Mount Windows Shared path on Linux Host(Compute Server) for storing/accessing huge datasets

[SangramjitPanda]

Hi All,

We have Installed SAS on Linux Host and we store the datasets and other physical content on Compute Server(linux host).

The Compute Server space is limited and we don't want to extend its capacity.

We need to store the huge datasets on the Windows Network Shared Path instead, so that we can use the Windows Network Shared path for storing/accessing the huge datasets/files.

Please do suggest us the workaround to mount the Windows Network Shared Path on the Compute Server(linux Host), so that we can use the libname statements to point those path for storing the datasets.

Quick response will be highly appreciated.

Thanks,

Sangramjit

[SASKiwi]

This is more a Linux-related question rather than a SAS one so you may find your answer quicker on Linux forums.

[PaulHomes]

Since you mentioned that the SAS datasets are huge, I might also ask if you have considered whether this method of providing additional storage space will provide adequate performance for your SAS users? It would be well worth testing before rolling it out. If you haven't already seen them, I'd recommend reading the following papers (and the associated resources mentioned within them):

• SAS Global Forum Paper SAS6761-2016: Best Practices for Configuring Your I/O Subsystem for SAS®9 App... by Tony Brown and Margaret Crevar.
• SAS Global Forum Paper SAS6201-2016: How to Maintain Happy SAS®9 Users by Margaret Crevar

Of course another consideration is to see if your users can take advantage of general SAS programming techniques to reduce the storage footprint for their SAS tables (appropriate variable lengths, compress option, etc).

[JuanS_OCS]

I can subscribe to the consideration provided by @PaulHomes.

And something else: if some of those huge datasets are used as relational databases, instead for pure data analysis, it probably would be interesting for you to consider to move some of that data to a relational database, and extract to SAS only the data to be used for analysis and reporting.

Then, a last interesting option, depending on the real amount of data you have behind, is the usage of Hadoop as data lake.

[Kurt_Bremser]

It has been my experience that network shares can never give you the performance of directly attached storage or SAN storage.

So you should think twice before doing that. In the end you'll waste a lot of time before you end up increasing the server's storage anyway.

Extending local storage on the server is also easier to handle and maintain in the future.

Mounting Windows(CIFS) shares on UNIX is usually done with Samba (client).

But IMO what you want to do is just a method of crippling your SAS server.