proxmox-rrd: implement new CBOR based format

Storing much more data points now got get better graphs.

Signed-off-by: Dietmar Maurer <dietmar@proxmox.com>
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
This commit is contained in:
Dietmar Maurer 2021-10-13 10:24:41 +02:00 committed by Thomas Lamprecht
parent 0355554905
commit bc68dee171
6 changed files with 623 additions and 324 deletions

View File

@ -14,19 +14,20 @@ pub enum RRDMode {
}
#[api()]
#[repr(u64)]
#[derive(Copy, Clone, Serialize, Deserialize)]
#[serde(rename_all = "lowercase")]
/// RRD time frame resolution
pub enum RRDTimeFrameResolution {
/// 1 min => last 70 minutes
Hour = 60,
/// 30 min => last 35 hours
Day = 60*30,
/// 3 hours => about 8 days
Week = 60*180,
/// 12 hours => last 35 days
Month = 60*720,
/// 1 week => last 490 days
Year = 60*10080,
/// Hour
Hour,
/// Day
Day,
/// Week
Week,
/// Month
Month,
/// Year
Year,
/// Decade (10 years)
Decade,
}

View File

@ -10,8 +10,12 @@ anyhow = "1.0"
bitflags = "1.2.1"
log = "0.4"
nix = "0.19.1"
serde = { version = "1.0", features = ["derive"] }
serde_json = "1.0"
serde_cbor = "0.11.1"
proxmox = { version = "0.14.0" }
proxmox-time = "1"
proxmox-schema = { version = "1", features = [ "api-macro" ] }
proxmox-rrd-api-types = { path = "../proxmox-rrd-api-types" }

View File

@ -13,7 +13,7 @@ use proxmox::tools::fs::{atomic_open_or_create_file, create_path, CreateOptions}
use proxmox_rrd_api_types::{RRDMode, RRDTimeFrameResolution};
use crate::{DST, rrd::RRD};
use crate::rrd::{DST, CF, RRD, RRA};
const RRD_JOURNAL_NAME: &str = "rrd.journal";
@ -81,6 +81,29 @@ impl RRDCache {
})
}
fn create_default_rrd(dst: DST) -> RRD {
let mut rra_list = Vec::new();
// 1min * 1440 => 1day
rra_list.push(RRA::new(CF::Average, 60, 1440));
rra_list.push(RRA::new(CF::Maximum, 60, 1440));
// 30min * 1440 => 30days = 1month
rra_list.push(RRA::new(CF::Average, 30*60, 1440));
rra_list.push(RRA::new(CF::Maximum, 30*60, 1440));
// 6h * 1440 => 360days = 1year
rra_list.push(RRA::new(CF::Average, 6*3600, 1440));
rra_list.push(RRA::new(CF::Maximum, 6*3600, 1440));
// 1week * 570 => 10years
rra_list.push(RRA::new(CF::Average, 7*86400, 570));
rra_list.push(RRA::new(CF::Maximum, 7*86400, 570));
RRD::new(dst, rra_list)
}
fn parse_journal_line(line: &str) -> Result<JournalEntry, Error> {
let line = line.trim();
@ -179,7 +202,7 @@ impl RRDCache {
if err.kind() != std::io::ErrorKind::NotFound {
log::warn!("overwriting RRD file {:?}, because of load error: {}", path, err);
}
RRD::new(entry.dst)
Self::create_default_rrd(entry.dst)
},
};
if entry.time > get_last_update(&entry.rel_path, &rrd) {
@ -246,7 +269,7 @@ impl RRDCache {
if err.kind() != std::io::ErrorKind::NotFound {
log::warn!("overwriting RRD file {:?}, because of load error: {}", path, err);
}
RRD::new(dst)
Self::create_default_rrd(dst)
},
};
rrd.update(now, value);
@ -264,13 +287,29 @@ impl RRDCache {
now: f64,
timeframe: RRDTimeFrameResolution,
mode: RRDMode,
) -> Option<(u64, u64, Vec<Option<f64>>)> {
) -> Result<Option<(u64, u64, Vec<Option<f64>>)>, Error> {
let state = self.state.read().unwrap();
let cf = match mode {
RRDMode::Max => CF::Maximum,
RRDMode::Average => CF::Average,
};
let now = now as u64;
let (start, resolution) = match timeframe {
RRDTimeFrameResolution::Hour => (now - 3600, 60),
RRDTimeFrameResolution::Day => (now - 3600*24, 60),
RRDTimeFrameResolution::Week => (now - 3600*24*7, 30*60),
RRDTimeFrameResolution::Month => (now - 3600*24*30, 30*60),
RRDTimeFrameResolution::Year => (now - 3600*24*365, 6*60*60),
RRDTimeFrameResolution::Decade => (now - 10*3600*24*366, 7*86400),
};
match state.rrd_map.get(&format!("{}/{}", base, name)) {
Some(rrd) => Some(rrd.extract_data(now, timeframe, mode)),
None => None,
Some(rrd) => Ok(Some(rrd.extract_data(start, now, cf, resolution)?)),
None => Ok(None),
}
}
}

View File

@ -1,23 +1,14 @@
//! # Simple Round Robin Database files with fixed format
//! # Round Robin Database files
//!
