Unable to create cmclconfig.ascii in HP-UX in SG CLUSTERING

[drteam]

kaveri:root-/>cmquerycl -C /etc/cmcluster/cmclconfig.ascii -n krishna -n kaveri

Warning: Unable to determine local domain name for Kaveri
Error: Node krishna is refusing Serviceguard communication.
Please make sure that the proper security access is configured on node
krishna through either file-based access (pre-A.11.16 version) or role-based
access (version A.11.16 or higher) and/or that the host name lookup
on node krishna resolves the IP address correctly.
Failed to gather configuration information.
kaveri:root-/>

[Ralf Seefeldt]

Hello drteam,

do you use a cluster lockdisk in your cmcl-configuration (only possible with a 2-node-cluster)?
If so, please check, whether the node krishna has access to it and refuses access from the other node?

Maybe krishna used vgahcnage -a y.

To give further hints, the colusterconfig-files would be very helpfull.

Bye
Ralf

[Stephen Doud]

1) If this is a new cluster creation (version A.11.16 or later), insure
/etc/cmcluster/cmclnodelist has been created on every node in the cluster.
The file format is akin to .rhosts. Example:

tic root
toc root

or use IP-based handles to expedite the authentication function. Eg:

16.113.32.42 root
16.113.32.43 root

NOTE: EVERY node must be listed in every cmclnodelist!

---
2) The error may occur because an old cluster binary file exists on the other
node and there is no cluster binary on the node where the configuration
command was executed. Removing the invalid cluster binary file will permit
the cmapplyconf to operate.

---
3) The identd service may not be enabled in /etc/inetd.conf.
# grep ident /etc/inetd.conf
#auth stream tcp6 wait bin /usr/lbin/identd identd

Removing the "#" comment and running 'inetd -c' clears the condition.

---
4) Insure /etc/hosts identifies every NIC that bears an IP with the hostname
of the server. Newer versions of Serviceguard depend on 'identd' to do a
"call-ID"-like lookup on the node requesting the operation of a
Serviceguard command - even the local node.
For that reason, every line in /etc/hosts that refers to a NIC on a
cluster node must bear the hostname.

Example:

123.45.67.89 helix.alf.cpqcorp.net helix # hostname
123.45.67.90 helix-hb.alf.cpqcorp.net helix-hb helix # heartbeat
123.45.67.91 helixbkup.alf.cpqcorp.net helixbkup helix # backup LAN

123.45.67.92 torus.alf.cpqcorp.net torus # hostname
123.45.67.93 torus-hb.alf.cpqcorp.net torus-hb torus # heartbeat
123.45.67.94 torusbkup.alf.cpqcorp.net torusbkup torus # backup LAN

Note the hostname is appended to every line.

[flower_2]

hi stephen and ralf,

I am one of member of drteam thanks for your inputs i am going to start new installation of service gaurd.If u have any documents for MC service guard clustering pls give me the link.it will help in installing the Cluster.

[Ralf Seefeldt]

Hello flower,

! have the following links or documentreferences about MCSG (I hope, the links are still up-to-date):

• Date : 2000 Oct 20
  
  Description : ServiceGuard Quick Reference
  
  Document Id : UXSGKBAN00000033
  
  Search String : UXSGKBAN00000033
  
  I don't have the link


• How do I test MC/ServiceGuard?
  
  DocId: UXSGKBAN00000471 Updated: 20020221


• Technical Documentation for High Availability


• Managing MC/ServiceGuard, HP Part Number:B3936-90065, March 2002


• Managing MC/ServiceGuard, HP Part Number:B3936-90024, January 1998


• High availability clustering information library


• HP Serviceguard Extension for RAC for Virtualization


• Designing Disaster Tolerant High Availability Clusters covering MC/ServiceGuard, Rel. 11.13

Bye
Ralf

[Stephen Doud]

If you have a software contract, you can access the ITRC knowledge database, where you can locate any document written for Serviceguard by Support Center engineers. Simply use the search engine to locate your particular need.

The Managing Serviceguard manual is the first document to review before building a cluster - and it's available outside of the database here:
http://docs.hp.com/en/ha.html#Serviceguard

[flower_2]

thnks for all your help,I find its very usefull.

