Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

add ebsnvme script #41

Open
wants to merge 1 commit into
base: 1.x
Choose a base branch
from
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
379 changes: 379 additions & 0 deletions ebsnvme
Original file line number Diff line number Diff line change
@@ -0,0 +1,379 @@
#!/usr/bin/env python

# Copyright Amazon.com, Inc. and its affiliates. All Rights Reserved.
#
# Licensed under the MIT License. See the LICENSE accompanying this file
# for the specific language governing permissions and limitations under
# the License.

"""
Usage:
Reads EBS information from EBS NVMe device
"""

from __future__ import print_function
import argparse
from ctypes import Structure, Array, c_uint8, c_uint16, c_uint32, c_uint64, \
c_char, addressof, sizeof, byref
from fcntl import ioctl
import json
import os
import signal
import sys
import time

NVME_ADMIN_IDENTIFY = 0x06
NVME_GET_LOG_PAGE = 0x02
NVME_IOCTL_ADMIN_CMD = 0xC0484E41

AMZN_NVME_EBS_MN = "Amazon Elastic Block Store"
AMZN_NVME_STATS_LOGPAGE_ID = 0xD0
AMZN_NVME_STATS_MAGIC = 0x3C23B510
AMZN_NVME_VID = 0x1D0F


class structure_dict_mixin:
def to_dict(self):
return {
field[0]: getattr(self, field[0])
for field in self._fields_
if not field[0].startswith("_") and
not isinstance(getattr(self, field[0]), (Structure, Array))
}


class nvme_admin_command(Structure):
_pack_ = 1
_fields_ = [("opcode", c_uint8),
("flags", c_uint8),
("cid", c_uint16),
("nsid", c_uint32),
("_reserved0", c_uint64),
("mptr", c_uint64),
("addr", c_uint64),
("mlen", c_uint32),
("alen", c_uint32),
("cdw10", c_uint32),
("cdw11", c_uint32),
("cdw12", c_uint32),
("cdw13", c_uint32),
("cdw14", c_uint32),
("cdw15", c_uint32),
("_reserved1", c_uint64)]


class nvme_identify_controller_amzn_vs(Structure):
_pack_ = 1
_fields_ = [("bdev", c_char * 32), # block device name
("_reserved0", c_char * (1024 - 32))]


class nvme_identify_controller_psd(Structure):
_pack_ = 1
_fields_ = [("mp", c_uint16), # maximum power
("_reserved0", c_uint16),
("enlat", c_uint32), # entry latency
("exlat", c_uint32), # exit latency
("rrt", c_uint8), # relative read throughput
("rrl", c_uint8), # relative read latency
("rwt", c_uint8), # relative write throughput
("rwl", c_uint8), # relative write latency
("_reserved1", c_char * 16)]


class nvme_identify_controller(Structure):
_pack_ = 1
_fields_ = [("vid", c_uint16), # PCI Vendor ID
("ssvid", c_uint16), # PCI Subsystem Vendor ID
("sn", c_char * 20), # Serial Number
("mn", c_char * 40), # Module Number
("fr", c_char * 8), # Firmware Revision
("rab", c_uint8), # Recommend Arbitration Burst
("ieee", c_uint8 * 3), # IEEE OUI Identifier
("mic", c_uint8), # Multi-Interface Capabilities
("mdts", c_uint8), # Maximum Data Transfer Size
("_reserved0", c_uint8 * (256 - 78)),
("oacs", c_uint16), # Optional Admin Command Support
("acl", c_uint8), # Abort Command Limit
("aerl", c_uint8), # Asynchronous Event Request Limit
("frmw", c_uint8), # Firmware Updates
("lpa", c_uint8), # Log Page Attributes
("elpe", c_uint8), # Error Log Page Entries
("npss", c_uint8), # Number of Power States Support
("avscc", c_uint8), # Admin Vendor Specific Command Configuration # noqa
("_reserved1", c_uint8 * (512 - 265)),
("sqes", c_uint8), # Submission Queue Entry Size
("cqes", c_uint8), # Completion Queue Entry Size
("_reserved2", c_uint16),
("nn", c_uint32), # Number of Namespaces
("oncs", c_uint16), # Optional NVM Command Support
("fuses", c_uint16), # Fused Operation Support
("fna", c_uint8), # Format NVM Attributes
("vwc", c_uint8), # Volatile Write Cache
("awun", c_uint16), # Atomic Write Unit Normal
("awupf", c_uint16), # Atomic Write Unit Power Fail
("nvscc", c_uint8), # NVM Vendor Specific Command Configuration # noqa
("_reserved3", c_uint8 * (704 - 531)),
("_reserved4", c_uint8 * (2048 - 704)),
("psd", nvme_identify_controller_psd * 32), # Power State Descriptor # noqa
("vs", nvme_identify_controller_amzn_vs)] # Vendor Specific