//! ## Features
//!
//! * One file stores a single data source
//! * Small/constant file size (6008 bytes)
//! * Stores avarage and maximum values
//! * Stores data for different time resolution ([RRDTimeFrameResolution](proxmox_rrd_api_types::RRDTimeFrameResolution))
//! * Stores data for different time resolution
//! * Simple cache implementation with journal support
mod rrd_v1;
pub mod rrd;
mod cache;
pub use cache::*;
/// RRD data source tyoe
#[repr(u8)]
#[derive(Copy, Clone)]
pub enum DST {
/// Gauge values are stored unmodified.
Gauge = 0,
/// Stores the difference to the previous value.
Derive = 1,
}

View File

@ -1,82 +1,175 @@
//! # Round Robin Database file format
//! # Proxmox RRD format version 2
//!
//! The new format uses
//! [CBOR](https://datatracker.ietf.org/doc/html/rfc8949) as storage
//! format. This way we can use the serde serialization framework,
//! which make our code more flexible, much nicer and type safe.
//!
//! ## Features
//!
//! * Well defined data format [CBOR](https://datatracker.ietf.org/doc/html/rfc8949)
//! * Plattform independent (big endian f64, hopefully a standard format?)
//! * Arbitrary number of RRAs (dynamically changeable)
use std::io::Read;
use std::path::Path;
use anyhow::{bail, Error};
use bitflags::bitflags;
use proxmox::tools::{fs::replace_file, fs::CreateOptions};
use serde::{Serialize, Deserialize};
use proxmox_rrd_api_types::{RRDMode, RRDTimeFrameResolution};
use proxmox::tools::fs::{replace_file, CreateOptions};
use proxmox_schema::api;
/// The number of data entries per RRA
pub const RRD_DATA_ENTRIES: usize = 70;
use crate::rrd_v1;
/// Proxmox RRD file magic number
// openssl::sha::sha256(b"Proxmox Round Robin Database file v1.0")[0..8];
pub const PROXMOX_RRD_MAGIC_1_0: [u8; 8] = [206, 46, 26, 212, 172, 158, 5, 186];
/// Proxmox RRD v2 file magic number
// openssl::sha::sha256(b"Proxmox Round Robin Database file v2.0")[0..8];
pub const PROXMOX_RRD_MAGIC_2_0: [u8; 8] = [224, 200, 228, 27, 239, 112, 122, 159];
use crate::DST;
bitflags!{
/// Flags to specify the data soure type and consolidation function
pub struct RRAFlags: u64 {
// Data Source Types
const DST_GAUGE = 1;
const DST_DERIVE = 2;
const DST_COUNTER = 4;
const DST_MASK = 255; // first 8 bits
// Consolidation Functions
const CF_AVERAGE = 1 << 8;
const CF_MAX = 2 << 8;
const CF_MASK = 255 << 8;
}
#[api()]
#[derive(Debug, Serialize, Deserialize, Copy, Clone, PartialEq)]
#[serde(rename_all = "kebab-case")]
/// RRD data source type
pub enum DST {
/// Gauge values are stored unmodified.
Gauge,
/// Stores the difference to the previous value.
Derive,
/// Stores the difference to the previous value (like Derive), but
/// detect counter overflow (and ignores that value)
Counter,
}
/// Round Robin Archive with [RRD_DATA_ENTRIES] data slots.
///
/// This data structure is used inside [RRD] and directly written to the
/// RRD files.