[flower_2]

i have followed the Service guard document for a newly installed HP-UX server .and installed the patches which was given in the document still the error is coming

Warning: Unable to determine local domain name for Kaveri
Error: Node krishna is refusing Serviceguard communication.
Please make sure that the proper security access is configured on node
krishna through either file-based access (pre-A.11.16 version) or role-based
access (version A.11.16 or higher) and/or that the host name lookup
on node krishna resolves the IP address correctly.
Failed to gather configuration information.
kaveri:root-/>

my inetd.conf entry
hacl-probe stream tcp nowait root /opt/cmom/lbin/cmomd /opt/cmom/lbin/cmomd -f /var/opt/cmom/cmomd.log -r /var/opt/cmom
registrar stream tcp nowait root /etc/opt/resmon/lbin/registrar /etc/opt/resmon/lbin/registrar
hacl-cfg dgram udp wait root /usr/lbin/cmclconfd cmclconfd -p
hacl-cfg stream tcp nowait root /usr/lbin/cmclconfd cmclconfd -c


swlist -l product | grep -i service
ATS-CORE A.11.16.00 Serviceguard Advanced Tape Services
CIFS-Server A.02.02 HP CIFS Server (Samba) File and Print Services
Ignite-UX C.6.3.324 HP-UX System Installation Services
LdapUxClient B.04.00.02 LDAP-UX Client Services
MCPS-COMMON A.01.01.01 HP Service Mission Critical Common Component
NFS B.11.11 ONC/NFS; Network-File System,Information Services,Utilities
PHCO_34348 1.0 VERITAS VM Mgmt Service Provider Patch 05
PHKL_27532 1.0 ; ServiceGuard/vsar incompatibility removed
PHKL_28114 1.0 timeout; ServiceGuard TOC
PHNE_31822 1.0 NTP timeservices upgrade plus utilities
ServiceGuard A.11.16.00 ServiceGuard
TechPrintServ B.11.11.0412 HP-UX Technical Image Printing Service
VRTSfspro 3.5-ga08 VERITAS File System Management Services Provider
VRTSob 3.0.2.261a VERITAS Enterprise Administrator Service
VRTSvmpro 3.5m VERITAS Volume Manager Management Services Provider
WBEMServices A.02.00.10 WBEM Services CORE Product for hp 9000 servers
#

Pls help in this

[Stephen Doud]

Error: Node krishna is refusing Serviceguard communication.

Several causes exist.

1) If this is a new cluster creation (version A.11.16 or later), insure
/etc/cmcluster/cmclnodelist has been created on every node in the cluster.
The file format is like .rhosts. Example:

tic root
toc root

NOTE: EVERY node must be listed in every cmclnodelist!


2) Remove any old cluster binary file on the system, /etc/cmcluster/cmclconfig

3) Enable identd in /etc/inetd.conf.
# grep ident /etc/inetd.conf
#auth stream tcp6 wait bin /usr/lbin/identd identd

or on some systems:
# grep ident /etc/inetd.conf
ident stream tcp wait bin /usr/lbin/identd identd

If it is commented out, remove the "#" comment and run 'inetd -c' .

4) Insure /etc/hosts identifies EVERY NIC that has an IP assigned to it and every line that has an IP also includes the hostname of the server.

Newer versions of Serviceguard depend on 'identd' to do a "caller-ID" like lookup on the node requesting the operation of a Serviceguard command - even the local node.
For that reason, every line in /etc/hosts that refers to a NIC on a cluster node must bear the hostname.

Example:

123.45.67.89 helix.alf.cpqcorp.net helix # hostname
123.45.67.90 helix-hb.alf.cpqcorp.net helix-hb helix # heartbeat
123.45.67.91 helixbkup.alf.cpqcorp.net helixbkup helix # backup LAN

123.45.67.92 torus.alf.cpqcorp.net torus # hostname
123.45.67.93 torus-hb.alf.cpqcorp.net torus-hb torus # heartbeat
123.45.67.94 torusbkup.alf.cpqcorp.net torusbkup torus # backup LAN

Note the hostname is appended to every line.

[flower_2]

Dear Stephen,
As per your causes these are the outputs
Server 1 palar

1) cat /etc/cmcluster/cmclnodelist
vaigai root
palar root

2) ll /etc/cmcluster
total 32
-rw-r--r-- 1 root sys 24 Apr 21 13:38 cmclnodelist
-r-------- 1 bin bin 0 Mar 16 2004 cmknowncmds
dr-xr-xr-x 2 bin bin 8192 Mar 23 2006 cvm
drwxr-xr-x 2 bin bin 96 Apr 21 13:28 sharedtape

3)cat /etc/inetd.conf

ntalk dgram udp wait root /usr/lbin/ntalkd ntalkd
ident stream tcp wait bin /usr/lbin/identd identd

4) vi /etc/hosts
172.17.112.171 palar
172.17.112.175 Vaigai

Server 2 vaigai

1) cat /etc/cmcluster/cmclnodelist
vaigai root
palar root

2) ll /etc/cmcluster
-rw-r--r-- 1 root sys 26 Apr 21 13:36 cmclnodelist
-r-------- 1 bin bin 0 Mar 16 2004 cmknowncmds
dr-xr-xr-x 2 bin bin 8192 Mar 23 2006 cvm
drwxr-xr-x 2 bin bin 96 Apr 23 15:59 sharedtape

3) cat /etc/inetd.conf
ntalk dgram udp wait root /usr/lbin/ntalkd ntalkd
ident stream tcp wait bin /usr/lbin/identd identd

4)cat /etc/hosts
172.17.112.175 Vaigai
172.17.112.171 palar

* rlogin is working perfectly in both the nodes.

