got an error in setting up the interfaces, it was trying to set up 2 though only 1 has ever been defined for this machine?

[Wim Van den Wyngaert]

KK,

I give up. As John suggested, open a case.

Wim

[Doug Phillips]

> Is it not any tool one can use to see what is in those tcpip databases?

Do you mean other than:

@sys$manager:tcpip$config

?

[Kjartan Konradsson]

Hi,
yes other than tcpip$config so one can see what is in the databases. How can one remove definitions for interfaces that are no longer valid for operations?

Kjartan

[Doug Phillips]

Via that same tcpip$config, select core then interfaces and go from there. That should work, no?

[Kjartan Konradsson]

Hi,
No, I have one node in my cluster, that is trying to initialize EWA1 that is not in the machine, when I start up tcpip, when I go into tcpip$config it does not ask me about this interface.

Kjartan

[Doug Phillips]

Stop or don't start TCPIP before you do tcpip$config; it really should rebuild the database properly when you run through the interface reconfigure option. If it doesn't, then I don't know what else to try.

I don't have v5.3 eco2 running anyplace so I can't try anything. Sorry.

[Kjartan Konradsson]

Hi,
Thanks for your input, I will use the next chance I get to stop and reconfig for each node to try get red of this profblem.

Kjartan

[Wim Van den Wyngaert]

KK,

To remove the interface use :
$ set config nointerface * /confirm
(or without config if only in the running version)

Wim

[Jan van den Ende]

Kjartan,

for THIS thread it is off-topic, but your thread
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=690881
you have closed.

It is considered 'civil" to close a thread if you are satisfied that it is solved,
BUT,
the rules of ITRC specify that a thread is considered 'solved' (and gets marked by a rabbit) IF, and ONLY IF, at least one of the answers is awarded at least 8 points.

And we in the VMS forum have been heralded as the forum with the highest percentage of rabbits, and we strive to stay on top.
So, please, before you CLOSE a thread, award at least one 8, 9, or 10.

(and of course this goes for any poster)

Sorry if this sounds a bit as a rant, that is definity NOT my intention!!

Happy foruming!


Jan

[Wim Van den Wyngaert]

Jan,

Unless he was expecting someone that says "I've got that situation and it works". In that case I would close it too without giving 8+.

Wim

[Kjartan Konradsson]

Sorry, i had forgotten this thread, because it didn´t solve my problem. I´m in the same situation every time I reboot. I will soon upgrade to 7.3-2 and then i will use the time to reconfigure the network.

Thanks

[Kjartan Konradsson]

Forgot to close..