#[repr(C)]
pub struct RRA {
/// Defined the data soure type and consolidation function
pub flags: RRAFlags,
/// Resulution (seconds) from [RRDTimeFrameResolution]
pub resolution: u64,
#[api()]
#[derive(Debug, Serialize, Deserialize, Copy, Clone, PartialEq)]
#[serde(rename_all = "kebab-case")]
/// Consolidation function
pub enum CF {
/// Average
Average,
/// Maximum
Maximum,
/// Minimum
Minimum,
}
#[derive(Serialize, Deserialize)]
pub struct DataSource {
/// Data source type
pub dst: DST,
/// Last update time (epoch)
pub last_update: f64,
/// Count values computed inside this update interval
pub last_count: u64,
/// Stores the last value, used to compute differential value for derive/counters
/// Stores the last value, used to compute differential value for
/// derive/counters
pub counter_value: f64,
/// Data slots
pub data: [f64; RRD_DATA_ENTRIES],
}
impl RRA {
fn new(flags: RRAFlags, resolution: u64) -> Self {
impl DataSource {
pub fn new(dst: DST) -> Self {
Self {
flags, resolution,
dst,
last_update: 0.0,
last_count: 0,
counter_value: f64::NAN,
data: [f64::NAN; RRD_DATA_ENTRIES],
}
}
fn delete_old(&mut self, time: f64) {
let epoch = time as u64;
let last_update = self.last_update as u64;
let reso = self.resolution;
fn compute_new_value(&mut self, time: f64, mut value: f64) -> Result<f64, Error> {
if time <= self.last_update {
bail!("time in past ({} < {})", time, self.last_update);
}
let min_time = epoch - (RRD_DATA_ENTRIES as u64)*reso;
if value.is_nan() {
bail!("new value is NAN");
}
// derive counter value
let is_counter = self.dst == DST::Counter;
if is_counter || self.dst == DST::Derive {
let time_diff = time - self.last_update;
let diff = if self.counter_value.is_nan() {
0.0
} else if is_counter && value < 0.0 {
bail!("got negative value for counter");
} else if is_counter && value < self.counter_value {
// Note: We do not try automatic overflow corrections, but
// we update counter_value anyways, so that we can compute the diff
// next time.
self.counter_value = value;
bail!("conter overflow/reset detected");
} else {
value - self.counter_value
};
self.counter_value = value;
value = diff/time_diff;
}
Ok(value)
}
}
#[derive(Serialize, Deserialize)]
pub struct RRA {
pub resolution: u64,
pub cf: CF,
/// Count values computed inside this update interval
pub last_count: u64,
/// The actual data
pub data: Vec<f64>,
}
impl RRA {
pub fn new(cf: CF, resolution: u64, points: usize) -> Self {
Self {
cf,
resolution,
last_count: 0,
data: vec![f64::NAN; points],
}
}
// directly overwrite data slots
// the caller need to set last_update value on the DataSource manually.
pub(crate) fn insert_data(
&mut self,
start: u64,
resolution: u64,
data: Vec<Option<f64>>,
) -> Result<(), Error> {
if resolution != self.resolution {
bail!("inser_data failed: got wrong resolution");
}
let num_entries = self.data.len() as u64;
let mut index = ((start/self.resolution) % num_entries) as usize;
for i in 0..data.len() {
if let Some(v) = data[i] {
self.data[index] = v;
}
index += 1;
if index >= self.data.len() { index = 0; }
}
Ok(())
}
fn delete_old_slots(&mut self, time: f64, last_update: f64) {
let epoch = time as u64;
let last_update = last_update as u64;
let reso = self.resolution;
let num_entries = self.data.len() as u64;
let min_time = epoch - num_entries*reso;
let min_time = (min_time/reso + 1)*reso;
let mut t = last_update.saturating_sub((RRD_DATA_ENTRIES as u64)*reso);
let mut index = ((t/reso) % (RRD_DATA_ENTRIES as u64)) as usize;
for _ in 0..RRD_DATA_ENTRIES {
t += reso; index = (index + 1) % RRD_DATA_ENTRIES;
let mut t = last_update.saturating_sub(num_entries*reso);
let mut index = ((t/reso) % num_entries) as usize;
for _ in 0..num_entries {
t += reso;
index = (index + 1) % (num_entries as usize);
if t < min_time {
self.data[index] = f64::NAN;
} else {
@ -85,13 +178,14 @@ impl RRA {
}
}
fn compute_new_value(&mut self, time: f64, value: f64) {