[freddy_21 · Accepted solution]

hello

please be carefull with capital word. Make sure your cmclnodelist, /etc/hosts and /.rlogin have consistency name.


thanks
freddy

[flower_2]

hi fredy,

I have tried with the same case sensitive hostnames in all the places.

rlogin is perfecty working.

regards,
arun

[flower_2]

i am attaching the output of the cmscan
of host vaigai

[flower_2]

i am attaching the output of cmscan of palar

[Stephen Doud]

The messages you reported were from another server??
Error: Node krishna is refusing Serviceguard communication

The /etc/hosts entries you reported did not include krishna - why is that?
172.17.112.175 Vaigai
172.17.112.171 palar

Are you actually working on a different cluster build now?

By the way, if the hostname is not capitalized, don't capitalize it in the file.
Also, insure /etc/nsswitch.conf sets the 'hosts:' line to:
hosts: files dns
This insures hostname lookup searched /etc/hosts first.
I also note how short /etc/hosts is. If there are any other IP-bearing NICs in the servers, they MUST be listed in /etc/hosts, AND the hostname must be added to the end of those lines too.

[freddy_21]

I need error message when you run cmquerycl for Vagai and palar. because at your question the node is krishna and kavari.

we cant troubleshooting if you never give us the error message.

i saw at ur cmclnodelist vagai not in capital, your hostname Vagai right. please put Vagai ( not vagai) at cmclnodelist.

i already test at my server.
# uname -a
HP-UX anakin B.11.23 U ia64 1826607440 unlimited-user license

# cat cmclnodelist
Anakin root

# cat /etc/hosts
16.178.0.113 anakin


# cmquerycl -C cluster -n Anakin
Warning: Unable to determine local domain name for anakin
Node Anakin is refusing Serviceguard communication.
Please make sure that the proper security access is configured on node
Anakin through either file-based access (pre-A.11.16 version) or role-based
access (version A.11.16 or higher) and/or that the host name lookup
on node Anakin resolves the IP address correctly.
Failed to gather configuration information.

# cmquerycl -C clusternew -n anakin
Warning: Unable to determine local domain name for anakin ---> ( ignore it )
Note: Disks were discovered which are not in use by either LVM or VxVM.
Use pvcreate(1M) to initialize a disk for LVM or,
use vxdiskadm(1M) to initialize a disk for VxVM.

Node Names: anakin

Bridged networks (local node information only - full probing was not performed):

1 lan0 (anakin)

IP subnets:
.
.
.


so the problem you not have consistency about ur node name. please refer all to your hostname ( Vagai not vagai )



thanks
freddy

[flower_2]

Dear Stephen and Freddy,

Now i am using different servers for testing
the cluster.

name of the servers are as follows and given the above configuration for the same.
vaigai
palar

i will check your inputs and let u know by hour or later.

thanks,
arunkumar

[flower_2]

Hi Freddy,

Pls find the error below

cmquerycl -C /etc/cmcluster/cmclconfig.ascii -n palar -n vaigai

Warning: Unable to determine local domain name for palar
Note: Disks were discovered which are not in use by either LVM or VxVM.
Use pvcreate(1M) to initialize a disk for LVM or,
use vxdiskadm(1M) to initialize a disk for VxVM.
Error: palar lan0 can communicate with palar lan2 over subnet 172.17.112.0
on the IP level, but not on the DLPI level.
There is possibly a network component between the two interfaces
that does not allow any data link level traffic through, which violates
a Serviceguard requirement.
Error: palar lan2 can communicate with palar lan0 over subnet 172.17.112.0
on the IP level, but not on the DLPI level.
There is possibly a network component between the two interfaces
that does not allow any data link level traffic through, which violates
a Serviceguard requirement.
Failed to evaluate network
Failed to gather configuration information.

Regards,
arun

[flower_2]

hi freddy,

as per your information i made all the host name entries identical now the command is working but giving the above error.

regards,
arun

[Rita C Workman]

This error here seems to indicate the problem is communication at this point.

Is there something between the servers that might be blocking communication...?
...try checking with your networking person(s) and see if they are blocking ports...

Rgrds,
Rita

[flower_2]

thanks for all your help ,Finally we have configured the cluster in the two servers.

happy with your help,thanks again sparing time for me.

Regards,
arun

[drteam]

thanks for the support for all the support
who spare the valuable time.

regards,
arun