adcfgclone.pl appsTier dualfs erroring out while Instantiating the OHS Config.

[ef_arg]

This is what oracle says:

ErmanArslansOracleBlog wrote

you have created a symbolic link for the file /etc/hosts --- EBS does not support sym links on binaries
lrwxrwxrwx 1 root root 12 Jun 6 12:56 /etc/hosts -> ./inet/hosts

this is what we have in our EBS internal instances servers
[oracle@celebscovm141 etc]$ ls -la hos*
-rw-r--r-- 1 root root 9 Jan 10 2017 host.conf
-rw-r--r--. 1 root root 214 Jun 16 2022 hosts <<<< this is not sym link --- this is a Linux platform however, not Solaris
-rw-r--r-- 1 root root 370 Jan 12 2010 hosts.allow
-rw-r--r-- 1 root root 460 Jan 12 2010 hosts.deny

>> please as your Solaris System Administrator to research how to make /etc/hosts as a flat file instead of a symbolic link

But to have a link sym instead of a plain text file is the default setup for Sparc Solaris.

[ErmanArslansOracleBlog]

Is the format used in that target file(inet/hosts) correct?

[ErmanArslansOracleBlog]

Send me a strace or truss output that is gathered for the apache process.. (start the process while gathering the trace, get the error and upload the trace..)

In the following thread, we used the strace for getting more info about the foundations of the problem.

Ref: http://erman-arslan-s-oracle-forum.124.s1.nabble.com/Apache-exiting-with-152-td11857.html

[ef_arg]

ErmanArslansOracleBlog wrote

Is the format used in that target file(inet/hosts) correct?

Yes, I think so:

bash-3.2# cd /etc
bash-3.2# ls -ltr host*
-r--r--r--   1 root     root         435 Sep 14  2015 hosts.allow
lrwxrwxrwx   1 root     root          12 Jun  6 12:56 hosts -> ./inet/hosts
-rw-r--r--   1 root     root          33 Jul 10 09:21 hostname.vnic2
bash-3.2# cat hosts
#
# Internet host table
#
127.0.0.1   localhost localhost.localdomain localhost4
::1         localhost localhost.localdomain localhost6
10.0.10.19 hostname.domainname.net hostname
bash-3.2# cat ./inet/hosts
#
# Internet host table
#
127.0.0.1   localhost localhost.localdomain localhost4
::1         localhost localhost.localdomain localhost6
10.0.10.19 hostname.domainname.net hostname
bash-3.2# cat hostname.vnic2
10.0.10.19 netmask 255.255.255.0
bash-3.2# cat hosts.allow
#
# Copyright (c) 2002 by Sun Microsystems, Inc.
# All rights reserved.
#
#ident  "@(#)hosts.allow 1.3     05/05/27     SMI"
#
# This file is supplied as part of the Solaris Security Toolkit and
# is used to grant access to specific services as part of the Solaris 9
# TCP Wrappers implementation.  This file should be customized based
# on individual site needs.
#

#sshd:          LOCAL
#sendmail:      localhost
ALL:            ALL
bash-3.2#

[ErmanArslansOracleBlog]

Something may be wrong with your hosts file.
your hosts file has some extra lines  (for ipv6) and it may also have non-ASCII, non-printing characters ..
ipv6 and that localhost4 shouldn't be a problem though.. But still I don't like them..
But anyways, they shouldn't cause any problem..

Example host file entries for those two;

127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1 localhost localhost.localdomain localhost6 localhost6.localdomain6

Again, there may some extra characters in that file..

Check that!

Also I already told you the following ->

Send me a strace or truss output that is gathered for the apache process.. (start the process while gathering the trace, get the error and upload the trace..)

In the following thread, we used the strace for getting more info about the foundations of the problem.

Ref: http://erman-arslan-s-oracle-forum.124.s1.nabble.com/Apache-exiting-with-152-td11857.html

[ef_arg]

Hi Erman!

The host file was fine, we tried several different formats, and still getting the same error. The case with oracle was not helpfull att all.. .they blamed the clone procedure, the DB restore using rman, using FND_CONC_CLONE.SETUP_CLEAN and some other random stuff.

We decided to do the exact same clone, using the exact same apps binaries bkp and rman restore on a different box, on  diferent zone (this is sparc solaris). Host file setup was exactly the same, nslookup was also not resolving (no dns server) and it worked just fine, no issues.

Firewalls, NW setup, kernell parameters, os users, etc, everything at OS level is exactly the same in both servers. I think there is some aditional security layer at container/zone/data center lever, that it is terminating the http process right after it is starterd... not sure but since we were able to proceed with this aproache, i wont waste more time here.

We are planning to do platform migration, from Sparc to Linux ASAP, and then upgrade our whole tech stack as we are vert behind already on CPUs and EBS patches and our DB is already no longer being supported.

Thanks again for your support!
Regards

[ErmanArslansOracleBlog]

The decision to migrate to Linux seems prudent given the current challenges. Being on Oracle Linux is better for both Oracle Databases and EBS.

Thank you for the detailed explanation of the issue. It's clear that you've invested significant effort in troubleshooting.

We're available to assist with the migration process if needed.