fn compute_new_value(&mut self, time: f64, last_update: f64, value: f64) {
let epoch = time as u64;
let last_update = self.last_update as u64;
let last_update = last_update as u64;
let reso = self.resolution;
let num_entries = self.data.len() as u64;
let index = ((epoch/reso) % (RRD_DATA_ENTRIES as u64)) as usize;
let last_index = ((last_update/reso) % (RRD_DATA_ENTRIES as u64)) as usize;
let index = ((epoch/reso) % num_entries) as usize;
let last_index = ((last_update/reso) % num_entries) as usize;
if (epoch - (last_update as u64)) > reso || index != last_index {
self.last_count = 0;
@ -112,258 +206,111 @@ impl RRA {
self.data[index] = value;
self.last_count = 1;
} else {
let new_value = if self.flags.contains(RRAFlags::CF_MAX) {
if last_value > value { last_value } else { value }
} else if self.flags.contains(RRAFlags::CF_AVERAGE) {
(last_value*(self.last_count as f64))/(new_count as f64)
+ value/(new_count as f64)
} else {
log::error!("rrdb update failed - unknown CF");
return;
let new_value = match self.cf {
CF::Maximum => if last_value > value { last_value } else { value },
CF::Minimum => if last_value < value { last_value } else { value },
CF::Average => {
(last_value*(self.last_count as f64))/(new_count as f64)
+ value/(new_count as f64)
}
};
self.data[index] = new_value;
self.last_count = new_count;
}
self.last_update = time;
}
// Note: This may update the state even in case of errors (see counter overflow)
fn update(&mut self, time: f64, mut value: f64) -> Result<(), Error> {
if time <= self.last_update {
bail!("time in past ({} < {})", time, self.last_update);
}
if value.is_nan() {
bail!("new value is NAN");
}
// derive counter value
if self.flags.intersects(RRAFlags::DST_DERIVE | RRAFlags::DST_COUNTER) {
let time_diff = time - self.last_update;
let is_counter = self.flags.contains(RRAFlags::DST_COUNTER);
let diff = if self.counter_value.is_nan() {
0.0
} else if is_counter && value < 0.0 {
bail!("got negative value for counter");
} else if is_counter && value < self.counter_value {
// Note: We do not try automatic overflow corrections, but
// we update counter_value anyways, so that we can compute the diff
// next time.
self.counter_value = value;
bail!("conter overflow/reset detected");
} else {
value - self.counter_value
};
self.counter_value = value;
value = diff/time_diff;
}
self.delete_old(time);
self.compute_new_value(time, value);
Ok(())
}
}
/// Round Robin Database file format with fixed number of [RRA]s
#[repr(C)]
// Note: Avoid alignment problems by using 8byte types only
pub struct RRD {
/// The magic number to identify the file type
pub magic: [u8; 8],
/// Hourly data (average values)
pub hour_avg: RRA,
/// Hourly data (maximum values)
pub hour_max: RRA,
/// Dayly data (average values)
pub day_avg: RRA,
/// Dayly data (maximum values)
pub day_max: RRA,
/// Weekly data (average values)
pub week_avg: RRA,
/// Weekly data (maximum values)
pub week_max: RRA,
/// Monthly data (average values)
pub month_avg: RRA,
/// Monthly data (maximum values)
pub month_max: RRA,
/// Yearly data (average values)
pub year_avg: RRA,
/// Yearly data (maximum values)
pub year_max: RRA,
}
impl RRD {
/// Create a new empty instance
pub fn new(dst: DST) -> Self {
let flags = match dst {
DST::Gauge => RRAFlags::DST_GAUGE,
DST::Derive => RRAFlags::DST_DERIVE,
};
Self {
magic: PROXMOX_RRD_MAGIC_1_0,
hour_avg: RRA::new(
flags | RRAFlags::CF_AVERAGE,
RRDTimeFrameResolution::Hour as u64,
),
hour_max: RRA::new(
flags | RRAFlags::CF_MAX,
RRDTimeFrameResolution::Hour as u64,
),
day_avg: RRA::new(
flags | RRAFlags::CF_AVERAGE,
RRDTimeFrameResolution::Day as u64,
),
day_max: RRA::new(
flags | RRAFlags::CF_MAX,
RRDTimeFrameResolution::Day as u64,
),
week_avg: RRA::new(
flags | RRAFlags::CF_AVERAGE,
RRDTimeFrameResolution::Week as u64,
),
week_max: RRA::new(
flags | RRAFlags::CF_MAX,