class nvme_histogram_bin(Structure, structure_dict_mixin):
_pack_ = 1
_fields_ = [("lower", c_uint64),
("upper", c_uint64),
("count", c_uint32),
("_reserved0", c_uint32)]

def to_human_readable(self):
print("[{0.lower:<8} - {0.upper:<8}] => {0.count}".format(self))


class ebs_nvme_histogram(Structure, structure_dict_mixin):
_pack_ = 1
_fields_ = [("num_bins", c_uint64),
("bins", nvme_histogram_bin * 64)]

def to_dict(self):
dict = super(ebs_nvme_histogram, self).to_dict()
dict["bins"] = [self.bins[i].to_dict() for i in range(self.num_bins)]

return dict

def to_human_readable(self):
print("Number of bins: {0}".format(self.num_bins))

print("=================================")
print("Lower Upper IO Count")
print("=================================")

for i in range(self.num_bins):
self.bins[i].to_human_readable()


class nvme_get_amzn_stats_logpage(Structure, structure_dict_mixin):
_pack_ = 1
_fields_ = [("_magic", c_uint32),
("_reserved0", c_char * 4),
("total_read_ops", c_uint64), # total number of read operations
("total_write_ops", c_uint64), # total number of write operations
("total_read_bytes", c_uint64), # total bytes read
("total_write_bytes", c_uint64), # total bytes written
("total_read_time", c_uint64), # total time spent on read operations (in microseconds)
("total_write_time", c_uint64), # total time spent on write operations (in microseconds)
("ebs_volume_performance_exceeded_iops", c_uint64), # time EBS volume IOPS limit was exceeded (in microseconds)
("ebs_volume_performance_exceeded_tp", c_uint64), # time EBS volume throughput limit was exceeded (in microseconds)
("ec2_instance_ebs_performance_exceeded_iops", c_uint64), # time EC2 instance EBS IOPS limit was exceeded (in microseconds)
("ec2_instance_ebs_performance_exceeded_tp", c_uint64), # time EC2 instance EBS throughput limit was exceeded (in microseconds)
("volume_queue_length", c_uint64), # current volume queue length
("_reserved1", c_char * 416),
("read_io_latency_histogram", ebs_nvme_histogram), # histogram of read I/O latencies (in microseconds)
("write_io_latency_histogram", ebs_nvme_histogram), # histogram of write I/O latencies (in microseconds)
("_reserved2", c_char * 496)]

def to_dict(self):
dict = super(nvme_get_amzn_stats_logpage, self).to_dict()
dict["read_io_latency_histogram"] = self.read_io_latency_histogram.to_dict()
dict["write_io_latency_histogram"] = self.write_io_latency_histogram.to_dict()

return dict

def to_json(self):
print(json.dumps(self.to_dict()))

def to_human_readable(self):
print("Total Ops")
print(" Read: {0}".format(self.total_read_ops))
print(" Write: {0}".format(self.total_write_ops))
print("Total Bytes")
print(" Read: {0}".format(self.total_read_bytes))
print(" Write: {0}".format(self.total_write_bytes))
print("Total Time (us)")
print(" Read: {0}".format(self.total_read_time))
print(" Write: {0}".format(self.total_write_time))

print("EBS Volume Performance Exceeded (us)")
print(" IOPS: {0}".format(self.ebs_volume_performance_exceeded_iops))
print(" Throughput: {0}".format(self.ebs_volume_performance_exceeded_tp))
print("EC2 Instance EBS Performance Exceeded (us)")
print(" IOPS: {0}".format(self.ec2_instance_ebs_performance_exceeded_iops))
print(" Throughput: {0}".format(self.ec2_instance_ebs_performance_exceeded_tp))

print("Queue Length (point in time): {0} \n".format(self.volume_queue_length))

print("Read IO Latency Histogram (us)")
self.read_io_latency_histogram.to_human_readable()

print("\nWrite IO Latency Histogram (us)")
self.write_io_latency_histogram.to_human_readable()


class ebs_nvme_device:
def __init__(self, device):
self.device = device

def _nvme_ioctl(self, admin_cmd):

with open(self.device, "r") as dev:
try:
ioctl(dev, NVME_IOCTL_ADMIN_CMD, admin_cmd)
except (OSError) as err:
print("Failed to issue nvme cmd, err: ", err)
sys.exit(1)


class ebs_nvme_device_stats(ebs_nvme_device):
def _query_stats_from_device(self):
stats = nvme_get_amzn_stats_logpage()
admin_cmd = nvme_admin_command(
opcode=NVME_GET_LOG_PAGE,
addr=addressof(stats),
alen=sizeof(stats),
nsid=1,
cdw10=(AMZN_NVME_STATS_LOGPAGE_ID | (1024 << 16))
)
self._nvme_ioctl(admin_cmd)

if stats._magic != AMZN_NVME_STATS_MAGIC:
raise TypeError("[ERROR] Not an EBS device: {0}".format(self.device))

return stats

def _get_stats_diff(self):
curr = self._query_stats_from_device()
if self.prev is None:
self.prev = curr
return curr

