Hello,
ran a test to check trhoughput of , I am seeing drop in
performance contrary to what is expected with multiple DS are added. Any
thoughts on why there is no scaling of read performace is appreciated.
./iozone -I -i 0 -i 2 -s 512m -r 4 -t 1
-I disable FS caching and run Direct I/O
-i 0 sequential write
-i 2 random read and write
-s 512m file size twice the size of RAM (256MB)
-r record size 4KB
-t 1 throughout mode and one thread
With one DS server enabled in spnfd.conf
One DS server Random read
Children see throughput for 1 random readers = 3355.27 KB/sec
Parent sees throughput for 1 random readers = 3355.21 KB/sec
Min throughput per process = 3355.27 KB/sec
Max throughput per process = 3355.27 KB/sec
Avg throughput per process = 3355.27 KB/sec
Min xfer = 524288.00 KB
two DS servers Random read
Children see throughput for 1 random readers = 3098.43 KB/sec
Parent sees throughput for 1 random readers = 3098.36 KB/sec
Min throughput per process = 3098.43 KB/sec
Max throughput per process = 3098.43 KB/sec
Avg throughput per process = 3098.43 KB/sec
Min xfer = 524288.00 KB
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