RRDTimeFrameResolution::Week as u64,
),
month_avg: RRA::new(
flags | RRAFlags::CF_AVERAGE,
RRDTimeFrameResolution::Month as u64,
),
month_max: RRA::new(
flags | RRAFlags::CF_MAX,
RRDTimeFrameResolution::Month as u64,
),
year_avg: RRA::new(
flags | RRAFlags::CF_AVERAGE,
RRDTimeFrameResolution::Year as u64,
),
year_max: RRA::new(
flags | RRAFlags::CF_MAX,
RRDTimeFrameResolution::Year as u64,
),
}
}
/// Extract data from the archive
pub fn extract_data(
fn extract_data(
&self,
time: f64,
timeframe: RRDTimeFrameResolution,
mode: RRDMode,
start: u64,
end: u64,
last_update: f64,
) -> (u64, u64, Vec<Option<f64>>) {
let epoch = time as u64;
let reso = timeframe as u64;
let end = reso*(epoch/reso + 1);
let start = end - reso*(RRD_DATA_ENTRIES as u64);
let last_update = last_update as u64;
let reso = self.resolution;
let num_entries = self.data.len() as u64;
let mut list = Vec::new();
let raa = match (mode, timeframe) {
(RRDMode::Average, RRDTimeFrameResolution::Hour) => &self.hour_avg,
(RRDMode::Max, RRDTimeFrameResolution::Hour) => &self.hour_max,
(RRDMode::Average, RRDTimeFrameResolution::Day) => &self.day_avg,
(RRDMode::Max, RRDTimeFrameResolution::Day) => &self.day_max,
(RRDMode::Average, RRDTimeFrameResolution::Week) => &self.week_avg,
(RRDMode::Max, RRDTimeFrameResolution::Week) => &self.week_max,
(RRDMode::Average, RRDTimeFrameResolution::Month) => &self.month_avg,
(RRDMode::Max, RRDTimeFrameResolution::Month) => &self.month_max,
(RRDMode::Average, RRDTimeFrameResolution::Year) => &self.year_avg,
(RRDMode::Max, RRDTimeFrameResolution::Year) => &self.year_max,
};
let rrd_end = reso*((raa.last_update as u64)/reso);
let rrd_start = rrd_end - reso*(RRD_DATA_ENTRIES as u64);
let rrd_end = reso*(last_update/reso);
let rrd_start = rrd_end.saturating_sub(reso*num_entries);
let mut t = start;
let mut index = ((t/reso) % (RRD_DATA_ENTRIES as u64)) as usize;
for _ in 0..RRD_DATA_ENTRIES {
let mut index = ((t/reso) % num_entries) as usize;
for _ in 0..num_entries {
if t > end { break; };
if t < rrd_start || t > rrd_end {
list.push(None);
} else {
let value = raa.data[index];
let value = self.data[index];
if value.is_nan() {
list.push(None);
} else {
list.push(Some(value));
}
}
t += reso; index = (index + 1) % RRD_DATA_ENTRIES;
t += reso; index = (index + 1) % (num_entries as usize);
}
(start, reso, list)
}
}
/// Create instance from raw data, testing data len and magic number
pub fn from_raw(mut raw: &[u8]) -> Result<Self, std::io::Error> {
let expected_len = std::mem::size_of::<RRD>();
if raw.len() != expected_len {
let msg = format!("wrong data size ({} != {})", raw.len(), expected_len);
return Err(std::io::Error::new(std::io::ErrorKind::Other, msg));
#[derive(Serialize, Deserialize)]
pub struct RRD {
pub source: DataSource,
pub rra_list: Vec<RRA>,
}
impl RRD {
pub fn new(dst: DST, rra_list: Vec<RRA>) -> RRD {
let source = DataSource::new(dst);
RRD {
source,
rra_list,
}
let mut rrd: RRD = unsafe { std::mem::zeroed() };
unsafe {
let rrd_slice = std::slice::from_raw_parts_mut(&mut rrd as *mut _ as *mut u8, expected_len);
raw.read_exact(rrd_slice)?;
}
if rrd.magic != PROXMOX_RRD_MAGIC_1_0 {
let msg = "wrong magic number".to_string();
return Err(std::io::Error::new(std::io::ErrorKind::Other, msg));
}
Ok(rrd)
}
/// Load data from a file
pub fn load(path: &Path) -> Result<Self, std::io::Error> {
let raw = std::fs::read(path)?;
Self::from_raw(&raw)
if raw.len() < 8 {
let msg = format!("not an rrd file - file is too small ({})", raw.len());
return Err(std::io::Error::new(std::io::ErrorKind::Other, msg));
}
if raw[0..8] == rrd_v1::PROXMOX_RRD_MAGIC_1_0 {
let v1 = rrd_v1::RRDv1::from_raw(&raw)?;
v1.to_rrd_v2()
.map_err(|err| {
let msg = format!("unable to convert from old V1 format - {}", err);
std::io::Error::new(std::io::ErrorKind::Other, msg)
})
} else if raw[0..8] == PROXMOX_RRD_MAGIC_2_0 {
serde_cbor::from_slice(&raw[8..])
