not working Advanced Server Pathwork after openVMS autogen and reboot

[CJBe]

Hi!

After a autogen with reboot on a three node cluster the Advanced Server (PathWork) is unable to startup on one of the cluster nodes.

In PWRK$LMMCP_NODENAME.LOG i found:

LISTER\listeninit: Unable to post the name on any networl: NODENAME

PCS_SetConnectionLevel: group name has not been declared

and after that

SYSTEM job terminated at ...

Is there a problem with the network or where can I look up next? Please help me, I am a "fresh" openVMS Admin.

Thanks!

[Shilpa K]

Hi,

It is likely that the pwrk$knbdaemon process is not starting up correctly. You may want to check if there are any errors in the log file PWRK$ROOT:[LOGS]PWRK$KNBDAEMON_.log;*

Regards,
Shilpa

[CJBe]

Thanks!

In PWRK$KNBDAEMON_NODENAME.LOG I found:

get_ip_addr: gethostname error: 0

I looked in TCPIP show host, there is a entry

IP-Adress Nodename

I can do ping, telnet usw. looks like network is on...

[Shilpa K]

Hi,

I am assuming that you are using TCPIP services provided HP. If so, can you execute the following command and see if the inet logicals are defined correctly?

$ sh log *inet*

Regards,
Shilpa

[CJBe]

Thanks!

I used your command to look up the logicals on both nodes from the cluster.

On the one where Advanced Server still working I found these logicals in LNM$SYSTEM_TABLE

"TCPIP$INET_DOMAIN" = "..."
"TCPIP$INET_HOST" = "..."
"TCPIP$INET_HOSTADDR" = "..."
"UCXIP$INET_DOMAIN" = "..."
"UCXIP$INET_HOST" = "..."
"UCXIP$INET_HOSTADDR" = "..."

On the node where Advanced Server isn't working the logicals are missing.

So please can You tell me, where and when these logicals are set up. Thanks a lot!!!

[marsh_1]

hi,

usually these would be defined after you have configured your tcpip setup with tcpip$config.com and started tcpip services.

fwiw

[Shilpa K]

Hi,

Looks like TCPIP is not configured correctly. As mentioned by Mark Hunt, you can use tcpip$config to reconfigure TCPIP.

Regards,
Shilpa

[CJBe]

I defined the logicals on the node where Advanced Server isn't running by hand. But it is still the same Problem and the same entrys in the logfiles...

Okay, so realy a problem with TCP/IP?

But when I say TCPIP show service I get Information. Also I can ping the node and connect to this via Telnet. Also TCPIP show interface gets the correct interface configuration...

[marsh_1]

hi,

can't remember if there can be case sensitivity issues with the host name with this, have you an upper and lower case nodename (alias) in the host table to tie in with the logicals ?

[CJBe]

The entrys in the host table all are like this:

IP NODENAME, nodename

and the logicals are once defined:

TCPIP$.... = "nodename"
UCX$... = "NODENAME"


I looked up the Processes running on both nodes an found an missing entry on the node where advanced server isn't running.

The missing process is TCPIP$PWIP_ACP, I looked up with TCPIP$CONFIG.COM and tried to stop and start but still missing... Is it possible this is the problem?

[marsh_1]

hi,

is there anything in the sys$manager:tcpip$pwipacp_'nodename'.log ?

[CJBe]

Yes, there are three lines:

getLocalAddr : getaddrinfo failure: 4
interfaceinit : could not getLocalAddr
main : PWIPdriver, PWIPacp interface init failure, abortin PWIP ACP startup

[Shilpa K]

Hi,

One of the versions of TCPIP release notes mentions that PWIP will not start if bind resolver is disabled though this is appicable if IPV6 is being used as per the release notes. So you may want to verify if bind resolver is disabled on your system. If so enable it and then try starting PWIP driver.

ECO 3 updates
-------------
ECO H 06-JUL-2004 Alpha and VAX and I64

Problem:

The PWIPacp process will not start if the bind resolver is
disabled:

|$ type sys$Manager:TCPIP$PWIPACP_KEOKA.LOG
|Wed Jun 30 13:27:54 2004: getLocalAddr : getaddrinfo failure, ipv6: 4
|Wed Jun 30 13:27:54 2004: interfaceInit : Could not getLocalAddr!
|Wed Jun 30 13:27:54 2004: main : PWIPdriver, PWIPacp interface init
|failure, aborting PWIP ACP startup...

Also, instead of manually defining inet logicals, I think itâ s a good idea to reconfigure TCPIP

Regards,
Shilpa

[Shilpa K]

Also, you may want to check this thread on the ITRC:

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1322118

Regards,
Shilpa

[CJBe]

Okay, so on. I look at the thread you send me.

When I say UCX SHOW NAME I get:

BIND RESOLVER PARAMETERS

Local domain: * Mismatch *

System

State: Started, Disabled
Transport: UDP
Domain: xxx <- same as on all nodes
...
Servers: xxx, xxx <- same as on all nodes
...

Process:

State: Disabled

I looked at the domain set up with @tcpip$config.com. It's the same like on all nodes on the network.

[marsh_1]

hi,

are all the systems at the same revision of tcpip services - tcpip show version ?

[CJBe]

Yes, all System are on
HP TCP/IP Service for OpenVMS Alpha Version V5.6 - ECO 2

[Shilpa K]

Hi,

The following thread seems to discuss a similar problem of Local Domain: *mismatch*.

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1184820

Regards,
Shilpa

[Shilpa K]

Can you check if the logical TCPIP$BIND_DOMAIN is defined:

$ sh log TCPIP$BIND_DOMAIN

Sorry I missed that you have already defined the domain.

Regards,
Shilpa

[CJBe]

No Problem, there are these three entrys. In each is the correct domain xxx.xxx in the last the entry is in Uppercases.

(LNM$SYSTEM_TABLE)

"TCPIP$BIND_DOMAIN" = "xxx.xxx"
"TCPIP$INET_DOMAIN" = "xxx.xxx"
"UCX$INET_DOMAIN" = "XXX.XXX"