diff = nvme_get_amzn_stats_logpage()
diff.volume_queue_length = curr.volume_queue_length

for field, _ in nvme_get_amzn_stats_logpage._fields_:
if field.startswith('_') or field == "volume_queue_length":
continue
if isinstance(getattr(self.prev, field), (int)):
setattr(diff, field, getattr(curr, field) - getattr(self.prev, field))

for histogram_field in ['read_io_latency_histogram', 'write_io_latency_histogram']:
self._calculate_histogram_diff(diff, curr, self.prev, histogram_field)

self.prev = curr
return diff

def _calculate_histogram_diff(self, diff, curr, prev, histogram_field):
prev_hist = getattr(prev, histogram_field)
curr_hist = getattr(curr, histogram_field)
diff_hist = getattr(diff, histogram_field)

diff_hist.num_bins = curr_hist.num_bins
for i in range(diff_hist.num_bins):
diff_hist.bins[i].lower = curr_hist.bins[i].lower
diff_hist.bins[i].upper = curr_hist.bins[i].upper
diff_hist.bins[i].count = curr_hist.bins[i].count - prev_hist.bins[i].count

def _print_stats(self, stats, json_format=False):
if json_format:
print(json.dumps(stats.to_dict()))
else:
stats.to_human_readable()

def _signal_handler(self, sig, frame):
sys.exit(0)

def get_stats(self, interval=0, json_format=False):
if interval > 0:
print("Polling EBS stats every {0} sec(s); press Ctrl+C to stop".format(interval))
signal.signal(signal.SIGINT, self._signal_handler)
self.prev = None

while True:
self._print_stats(self._get_stats_diff(), json_format)
time.sleep(interval)
print("\n")

else:
self._print_stats(self._query_stats_from_device(), json_format)

class ebs_nvme_device_id(ebs_nvme_device):
def get_id(self, volume=False, block_dev=False, udev=False):
id_ctrl = self._query_id_ctrl_from_device()

if not (volume or block_dev or udev):
print("Volume ID: {0}".format(self._get_volume_id(id_ctrl)))
print(self._get_block_device(id_ctrl))
else:
if volume:
print("Volume ID: {0}".format(self._get_volume_id(id_ctrl)))
if block_dev or udev:
print(self._get_block_device(id_ctrl, udev))

def _query_id_ctrl_from_device(self):
id_ctrl = nvme_identify_controller()
admin_cmd = nvme_admin_command(
opcode=NVME_ADMIN_IDENTIFY,
addr=addressof(id_ctrl),
alen=sizeof(id_ctrl),
cdw10=1
)
self._nvme_ioctl(admin_cmd)

if id_ctrl.vid != AMZN_NVME_VID or id_ctrl.mn.decode().strip() != AMZN_NVME_EBS_MN:
raise TypeError("[ERROR] Not an EBS device: ", self.device)

return id_ctrl

def _get_volume_id(self, id_ctrl):
vol = id_ctrl.sn.decode()
if vol.startswith("vol") and vol[3] != "-":
vol = "vol-" + vol[3:]
return vol

def _get_block_device(self, id_ctrl, stripped=False):
dev = id_ctrl.vs.bdev.decode().strip()
if stripped and dev.startswith("/dev/"):
dev = dev[5:]
return dev


if __name__ == "__main__":

# check if the script is being called as ebsnvme-id
if os.path.basename(sys.argv[0]) == 'ebsnvme-id':
sys.argv.insert(1, 'id')

parser = argparse.ArgumentParser(description="Reads EBS information from EBS NVMe devices.")
cmd_parser = parser.add_subparsers(dest="cmd", help="Available commands")
cmd_parser.required = True

stats_parser = cmd_parser.add_parser("stats", help="Get EBS NVMe stats")
stats_parser.add_argument("device", help="Device to query")
stats_parser.required = True
stats_parser.add_argument("-j", "--json", action="store_true",
help="Output in json format")
stats_parser.add_argument("-i", "--interval", type=int, default=0,
help='Interval in seconds to poll ebs stats')

id_parser = cmd_parser.add_parser("id", help="Display id options")
id_parser.add_argument("device", help="Device to query")
id_parser.required = True
id_parser.add_argument("-v", "--volume", action="store_true",
help="Return volume-id")
id_parser.add_argument("-b", "--block-dev", action="store_true",
help="Return block device mapping")
id_parser.add_argument("-u", "--udev", action="store_true",
help="Output data in format suitable for udev rules")

args = parser.parse_args()

try:
if args.cmd == "stats":
stats = ebs_nvme_device_stats(args.device)
stats.get_stats(interval=args.interval, json_format=args.json)

elif args.cmd == "id":
id_info = ebs_nvme_device_id(args.device)
id_info.get_id(volume=args.volume, block_dev=args.block_dev, udev=args.udev)
except (IOError, TypeError) as err:
print(err, file=sys.stderr)
sys.exit(1)