.map_err(|err| {
let msg = format!("unable to decode RRD file - {}", err);
std::io::Error::new(std::io::ErrorKind::Other, msg)
})
} else {
let msg = format!("not an rrd file - unknown magic number");
return Err(std::io::Error::new(std::io::ErrorKind::Other, msg));
}
}
/// Store data into a file (atomic replace file)
pub fn save(&self, filename: &Path, options: CreateOptions) -> Result<(), Error> {
let rrd_slice = unsafe {
std::slice::from_raw_parts(self as *const _ as *const u8, std::mem::size_of::<RRD>())
};
replace_file(filename, rrd_slice, options)
let mut data: Vec<u8> = Vec::new();
data.extend(&PROXMOX_RRD_MAGIC_2_0);
serde_cbor::to_writer(&mut data, self)?;
replace_file(filename, &data, options)
}
pub fn last_update(&self) -> f64 {
let mut last_update = 0.0;
{
let mut check_last_update = |rra: &RRA| {
if rra.last_update > last_update {
last_update = rra.last_update;
}
};
check_last_update(&self.hour_avg);
check_last_update(&self.hour_max);
check_last_update(&self.day_avg);
check_last_update(&self.day_max);
check_last_update(&self.week_avg);
check_last_update(&self.week_max);
check_last_update(&self.month_avg);
check_last_update(&self.month_max);
check_last_update(&self.year_avg);
check_last_update(&self.year_max);
}
last_update
self.source.last_update
}
/// Update the value (in memory)
@ -371,32 +318,53 @@ impl RRD {
/// Note: This does not call [Self::save].
pub fn update(&mut self, time: f64, value: f64) {
let mut log_error = true;
let mut update_rra = |rra: &mut RRA| {
if let Err(err) = rra.update(time, value) {
if log_error {
log::error!("rrd update failed: {}", err);
// we only log the first error, because it is very
// likely other calls produce the same error
log_error = false;
}
let value = match self.source.compute_new_value(time, value) {
Ok(value) => value,
Err(err) => {
log::error!("rrd update failed: {}", err);
return;
}
};
update_rra(&mut self.hour_avg);
update_rra(&mut self.hour_max);
let last_update = self.source.last_update;
self.source.last_update = time;
update_rra(&mut self.day_avg);
update_rra(&mut self.day_max);
update_rra(&mut self.week_avg);
update_rra(&mut self.week_max);
update_rra(&mut self.month_avg);
update_rra(&mut self.month_max);
update_rra(&mut self.year_avg);
update_rra(&mut self.year_max);
for rra in self.rra_list.iter_mut() {
rra.delete_old_slots(time, last_update);
rra.compute_new_value(time, last_update, value);
}
}
/// Extract data from the archive
///
/// This selects the RRA with specified [CF] and (minimum)
/// resolution, and extract data from `start` to `end`.
pub fn extract_data(
&self,
start: u64,
end: u64,
cf: CF,
resolution: u64,
) -> Result<(u64, u64, Vec<Option<f64>>), Error> {
let mut rra: Option<&RRA> = None;
for item in self.rra_list.iter() {
if item.cf != cf { continue; }
if item.resolution > resolution { continue; }
if let Some(current) = rra {
if item.resolution > current.resolution {
rra = Some(item);
}
} else {
rra = Some(item);
}
}
match rra {
Some(rra) => Ok(rra.extract_data(start, end, self.source.last_update)),
None => bail!("unable to find RRA suitable ({:?}:{})", cf, resolution),
}
}
}

296
proxmox-rrd/src/rrd_v1.rs Normal file
View File

@ -0,0 +1,296 @@
use std::io::Read;
use anyhow::Error;
use bitflags::bitflags;
/// The number of data entries per RRA
pub const RRD_DATA_ENTRIES: usize = 70;
/// Proxmox RRD file magic number
// openssl::sha::sha256(b"Proxmox Round Robin Database file v1.0")[0..8];
pub const PROXMOX_RRD_MAGIC_1_0: [u8; 8] = [206, 46, 26, 212, 172, 158, 5, 186];
use crate::rrd::{RRD, RRA, CF, DST, DataSource};
bitflags!{
/// Flags to specify the data soure type and consolidation function
pub struct RRAFlags: u64 {
// Data Source Types
const DST_GAUGE = 1;
const DST_DERIVE = 2;
const DST_COUNTER = 4;
const DST_MASK = 255; // first 8 bits
// Consolidation Functions
const CF_AVERAGE = 1 << 8;
const CF_MAX = 2 << 8;
const CF_MASK = 255 << 8;
}
}
/// Round Robin Archive with [RRD_DATA_ENTRIES] data slots.
