David Crayford Wrote:

If you intend to set up zCX for high availability, it would involve running
zCX instances on different LPARs. To achieve this, you'll likely need
Docker Swarm for creating a cluster that operates across multiple nodes.
While OpenShift (Kubernetes) is a more advanced option, it comes with much
higher operational costs in terms of resource usage. OpenShift 4.13 offers
incremental improvements, we've been told it's expected that further
enhancements will be introduced over time.

Thanks David.  I am actually working with one of your associates at Rocket on 
this.  I get the impression that while Rocket has tested RTE in docker (and 
documented as such), clustering never was.  They are going to work on that, and 
provide some example/samples.


Timothy Sipples wrote:

It provides an example using nginx, a popular HTTP(S) server. The example uses 
this startup command:

docker run -p 8080:80 -d nginx

The -p parameter is crucial. In this example it means, “Expose port 8080 to the 
outside world, and any traffic to/from port 8080 should be directed to/from 
port 80 within this nginx container image.” So if you’re trying to get two 
container images (on two different z/OS LPARs, as Dave Crayford suggested) to 
talk to each other you’d start them up with the -p option and then tell them to 
talk to each other on the respective external ports you’ve chosen. Hopefully 
obviously you should pick external ports that aren’t already occupied or 
reserved for other z/OS uses in that LPAR.

Thanks Timothy.  Yep found all that, have the instance up and working just fine 
it’s the peer to peer networking that is not working.   The fine folks at 
Rocket indicate that their software is picking up the internal container IP, 
and not using the Host IP causing the problem.   They are working up their own 
testing, and believe that docker overlay networking can resolve this.

Yes indeed, the reason for zCX vs straight z/OS Unix is the fact that it all 
runs on ZIIPS.   Plus its cool technology to learn and play with.  It’s all a 
part of changing perceptions here that mainframe is just a COBOL machine.   
Separately, we are doing a zCX Openshift POC, mainly because there wasn’t any 
additional infrastructure to provision.   If we were to do Openshift in any 
larger capacity, it would likely be on a LinuxOne box(es).   Our next move for 
DB2 IDAA has to be LinuxOne boxes as well.   I mention all of this as is to 
leverage our investment in GDPS and being able to Hiperswap/Region swap 
everything on  Z or LinuxOne boxes in a coordinated fashion for Disaster 
recovery or facilities maintenance.





This e-mail transmission contains information that is confidential and may be 
privileged.   It is intended only for the addressee(s) named above. If you 
receive this e-mail in error, please do not read, copy or disseminate it in any 
manner. If you are not the intended recipient, any disclosure, copying, 
distribution or use of the contents of this information is prohibited. Please 
reply to the message immediately by informing the sender that the message was 
misdirected. After replying, please erase it from your computer system. Your 
assistance in correcting this error is appreciated.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to