BookmarkSubscribeRSS Feed
MargaretC
SAS Employee

SAS has been working with DDN on deployments of their EXAScaler Cloud parallel filesystem solution running on Microsoft Azure to optimize SAS GRID in the cloud for data-intensive customers.  We recently concluded testing of SAS 9.4 on MS Azure instances using EXAScaler Cloud which is available today in the Azure Marketplace - EXAScaler Cloud using an automated utility to test IO operations that approximate the behavior of SAS. The results of the testing were very favorable.

 

The test infrastructure consisted of the 11 MS Azure instances configured as follows: 

  • Three E64-32ds_v4 instances for the SAS clients – each with six P30 premium drives striped together for local storage.
  • Six D64s_v3 instances for the OSS hosts – each with five P60 premium drives; one E8s_v4 for the MGS host – with one P60 premium drive; and one E16s_v4 for the MDS host – with one P60 premium drive. 
  • EXAScaler 5.2.0 was used for the testing.

 

For the test, the following EXAScaler settings were applied:

  • Tuning command for all OSS nodes:

lctl set_param osd-ldiskfs.*.read_cache_enable=0 osd-ldiskfs.*.writethrough_cache_enable=0 obdfilter.*.brw_size=16 obdfilter.*.precreate_batch=1024

  • Tuning command for all client nodes:

lctl set_param mdc.*.max_rpcs_in_flight=128 osc.*.max_pages_per_rpc=16M osc.*.max_rpcs_in_flight=16 osc.*.max_dirty_mb=1024 llite.*.max_read_ahead_mb=2048 osc.*.checksums=0  llite.*.max_read_ahead_per_file_mb=256

 

Using rhel_iotest, we achieved the following results:

  • The 1-node test gave us 237 MB per second per physical core for WRITEs and 267 MB per second per physical core for READs.
  • The 2-node test gave us 225 MB per second per physical core for WRITEs and 218 MB per second per physical core for READs.
  • The 3-node test gave us 167 MB per second per physical core for WRITEs and 150 MB per second per physical core for READs.

 

SAS recommends a minimum of 100 MB per second per physical core for WRITEs and READs. As you can see, the EXAScaler Cloud infrastructure, as configured above, can produce throughput exceeding the minimum needs of a 64-core SAS Grid application.  Please reach out to DDN for more details on EXAScaler Cloud and how to configure it for optimal performance in MS Azure.

 

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
  • 0 replies
  • 5385 views
  • 7 likes
  • 1 in conversation