///
/// This data structure is used inside [RRD] and directly written to the
/// RRD files.
#[repr(C)]
pub struct RRAv1 {
/// Defined the data soure type and consolidation function
pub flags: RRAFlags,
/// Resulution (seconds) from [RRDTimeFrameResolution]
pub resolution: u64,
/// Last update time (epoch)
pub last_update: f64,
/// Count values computed inside this update interval
pub last_count: u64,
/// Stores the last value, used to compute differential value for derive/counters
pub counter_value: f64,
/// Data slots
pub data: [f64; RRD_DATA_ENTRIES],
}
impl RRAv1 {
fn extract_data(
&self,
) -> (u64, u64, Vec<Option<f64>>) {
let reso = self.resolution;
let mut list = Vec::new();
let rra_end = reso*((self.last_update as u64)/reso);
let rra_start = rra_end - reso*(RRD_DATA_ENTRIES as u64);
let mut t = rra_start;
let mut index = ((t/reso) % (RRD_DATA_ENTRIES as u64)) as usize;
for _ in 0..RRD_DATA_ENTRIES {
let value = self.data[index];
if value.is_nan() {
list.push(None);
} else {
list.push(Some(value));
}
t += reso; index = (index + 1) % RRD_DATA_ENTRIES;
}
(rra_start, reso, list)
}
}
/// Round Robin Database file format with fixed number of [RRA]s
#[repr(C)]
// Note: Avoid alignment problems by using 8byte types only
pub struct RRDv1 {
/// The magic number to identify the file type
pub magic: [u8; 8],
/// Hourly data (average values)
pub hour_avg: RRAv1,
/// Hourly data (maximum values)
pub hour_max: RRAv1,
/// Dayly data (average values)
pub day_avg: RRAv1,
/// Dayly data (maximum values)
pub day_max: RRAv1,
/// Weekly data (average values)
pub week_avg: RRAv1,
/// Weekly data (maximum values)
pub week_max: RRAv1,
/// Monthly data (average values)
pub month_avg: RRAv1,
/// Monthly data (maximum values)
pub month_max: RRAv1,
/// Yearly data (average values)
pub year_avg: RRAv1,
/// Yearly data (maximum values)
pub year_max: RRAv1,
}
impl RRDv1 {
pub fn from_raw(mut raw: &[u8]) -> Result<Self, std::io::Error> {
let expected_len = std::mem::size_of::<RRDv1>();
if raw.len() != expected_len {
let msg = format!("wrong data size ({} != {})", raw.len(), expected_len);
return Err(std::io::Error::new(std::io::ErrorKind::Other, msg));
}
let mut rrd: RRDv1 = unsafe { std::mem::zeroed() };
unsafe {
let rrd_slice = std::slice::from_raw_parts_mut(&mut rrd as *mut _ as *mut u8, expected_len);
raw.read_exact(rrd_slice)?;
}
if rrd.magic != PROXMOX_RRD_MAGIC_1_0 {
let msg = "wrong magic number".to_string();
return Err(std::io::Error::new(std::io::ErrorKind::Other, msg));
}
Ok(rrd)
}
pub fn to_rrd_v2(&self) -> Result<RRD, Error> {
let mut rra_list = Vec::new();
// old format v1:
//
// hour 1 min, 70 points
// day 30 min, 70 points
// week 3 hours, 70 points
// month 12 hours, 70 points
// year 1 week, 70 points
//
// new default for RRD v2:
//
// day 1 min, 1440 points
// month 30 min, 1440 points
// year 365 min (6h), 1440 points
// decade 1 week, 570 points
// Linear extrapolation
fn extrapolate_data(start: u64, reso: u64, factor: u64, data: Vec<Option<f64>>) -> (u64, u64, Vec<Option<f64>>) {
let mut new = Vec::new();
for i in 0..data.len() {
let mut next = i + 1;
if next >= data.len() { next = 0 };
let v = data[i];
let v1 = data[next];
match (v, v1) {
(Some(v), Some(v1)) => {
let diff = (v1 - v)/(factor as f64);
for j in 0..factor {
new.push(Some(v + diff*(j as f64)));
}
}
(Some(v), None) => {
new.push(Some(v));
for _ in 0..factor-1 {
new.push(None);
}
}
(None, Some(v1)) => {
for _ in 0..factor-1 {
new.push(None);
}
new.push(Some(v1));
}
(None, None) => {
for _ in 0..factor {
new.push(None);
}
}
}
}
