BookmarkSubscribeRSS Feed
🔒 This topic is solved and locked. Need further help from the community? Please sign in and ask a new question.
IqbalS
Calcite | Level 5

Hi,

 

I have a few questions related to the configuration of a newly cloned SAS server…a few key facts:

 

- SAS 9.3 platform used primarily to host Customer Intelligence Studio

- Mid Tier, Metadata and Workspace tiers are all on a single server

- The newly cloned environment mirrors the source environment, however the host information (Host Name/IP) needs to be updated, which I intend to update using the Update Host Name References tool (via SAS Deployment Wizard)

 

Questions:

 

The existing/new deployments will need to run concurrently, I believe this means that the new server will require manual steps to update its multicast parameters to avoid conflict? Can you advise what this entails, is it just a case of updating the IP address within the sasserver1.bat and environment.properties files?

 

Can I postpone running the Update Host Name References tool on client machines until a few days after the new server has been configured; I assume doing so would not affect client access to existing environments, just the new environment I’m setting up?

 

Thanks in advance

 

 

Iqbal

1 ACCEPTED SOLUTION

Accepted Solutions
anja
SAS Employee

Hi,

 

since it is a failover machine, let me throw in another option: Metadata Server Clustering

http://support.sas.com/documentation/cdl/en/omaref/68021/HTML/default/viewer.htm#p1jmsj59l1lnlxn1ji5...

Maybe it would be another option.

 

Going back to the great info provided in the last post, i'd like to add a bit (I hope you don't mind).

Most important when setting up a failover machine is to use the exact same hostname as the prod machine you are cloning.

You can use either DNS or actual host names. Also important is that you keep both systems the same at any point in time; when you upgrade production (new maintenance release, migration, system update, hot fxes, fixes etc.) you always have to do the same on the cloned machine.

 

The Metadata Server cluster - as one example - uses nodes for load balancing, and for keeping all machines on the same content all the time. The backups are stored in a vault location, which is accessible by all nodes.

 

If it is really a true clone, you could restore a back up on this new machine.

Please keep in mind, this really only works if both machines are exact clones.

As Juan mentioned in his post, you could then use promotion (export/import wizards) to keep the environments in sync.

http://support.sas.com/documentation/cdl/en/bisag/68240/HTML/default/viewer.htm#p04l8c3fo8tpyhn103dp...

 

Given that you have 2 machines, maybe virtualization options might be a solution (Containers for example).

This would give you the option to run your Metadata server in protable containers.

 

Here are some pointer

http://support.sas.com/kb/55/919.html

 

It is also important to think about physical data or DBMSs. The config files/metadata might have to be changed if the access to these might change.

 

Hope this helps

Thanks

Anja

View solution in original post

5 REPLIES 5
anja
SAS Employee

Hi,

 

hopully understanding your question correctly ... your "old" environment just keeps running as is, so all your clients access the old environment, and therefore, nothing changes, so yes, you can run the hostname change util later on if needed.

As a best practice, and if daily business allowes, I'd recommend to run it rather sooner than later.


If you want to make the new cloned server production and would change the hotname etc etc. the clients have to be updated at the same time, or else they'd run "against" the old environment.

 

Is the "clone" a new server which will become a new prod sever, and the now current prod server is going to "retire", ?

Or, are you setting up a failover environment??

 

Thanks

Anja

IqbalS
Calcite | Level 5

Thanks for replying Anja.

 

This new/cloned environment is a failover so the client machine's don't need imminent access. Good to know that client access to existing environment won't be affected.

 

I'm not 100% clear on how to update multicast parameters, do you know if the report generated at the end of the Host Name Update wizard is detailed enough to guide me or do I need to bury my head into the Mid-Tier Admin Guide for a while?

 

 

Iqbal

JuanS_OCS
Amethyst | Level 16

@IqbalS,

 

Very interesting and useful questions from @anja that can help to provide the best advise for your scenario.

 

Some general practices I take for those kind of environments:

 

- Stop all the services, make the clone, and disconnect the clone completely from the network, so the current Production machine can keep working for your users, and you can work on the new machine.

 

- You will need to keep track of the major changes on the current server, so when you will be finished with the migration, you will need to update/promote metadata, data and other contents to the cloned server.

 

- On the cloned server, before anything else, protect the dns names used by the original server, by modifying the hosts file of your operating system. On this way, this cloned server will never try to resolve a dns towards the original server.

 

- Indeed, you will need to run the Update Hostnames tool. In other hand, this probably will be sufficient to make the SAS system to be runing, but you might find issues specially when migrating or updating/upgrading the clonned server. Therefore, you need to take some additional actions, very carefully.

 

- The first action, is to run the Update hostname utility for each hostname dns that will change (your server, the mail server, database server, etc), the FQDNs (Full hostnames) and short hostnames. And I would even do it for the IPs.... just in case.

 

- Second action, use your company's prefered grep utility, to search and update each of the previous hostnames, short hostnames, ips... and the multicast address.

 

- Only when you are finished, and with no connection to the network yet from the clonned server, you can tryto start the services and start the technical validations, ensuring everything works fine as expected. 

 

- If still there is some component that does not start properly, please contact SAS Technical Support or come to us again.

 

- When everything works properly, you will need to stop the current server, have the promotion/move/update of the metadata and data from the current server to the clonned server, open the network, execute again the technical validations, let one or 2 key users to validate functionaly your new server. If they give the OK, you can open the connections o all your users.

 

Hope this helps a bit.

Best regards,

Juan

IqbalS
Calcite | Level 5

Thanks for the detailed input. I hadn't thought of it myself but the grep tool in particular will come in handy for manual updates after the Update Host Name Tool has been run.

 

 

Iqbal

 

 

anja
SAS Employee

Hi,

 

since it is a failover machine, let me throw in another option: Metadata Server Clustering

http://support.sas.com/documentation/cdl/en/omaref/68021/HTML/default/viewer.htm#p1jmsj59l1lnlxn1ji5...

Maybe it would be another option.

 

Going back to the great info provided in the last post, i'd like to add a bit (I hope you don't mind).

Most important when setting up a failover machine is to use the exact same hostname as the prod machine you are cloning.

You can use either DNS or actual host names. Also important is that you keep both systems the same at any point in time; when you upgrade production (new maintenance release, migration, system update, hot fxes, fixes etc.) you always have to do the same on the cloned machine.

 

The Metadata Server cluster - as one example - uses nodes for load balancing, and for keeping all machines on the same content all the time. The backups are stored in a vault location, which is accessible by all nodes.

 

If it is really a true clone, you could restore a back up on this new machine.

Please keep in mind, this really only works if both machines are exact clones.

As Juan mentioned in his post, you could then use promotion (export/import wizards) to keep the environments in sync.

http://support.sas.com/documentation/cdl/en/bisag/68240/HTML/default/viewer.htm#p04l8c3fo8tpyhn103dp...

 

Given that you have 2 machines, maybe virtualization options might be a solution (Containers for example).

This would give you the option to run your Metadata server in protable containers.

 

Here are some pointer

http://support.sas.com/kb/55/919.html

 

It is also important to think about physical data or DBMSs. The config files/metadata might have to be changed if the access to these might change.

 

Hope this helps

Thanks

Anja

suga badge.PNGThe SAS Users Group for Administrators (SUGA) is open to all SAS administrators and architects who install, update, manage or maintain a SAS deployment. 

Join SUGA 

CLI in SAS Viya

Learn how to install the SAS Viya CLI and a few commands you may find useful in this video by SAS’ Darrell Barton.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 5 replies
  • 1914 views
  • 1 like
  • 3 in conversation