== sanityn test 77ud: TBF: a rate change on a BPS rule leaves an IOPS rate ========================================================== 21:49:27 (1789264167)
ost.OSS.ost_io.nrs_policies=tbf nid+opcode
ost.OSS.ost_io.nrs_tbf_rule=start nidrd nid={0@lo 192.168.202.36@tcp}&opcode={ost_read} bps=5MiB/s
ost.OSS.ost_io.nrs_tbf_rule=start nidwr nid={0@lo 192.168.202.36@tcp}&opcode={ost_write} bps=5MiB/s
ost.OSS.ost_io.nrs_tbf_rule=
regular_requests:
CPT 0:
nidwr nid={0@lo 192.168.202.36@tcp}&opcode={ost_write} 5, seq 3, type=nid+opcode, ref 0 bps=5242880
nidrd nid={0@lo 192.168.202.36@tcp}&opcode={ost_read} 5, seq 2, type=nid+opcode, ref 0 bps=5242880
default * 10000, seq 1, type=nid+opcode, ref 0
high_priority_requests:
CPT 0:
nidwr nid={0@lo 192.168.202.36@tcp}&opcode={ost_write} 5, seq 3, type=nid+opcode, ref 0 bps=5242880
nidrd nid={0@lo 192.168.202.36@tcp}&opcode={ost_read} 5, seq 2, type=nid+opcode, ref 0 bps=5242880
default * 10000, seq 1, type=nid+opcode, ref 0
ost.OSS.ost_io.nrs_tbf_rule=change nidrd rate=100
ost.OSS.ost_io.nrs_tbf_rule=change nidwr rate=100
ost.OSS.ost_io.nrs_tbf_rule=
regular_requests:
CPT 0:
nidwr nid={0@lo 192.168.202.36@tcp}&opcode={ost_write} 100, seq 3, type=nid+opcode, ref 0
nidrd nid={0@lo 192.168.202.36@tcp}&opcode={ost_read} 100, seq 2, type=nid+opcode, ref 0
default * 10000, seq 1, type=nid+opcode, ref 0
high_priority_requests:
CPT 0:
nidwr nid={0@lo 192.168.202.36@tcp}&opcode={ost_write} 100, seq 3, type=nid+opcode, ref 0
nidrd nid={0@lo 192.168.202.36@tcp}&opcode={ost_read} 100, seq 2, type=nid+opcode, ref 0
default * 10000, seq 1, type=nid+opcode, ref 0
ost.OSS.ost_io.nrs_tbf_rule=change nidrd rate=20
ost.OSS.ost_io.nrs_tbf_rule=change nidwr rate=20
not_under_control: 0
cpu_npartitions on ost1 is 1
Limited write rate: 20, read rate: 20, bs: 1048576
Verify the write rate is under TBF control
100+0 records in
100+0 records out
104857600 bytes (105 MB, 100 MiB) copied, 6.45064 s, 16.3 MB/s
Write runtime is 8 s, speed is 12.500000 vs. 20
Verify the read rate is under TBF control
100+0 records in
100+0 records out
104857600 bytes (105 MB, 100 MiB) copied, 5.00152 s, 21.0 MB/s
Read runtime is 6 s, speed is 16.666666 vs. 20
cleanup_tbf /mnt/lustre/d77ud.sanityn
sleep 5 for ZFS MDS
sleep 5 for ZFS OST
Waiting for MDT destroys to complete
Cleanup TBF
ost.OSS.ost_io.nrs_tbf_rule=stop nidrd
ost.OSS.ost_io.nrs_tbf_rule=stop nidwr
ost.OSS.ost_io.nrs_policies=fifo