(start, reso/factor, new)
}
// Try to convert to new, higher capacity format
// compute daily average (merge old self.day_avg and self.hour_avg
let mut day_avg = RRA::new(CF::Average, 60, 1440);
let (start, reso, data) = self.day_avg.extract_data();
let (start, reso, data) = extrapolate_data(start, reso, 30, data);
day_avg.insert_data(start, reso, data)?;
let (start, reso, data) = self.hour_avg.extract_data();
day_avg.insert_data(start, reso, data)?;
// compute daily maximum (merge old self.day_max and self.hour_max
let mut day_max = RRA::new(CF::Maximum, 60, 1440);
let (start, reso, data) = self.day_max.extract_data();
let (start, reso, data) = extrapolate_data(start, reso, 30, data);
day_max.insert_data(start, reso, data)?;
let (start, reso, data) = self.hour_max.extract_data();
day_max.insert_data(start, reso, data)?;
// compute montly average (merge old self.month_avg,
// self.week_avg and self.day_avg)
let mut month_avg = RRA::new(CF::Average, 30*60, 1440);
let (start, reso, data) = self.month_avg.extract_data();
let (start, reso, data) = extrapolate_data(start, reso, 24, data);
month_avg.insert_data(start, reso, data)?;
let (start, reso, data) = self.week_avg.extract_data();
let (start, reso, data) = extrapolate_data(start, reso, 6, data);
month_avg.insert_data(start, reso, data)?;
let (start, reso, data) = self.day_avg.extract_data();
month_avg.insert_data(start, reso, data)?;
// compute montly maximum (merge old self.month_max,
// self.week_max and self.day_max)
let mut month_max = RRA::new(CF::Maximum, 30*60, 1440);
let (start, reso, data) = self.month_max.extract_data();
let (start, reso, data) = extrapolate_data(start, reso, 24, data);
month_max.insert_data(start, reso, data)?;
let (start, reso, data) = self.week_max.extract_data();
let (start, reso, data) = extrapolate_data(start, reso, 6, data);
month_max.insert_data(start, reso, data)?;
let (start, reso, data) = self.day_max.extract_data();
month_max.insert_data(start, reso, data)?;
// compute yearly average (merge old self.year_avg)
let mut year_avg = RRA::new(CF::Average, 6*3600, 1440);
let (start, reso, data) = self.year_avg.extract_data();
let (start, reso, data) = extrapolate_data(start, reso, 28, data);
year_avg.insert_data(start, reso, data)?;
// compute yearly maximum (merge old self.year_avg)
let mut year_max = RRA::new(CF::Maximum, 6*3600, 1440);
let (start, reso, data) = self.year_max.extract_data();
let (start, reso, data) = extrapolate_data(start, reso, 28, data);
year_max.insert_data(start, reso, data)?;
// compute decade average (merge old self.year_avg)
let mut decade_avg = RRA::new(CF::Average, 7*86400, 570);
let (start, reso, data) = self.year_avg.extract_data();
decade_avg.insert_data(start, reso, data)?;
// compute decade maximum (merge old self.year_max)
let mut decade_max = RRA::new(CF::Maximum, 7*86400, 570);
let (start, reso, data) = self.year_max.extract_data();
decade_max.insert_data(start, reso, data)?;
rra_list.push(day_avg);
rra_list.push(day_max);
rra_list.push(month_avg);
rra_list.push(month_max);
rra_list.push(year_avg);
rra_list.push(year_max);
rra_list.push(decade_avg);
rra_list.push(decade_max);
// use values from hour_avg for source (all RRAv1 must have the same config)
let dst = if self.hour_avg.flags.contains(RRAFlags::DST_COUNTER) {
DST::Counter
} else if self.hour_avg.flags.contains(RRAFlags::DST_DERIVE) {
DST::Derive
} else {
DST::Gauge
};
let source = DataSource {
dst,
counter_value: f64::NAN,
last_update: self.hour_avg.last_update, // IMPORTANT!
};
Ok(RRD {
source,
rra_list,
})
}
}