mirror of
https://github.com/kemko/nomad.git
synced 2026-01-03 17:05:43 +03:00
fec91d1dc8
In #20619 we overhauled how we were gathering stats for Windows processes. Unlike in Linux where we can ask for processes in a cgroup, on Windows we have to make a single expensive syscall to get all the processes and then build the tree ourselves. Our algorithm to do so is recursive and quadratic in both steps and space with the number of processes on the host. For busy hosts this hits the stack limit and panics the Nomad client. We already build a map of parent PID to PID, so modify this to be a map of parent PID to slice of children and then traverse that tree only from the root we care about (the executor PID). This moves the allocations to the heap but makes the stats gathering linear in steps and space required. This changeset also moves as much of this code as possible into an area not conditionally-compiled by OS, as the tagged test file was not being run in CI. Fixes: https://github.com/hashicorp/nomad/issues/23984
32 lines
1.1 KiB
Go
32 lines
1.1 KiB
Go
// Copyright (c) HashiCorp, Inc.
|
|
// SPDX-License-Identifier: MPL-2.0
|
|
|
|
//go:build windows
|
|
|
|
package procstats
|
|
|
|
import (
|
|
"github.com/hashicorp/go-set/v3"
|
|
"github.com/mitchellh/go-ps"
|
|
)
|
|
|
|
// List will scan the process table and return a set of the process family
|
|
// tree starting with executorPID as the root.
|
|
//
|
|
// The implementation here specifically avoids using more than one system
|
|
// call. Unlike on Linux where we just read a cgroup, on Windows we must build
|
|
// the tree manually. We do so knowing only the child->parent relationships.
|
|
//
|
|
// So this turns into a fun leet code problem, where we invert the tree using
|
|
// only a bucket of edges pointing in the wrong direction. Basically we just
|
|
// iterate every process, recursively follow its parent, and determine whether
|
|
// executorPID is an ancestor.
|
|
//
|
|
// See https://github.com/hashicorp/nomad/issues/20042 as an example of what
|
|
// happens when you use syscalls to work your way from the root down to its
|
|
// descendants.
|
|
func List(executorPID int) set.Collection[ProcessID] {
|
|
procs, _ := list(executorPID, ps.Processes)
|
|
return procs
|
|
}
|