BookmarkSubscribeRSS Feed
moorsd
Obsidian | Level 7
Is it possible to run two instances of SPD Server in the same SAS environment (Lev)? I'm thinking on SPDS 4.52 and 5.2. Has anyone done this? Are there any best practices or drawbacks to doing this type of thing. Does anyone have any war stories along this line, they could pass on?
8 REPLIES 8
LinusH
Tourmaline | Level 20

Not 100% sure, but it sounds awkward. Spontaneously, it sounds like you need different SAS versions (9.2 - 9.4?), and that means different Lev.

But you could do it on separate Lev but on the same server.

What is the requirement? Testing an upgrade without investing in new servers?

Data never sleeps
moorsd
Obsidian | Level 7
The client I'm at has upgraded to SAS 9.4, but due to the volumes of data didn't upgrade SPDS 4.52. It's compatible with SAS 9.4. However we're now looking at some of the new Hadoop features on SPDS 5.2 and wondered could this be run alongside each other in a sandpit type environment. No critical data is in that environment.
LinusH
Tourmaline | Level 20

If that's the case, there might be a chance to make it work. But I should verify with SAS that this setup will work. I'm especially thinking about the SPD Server SAS engine, that needs to backward compatible as well.

Data never sleeps
JuanS_OCS
Amethyst | Level 16

Never tried. It sounds strange to me, and I feel like if the integration is between 2 different SPDS on different machines, it should not be a problem. But if it is planned to deploy on the same machine, it would not expect a conflict in the metadata definition or even on the configuration, but I would expect a severe conflict on the binaries

 

Maybe it is possible to make it work under a Unix/Linux environment, then you can play a lot with the deployment. Of course, if you do, problably there would be issues during your next software maintenance or migration.

LinusH
Tourmaline | Level 20

There should work to have them installed on the same machine, as long as you keep binaries, port no and data directories seperate.

But in the same metadata Lev...nah...

Data never sleeps
anja
SAS Employee

Hi there,

 

just to make sure: you have 1 SAS 9.4. instances, and you'd like run SPDS 4x and 5 on the same machine,

9.4 accessing both versions (one Metadata Server)? Or do you have a dev and test for example on the same machine and you 'd like to install both SPDS' in parallel, one 9.4 (DEV for example) using SPDS 4 and one 9.4 using SPDS 5?

 

I know that you can install different maintenance releases on the same box, the only thing you'd have to keep in mind is to use different service names and ports. In that case, you would, for example, modify the service name in the .bat file prior to install.

 

The only thing that would keep SPDS 4x and 5 from running in parallel is format incompatibility.

Let me see what i can find out!

 

Thanks

Anja

jklaverstijn
Rhodochrosite | Level 12

There is no technical obstacle to do this, provided you separate the two SPDS instances at the filesystem and port number levels. Unlike you suggest there is no parallel between SPDS and SAS and SPDS does not live in a Lev{n} configuration. By default it uses the site subdirectory in SASHome to store your local configuration settings. You can regard it conceptually as a 3rd party DBMS. So you can install and run them separately (we have done that to separate DEV and PROD) and have both servers in the metadata.

 

If it is a good idea from a resource point of view is another story altogether. SPDS is meant to be a high volume, high throughput performance beast. Having two of those potential monsters on a single server instance sounds a bit counter intuitive and could easily cause all sorts of havoc, so you need to tread lightly. But again, nothing is really stopping you until the data starts flowing. It's ultimately up to you.

 

Hope this helps,

- Jan

SylviaPowell
SAS Employee

Keep this in mind if you are going to try to use two versions of the server: SPD Server 5.2 can read earlier SPD Server 5.x tables and 4.x tables. If SPD Server 5.2 modifies an earlier 5.1 or 4.x table, that table is no longer usable by a version older than SPD Server 5.2. SPD Server 5.2 introduced table modifications in support of Hadoop. Any tables updated by a 5.2 server will include these modifications.

 

 

 

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 

Get Started with SAS Information Catalog in SAS Viya

SAS technical trainer Erin Winters shows you how to explore assets, create new data discovery agents, schedule data discovery agents, and much more.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 8 replies
  • 2441 views
  • 1 like
  • 6 in conversation