<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
<META NAME="GENERATOR" CONTENT="GtkHTML/3.32.2">
</HEAD>
<BODY>
Hi,<BR>
<BR>
After researching this in the archives it seems 85% of Lustre's native performance is reasonable via NFS. Over GigE I'm seeing 114 MB/sec read/write via Lustre native. NFS writes at 44 MB/sec via NFS (about 38%) but reads at 95 MB/sec. CIFS is lucky to see 18 MB/sec writes after tuning Samba (12 MB/sec prior). I've tried tuning the Lustre client and NFS client and server but nothing impacts the performance except setting noatime on NFS, which gets another 1.5 MB/sec on writes. What are some of you seeing in this configuration? Suggestions?<BR>
<BR>
<BR>
My configuration is a separate MDS and OSS with one OST running Lustre 1.8.6 on GigE only. Testing is currently simple using dd. Obdfilter-survey results below (single Areca 1680 w/ 10 3TB disks in RAID 6, no stripe/stride tuning or external journal, using the stock kernel driver w/ latest firmware):<BR>
<BR>
Wed Sep 21 20:07:02 PDT 2011 Obdfilter-survey for case=disk from localhost.localdomain<BR>
ost 1 sz 16777216K rsz 1024K obj 1 thr 1 write 183.91 [ 126.86, 241.73] rewrite 202.90 [ 163.57, 241.73] read 753.34 [ 152.83,3440.19] <BR>
ost 1 sz 16777216K rsz 1024K obj 1 thr 2 write 384.32 [ 255.70, 486.45] rewrite 516.70 [ 290.68, 598.34] read 1570.43 [ 376.59,3589.96] <BR>
ost 1 sz 16777216K rsz 1024K obj 1 thr 4 write 415.34 [ 332.63, 488.45] rewrite 653.74 [ 38.96, 877.01] read 1915.92 [ 554.28,4865.86] <BR>
ost 1 sz 16777216K rsz 1024K obj 1 thr 8 write 401.46 [ 305.66, 541.39] rewrite 664.63 [ 383.57,1030.43] read 2081.41 [ 803.67,5404.30] <BR>
ost 1 sz 16777216K rsz 1024K obj 1 thr 16 write 406.12 [ 290.68, 541.39] rewrite 700.57 [ 0.00,1112.77] read 1563.54 [ 63.93,1936.96] <BR>
ost 1 sz 16777216K rsz 1024K obj 2 thr 2 write 373.83 [ 293.66, 458.50] rewrite 569.01 [ 60.99, 806.09] read 1380.53 [ 307.19,3737.88] <BR>
ost 1 sz 16777216K rsz 1024K obj 2 thr 4 write 443.58 [ 351.64, 517.43] rewrite 591.88 [ 334.62, 852.04] read 1633.55 [ 7.99,4560.89] <BR>
ost 1 sz 16777216K rsz 1024K obj 2 thr 8 write 459.65 [ 374.58, 532.95] rewrite 602.29 [ 358.60,1039.84] read 1416.61 [ 37.96,3275.32] <BR>
ost 1 sz 16777216K rsz 1024K obj 2 thr 16 write 461.84 [ 366.59, 517.42] rewrite 663.96 [ 262.71,1056.77] read 1366.43 [ 51.94,2379.27] <BR>
ost 1 sz 16777216K rsz 1024K obj 4 thr 4 write 434.83 [ 281.67, 518.41] rewrite 561.85 [ 326.63, 859.03] read 1225.48 [ 4.99,3352.31] <BR>
ost 1 sz 16777216K rsz 1024K obj 4 thr 8 write 481.20 [ 378.57, 566.37] rewrite 602.96 [ 143.84, 989.18] read 1478.49 [ 14.98,4548.44] <BR>
ost 1 sz 16777216K rsz 1024K obj 4 thr 16 write 493.07 [ 373.58, 583.34] rewrite 628.67 [ 304.66,1044.05] read 1475.33 [ 57.93,3221.92] <BR>
ost 1 sz 16777216K rsz 1024K obj 8 thr 8 write 436.75 [ 4.00, 535.33] rewrite 556.52 [ 249.72,1003.87] read 1473.04 [ 20.98,3419.10] <BR>
ost 1 sz 16777216K rsz 1024K obj 8 thr 16 write 483.68 [ 383.57, 588.34] rewrite 627.79 [ 226.74,1194.59] read 1490.97 [ 41.95,4202.74] <BR>
ost 1 sz 16777216K rsz 1024K obj 16 thr 16 write 469.65 [ 377.57, 552.84] rewrite 544.70 [ 257.71,1216.66] read 1375.23 [ 31.96,1133.43] <BR>
done!<BR>
<BR>
Thank you,<BR>
<BR>
Dan
</BODY>
</HTML>