Serviceguard package not failing over

[Mr Zer0]

Hi All,

Package failover was working previously, nothing has been changed, no patched have been applied,

I see the following message in the syslog-
Aug 3 20:44:14 nnm2srv automountd[1100]: match_addr: getaddrinfo failed: host nor service provided, or not known
Aug 3 20:45:19 nnm2srv automountd[1100]: match_addr: getaddrinfo failed: host nor service provided, or not known
Aug 3 20:45:50 nnm2srv above message repeats 48 times

The machine has 3 IP's(A, B, and C) and the name was resolving with some other IP(B) but i have a product that's license was bound to the IP(A). So i updated the /etc/hosts so that the name lookup resolves to IP A. I have restarted the node and even restarted the cluster, but the package is unable to start.

Serviceguard version is A.11.18.00 and OS is HPUX 11.31

[Mark S Meadows · Accepted solution]

When you say "the package is not able to start" please can you supply the message returned by the system.

Also please supply the output to cmviewcl -v along with the appropriate entries in the package control log file, i.e. /etc/cmcluster/.cntl.log with regards to the last entries where it attempted to start.

Regards,

Mark

[SoorajCleris]

Hi,

What the cluster logs and package logs are saying?

Regards,
Soraj

[Mr Zer0]

Hi All,

Thanks for the replies.

The package run script and the halt script do not have timeout's, so the package just stays in "starting" state when i move it to the second system.

And there is no entry in the package log file, but only the above message in the syslog

[Lijeesh N G_1]

Hi,

May know what activity(recent change) did in that particular server.
With respect to above,I could see you did some network changes,provide some more details about that, you may check syslog also.

Also some more details about your cluster and package.

Regards,
LIJEESH N G

[Mr Zer0]

Hi,

As i said - nothing was done, no patches installed and all, all i did was there were 3 entries in /etc/hosts like

B hostname
C hostname
A hostname

where A, B and C are the 3 ip's of server.
I deleted the entries B and C. Because application does a nslookup and it finds a different IP for the host and license of the application is bound to address A.

[Mr Zer0]

Package is active/passive, failover kind. It was all working before this change.

[Mr Zer0]

The problem was with the application which was unable to start a process, it's resolved. Thanks everyone for ur help :D

[Mr Zer0]

`