ORDS 503 error

[suttontools]

Hi,
Got this ORDS error after install on Centos 7.6 minimum install with build-utilities added:

The connection pool named: |apex|| is not correctly configured, due to the following error(s): IO Error: The Network Adapter could not establish the connection

Tried initially with latest ORDS but script rejected it. Configured with listed version, script ran okay. With firewall off I can make basic connection to DB via SQL Developer, with firewall on SSH Hosts connection works fine.

I had changed the hostname to one with hyphens, some forums suggest not using this,s o I changed it to one without. I changed ords_params.properties, tnsnames.ora and listener.ora to the actual new hostname but it made no difference (rather than localhost). I will try and enable the debug messages in ORDS too. I notice that the APEX_INSTANCE_ADMIN_USER is password expired and the ORDS_METADATA user is expired and locked. I ran the install under screen and did not watch it the second time through. I'll look and run the script for the APEX admin user when I find it,
Thanks,
Steve
PS ran sudo find / -name apex_config.sql and the only instance was in the /tmp/oxar/apex/apex_config.sql
Also the APEX_040000 is still in the DB - perhaps APEX didn't install properly (hence the ORDS error)
Also note I left the AOP printing to install

These are the files:
oracle-xe-11.2.0-1.0.x86_64.rpm.zip
sqlcl-18.4.0.007.1818.zip
ords-18.3.0.270.1456.zip
apex_18.2_en.zip

This is one error on install (seems to be due to a too big an integer for bash?):

• Installing Oracle XE *

######################################################################## 100.0%
/var/tmp/rpm-tmp.kY321c: line 257: [: 18446744073692774399: integer expression expected
/var/tmp/rpm-tmp.kY321c: line 271: [: 18446744073692774399: integer expression expected

• Oracle Config *

[suttontools]

FWIW:
I upgraded the kernel to 4.20.. from elrepo - might not have been necessary as error persisted. USed the example sysctl.conf from here:
https://community.oracle.com/thread/4154293
and reinstalled oracle-xe from the rpm package - note the rpm -e oracle-xe didn't remove everything.

This removed the errors but - not even sure if this was'the' problem yet...

[suttontools]

In line 18 of /scripts/ords.sh there is:

#move the parameters file from ords to tmp/ords/ (where ords files are)
mv -f ${OOS_SOURCE_DIR}/ords/ords_params.properties ${ORDS_PARAMS}

However, if you unzip ords-18.3.0.270.1456.zip, the ords_param.properties file is in the sub directory params. I think this leads to:

• Installing ORDS *

######################################################################## 100.0%
mv: cannot stat ‘/tmp/oxar/ords/ords_params.properties’: No such file or directory
Mar 05, 2019 9:26:29 PM
INFO: Set config.dir to /etc in: /tmp/oxar/tmp/ords/ords.war

Actually the error is there, but I don't really know what 'cannot stat' means!

[suttontools]

Just a final comment. It is quite possible that if I started with the last supported ords version, this may have worked fine. However the recovery or uninstall of failed updates or trying to re-run the build.sh leads to the problems I came across.
I would just like to add that I had used oxar in the past with no install issues whatsoever, so I am grateful for the hard work that you put into it.
Perhaps I should try Docker? :-)

[GasparYYC]

@suttontools I've noticed that if the Network MAC changes for whatever reason (e.g moving the VM, cloning the VM, etc) then the listener gets cranky and the consequence is that ORDS give the error you mentioned.

The way I fix it is:
In the file: /u01/app/oracle/product/11.2.0/xe/network/admin/listener.ora

comment out this line:
(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC_FOR_XE))

And that cures the problem all the time.

Good luck!