1
0
Fork 0
mirror of https://github.com/retspen/webvirtcloud synced 2026-11-09 03:36:48 +00:00
webvirtcloud/computes/views.py
catborise 1b99d2eac5 feat(computes): the host overview shows its VMs, capacity and load, also when it cannot be reached
A host the panel could not reach made its overview the 500 page,
without the host's sections, and the Computes list gave such a host an
Overview button that led nowhere. The page now keeps the host's name,
sections, "Edit connection" and "Remove from panel" (the existing
confirmation), and says "Not connected" with the reason and a "Try
again". Details that cannot be read are said in an alert; the VMs,
capacity and load still show, as they are read on their own.

A reachable host's page leads with what it runs and what it gave:
- its VMs by state (running, shut off, paused, other), linked to
  Instances;
- the vCPUs, memory and storage pools given, as on the Computes list,
  which now shares those parts (computes/capacity_*.html) and their
  preparation (views.shown_capacity) with it;
- CPU and memory in use, side by side, with the current reading in
  words, a note on what the charts hold and when they were last
  updated, or since when not; memory in GB like the rest of the page;
- Hardware and Software and connection: logical CPUs and clock from
  virNodeGetInfo, but physical cores, sockets and NUMA nodes counted
  from the CPUs of the capabilities' NUMA cells, as virNodeGetInfo
  makes a topology up for hosts it cannot describe (asymmetric NUMA);
  the hypervisor's own name and version, libvirt, emulator, the guest
  architectures as text instead of menus of dead links. A host without
  sysinfo shows no processor instead of failing.
Empty notes no longer show "None"; labels stack over their values on a
phone instead of running into them.

Pools that cannot be read no longer drop a host's vCPUs and memory from
its capacity: the storage part says so. get_capacity also counts the
VMs by state.
2026-10-27 14:51:18 +03:00

278 lines
10 KiB
Python

import json
from admin.decorators import superuser_only
from django.conf import settings
from django.core import signing
from django.http import HttpResponse
from django.shortcuts import get_object_or_404, redirect, render
from django.urls import reverse
from django.utils import timezone
from django.utils.translation import gettext_lazy as _
from instances.models import Instance
from libvirt import libvirtError
from logs.audit import audit, changed_fields, snapshot
from operations import claims
from vrtManager.hostdetails import wvmHostDetails
from computes.forms import (
SocketComputeForm,
SshComputeForm,
TcpComputeForm,
TlsComputeForm,
)
from computes.models import Compute
from . import utils
# The host CPU graph sends back the CPU sample of its previous answer; usage
# is taken since then when it is this recent, otherwise over one second.
CPU_SAMPLE_MAX_AGE = 15
CPU_SAMPLE_SIGNER = signing.TimestampSigner(salt="computes.cpu-sample")
def shown_capacity(compute):
"""The compute's capacity, with what its bars show; None when it cannot
be read."""
given = compute.capacity
if given:
# the memory bar: running and shut-off VMs' memory, and a mark at the
# host's memory, on a scale that holds whichever is larger
assigned, memory = given["assigned"], given["memory"]
scale = max(assigned["all"], memory) or 1
given["bar"] = {"running": assigned["running"] * 100 / scale,
"stopped": (assigned["all"] - assigned["running"]) * 100 / scale,
"host": memory * 100 / scale, "over": assigned["all"] > memory}
given["memory_percent"] = round(assigned["all"] * 100 / memory) if memory else 0
given["vm_count"] = sum(given["vms"].values())
# the default pool (else the first) shows; the others (a host may
# have dozens) open on a click
pools = given["pools"] or []
for pool in pools:
pool["used_percent"] = (pool["size"] - pool["free"]) * 100 / pool["size"] if pool["size"] else 0
given["pool"] = pools[0] if pools else None
given["more_pools"] = pools[1:]
return given
@superuser_only
def computes(request):
"""The computes, with what each has given its VMs."""
computes = list(Compute.objects.order_by("name"))
for compute in computes:
compute.shown_capacity = shown_capacity(compute)
return render(request, "computes/list.html", {"computes": computes})
@superuser_only
def capacity(request, compute_id):
"""One line of what the compute has given its VMs, as plain text, for the
new-VM dialog."""
compute = get_object_or_404(Compute, pk=compute_id)
return render(request, "computes/capacity_line.txt", {"given": compute.capacity}, content_type="text/plain; charset=utf-8")
@superuser_only
def overview(request, compute_id):
"""The host's page: whether the panel reaches it, its VMs, what it has
given them, its load, hardware and software. A host it cannot reach
keeps the page, with the reason and its connection to edit."""
compute = get_object_or_404(Compute, pk=compute_id)
# the connection itself, not a TCP probe: ssh aliases and ProxyJump hosts
# are not reachable by name and port from here
context = {"compute": compute, "connected": compute.status is True}
if context["connected"]:
try:
conn = wvmHostDetails(compute.hostname, compute.login, compute.password, compute.type)
hostname, arch, memory, _cpus, processor, uri = conn.get_node_info()
context["host"] = {
"hostname": hostname, "arch": arch, "memory": memory, "processor": processor, "uri": uri,
"topology": conn.get_cpu_topology(),
"hypervisor": conn.wvm.getType(), "hypervisor_version": conn.get_version(),
"libvirt_version": conn.get_lib_version(),
"emulator": conn.get_emulator(arch), "guest_archs": conn.get_hypervisors_domain_types(),
}
conn.close()
except libvirtError as error:
context["read_error"] = str(error)
context["given"] = shown_capacity(compute)
return render(request, "overview.html", context)
@superuser_only
def instances(request, compute_id):
compute = get_object_or_404(Compute, pk=compute_id)
utils.refresh_instance_database(compute)
instances = Instance.objects.filter(compute=compute).prefetch_related(
"userinstance_set"
)
return render(
request, "computes/instances.html",
{"compute": compute, "instances": instances, "job_resource": claims.compute_key(compute.id)},
)
@superuser_only
def compute_create(request, FormClass):
form = FormClass(request.POST or None)
if form.is_valid():
compute = form.instance
with audit(request, compute, f"Compute {compute.name} ({compute.hostname})", "add"):
form.save()
return redirect(reverse("computes"))
return render(request, "computes/form.html", {"form": form})
@superuser_only
def compute_update(request, compute_id):
compute = get_object_or_404(Compute, pk=compute_id)
if compute.type == 1:
FormClass = TcpComputeForm
elif compute.type == 2:
FormClass = SshComputeForm
elif compute.type == 3:
FormClass = TlsComputeForm
elif compute.type == 4:
FormClass = SocketComputeForm
subject = f"Compute {compute.name} ({compute.hostname})"
before = snapshot(compute)
form = FormClass(request.POST or None, instance=compute)
if form.is_valid():
with audit(request, compute, subject, "change") as entry:
form.save()
fields = changed_fields(before, snapshot(compute))
entry.action = "change " + ", ".join(fields)
entry.skip = not fields
return redirect(reverse("computes"))
return render(request, "computes/form.html", {"form": form})
@superuser_only
def compute_delete(request, compute_id):
compute = get_object_or_404(Compute, pk=compute_id)
if request.method == "POST":
with audit(request, compute, f"Compute {compute.name} ({compute.hostname})", "delete"):
compute.delete()
return redirect("computes")
return render(
request,
"common/confirm_delete.html",
{"object": compute,
"consequence": _("Its VMs keep running on the host and leave the panel; their owners are kept for "
"%(days)s days on the Removed VMs page and come back if the host is added again.")
% {"days": getattr(settings, "INSTANCE_OWNERSHIP_RETENTION_DAYS", 30)}},
)
def compute_graph(request, compute_id):
"""
:param request:
:param compute_id:
:return:
"""
comp_mgr = ComputeManager(compute_id)
data = comp_mgr.compute_graph(request.GET.get("cpu_sample"))
return HttpResponse(data, content_type="application/json")
def get_compute_disk_buses(request, compute_id, arch, machine, disk):
"""
:param request:
:param compute_id:
:param arch:
:param machine:
:param disk:
:return:
"""
comp_mgr = ComputeManager(compute_id)
return HttpResponse(comp_mgr.get_disk_buses(arch, machine, disk), content_type="application/json")
def get_compute_machine_types(request, compute_id, arch):
"""
:param request:
:param compute_id:
:param arch:
:return:
"""
comp_mgr = ComputeManager(compute_id)
return HttpResponse(comp_mgr.get_machine_types(arch), content_type="application/json")
class ComputeManager:
def __init__(self, compute_id):
self.compute = get_object_or_404(Compute, pk=compute_id)
def get_machine_types(self, arch):
data = dict()
try:
data["machines"] = self.compute.proxy.get_machine_types(arch)
except libvirtError:
pass
return json.dumps(data)
def get_disk_buses(self, arch, machine, disk):
data = dict()
try:
disk_device_types = self.compute.proxy.get_disk_device_types(arch, machine)
if disk in disk_device_types:
if disk == "disk":
data["bus"] = sorted(self.compute.proxy.get_disk_bus_types(arch, machine))
elif disk == "cdrom":
data["bus"] = ["ide", "sata", "scsi"]
elif disk == "floppy":
data["bus"] = ["fdc"]
elif disk == "lun":
data["bus"] = ["scsi"]
except libvirtError:
pass
return json.dumps(data)
def _cpu_sample(self, token):
"""The (idle, total, time) sample in a token of this compute's graph,
or None when there is none, it is altered or too old."""
if not token:
return None
try:
data = CPU_SAMPLE_SIGNER.unsign_object(token, max_age=CPU_SAMPLE_MAX_AGE)
except signing.BadSignature:
return None
if data.get("compute") != self.compute.id:
return None
return data["idle"], data["total"], data["time"]
def compute_graph(self, cpu_sample=None):
result = {"cpudata": None, "memdata": {"total": None, "usage": None}, "timeline": timezone.now().strftime("%H:%M:%S")}
try:
conn = wvmHostDetails(
self.compute.hostname,
self.compute.login,
self.compute.password,
self.compute.type,
)
cpu = conn.get_cpu_usage(previous=self._cpu_sample(cpu_sample))
idle, total, taken = cpu["sample"]
result = {
"cpudata": cpu["usage"],
"memdata": conn.get_memory_usage(),
"timeline": timezone.now().strftime("%H:%M:%S"),
"window": round(cpu["window"], 1),
"cpu_sample": CPU_SAMPLE_SIGNER.sign_object(
{"compute": self.compute.id, "idle": idle, "total": total, "time": taken}
),
}
conn.close()
except libvirtError:
pass
return json.dumps(result)