From 1b99d2eac5c86bc8e17b9d571eec86f6779d672b Mon Sep 17 00:00:00 2001 From: catborise Date: Tue, 27 Oct 2026 14:26:03 +0300 Subject: [PATCH] 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. --- computes/models.py | 14 +- .../templates/computes/capacity_memory.html | 10 + .../templates/computes/capacity_storage.html | 16 + .../templates/computes/capacity_vcpus.html | 5 + computes/templates/computes/list.html | 32 +- computes/templates/overview.html | 282 +++++++++++------- computes/tests/test_capacity.py | 15 +- computes/tests/test_cpu_graph.py | 23 +- computes/tests/test_overview.py | 94 ++++++ computes/tests/test_overview_status.py | 11 - computes/views.py | 86 +++--- .../arrow-clockwise_None_None_None.svg | 11 + static/icon_cache/x-circle_None_None_None.svg | 11 + vrtManager/hostdetails.py | 53 +++- vrtManager/tests/test_host_reads.py | 25 ++ 15 files changed, 487 insertions(+), 201 deletions(-) create mode 100644 computes/templates/computes/capacity_memory.html create mode 100644 computes/templates/computes/capacity_storage.html create mode 100644 computes/templates/computes/capacity_vcpus.html create mode 100644 computes/tests/test_overview.py create mode 100644 static/icon_cache/arrow-clockwise_None_None_None.svg create mode 100644 static/icon_cache/x-circle_None_None_None.svg diff --git a/computes/models.py b/computes/models.py index 407f63ec..ec7c769b 100644 --- a/computes/models.py +++ b/computes/models.py @@ -91,18 +91,22 @@ class Compute(Model): @cached_property def capacity(self): """What the host has given its VMs (wvmHostDetails.get_capacity), its - vCPUs per CPU, and its active storage pools, the default one first; or - None when it cannot be read.""" + vCPUs per CPU, and its active storage pools, the default one first + (None when the pools cannot be read); or None when the host cannot be + read.""" if self.status is not True: return None try: given = self.proxy.get_capacity() - storages = wvmStorages(self.hostname, self.login, self.password, self.type) - pools = [pool for pool in storages.get_storages_info() if pool["status"]] except libvirtError: return None given["ratio"] = given["vcpus"]["all"] / given["cpus"] if given["cpus"] else 0 - given["pools"] = sorted(pools, key=lambda pool: (pool["name"] != "default", pool["name"])) + try: + storages = wvmStorages(self.hostname, self.login, self.password, self.type) + pools = [pool for pool in storages.get_storages_info() if pool["status"]] + given["pools"] = sorted(pools, key=lambda pool: (pool["name"] != "default", pool["name"])) + except libvirtError: + given["pools"] = None return given def __str__(self): diff --git a/computes/templates/computes/capacity_memory.html b/computes/templates/computes/capacity_memory.html new file mode 100644 index 00000000..ff436b5b --- /dev/null +++ b/computes/templates/computes/capacity_memory.html @@ -0,0 +1,10 @@ +{% load i18n %} +{# the memory a host has given its VMs: given, from views.shown_capacity #} +
{% blocktrans with all=given.assigned.all|filesizeformat memory=given.memory|filesizeformat %}{{ all }} given of {{ memory }}{% endblocktrans %}
+ +
{% blocktrans with running=given.assigned.running|filesizeformat %}{{ running }} of it running{% endblocktrans %}
diff --git a/computes/templates/computes/capacity_storage.html b/computes/templates/computes/capacity_storage.html new file mode 100644 index 00000000..569b5d7e --- /dev/null +++ b/computes/templates/computes/capacity_storage.html @@ -0,0 +1,16 @@ +{% load i18n %} +{# a host's storage pools: given, from views.shown_capacity #} +{% if given.pools is None %} +
{% trans "The storage pools could not be read." %}
+{% elif not given.pool %} +
{% trans "No active storage pool." %}
+{% endif %} +{% if given.pool %}{% include "computes/pool_bar.html" with pool=given.pool %}{% endif %} +{% if given.more_pools %} +
+ {% blocktrans count counter=given.more_pools|length %}{{ counter }} more pool{% plural %}{{ counter }} more pools{% endblocktrans %} + {% for pool in given.more_pools %} +
{% include "computes/pool_bar.html" %}
+ {% endfor %} +
+{% endif %} diff --git a/computes/templates/computes/capacity_vcpus.html b/computes/templates/computes/capacity_vcpus.html new file mode 100644 index 00000000..2730fdae --- /dev/null +++ b/computes/templates/computes/capacity_vcpus.html @@ -0,0 +1,5 @@ +{% load i18n %} +{# the vCPUs a host has given its VMs: given, what its list cell and its overview show #} +
{{ given.ratio|floatformat:1 }}:1
+
{% blocktrans count counter=given.vcpus.all %}{{ counter }} vCPU{% plural %}{{ counter }} vCPUs{% endblocktrans %} {% blocktrans count counter=given.cpus %}on {{ counter }} CPU{% plural %}on {{ counter }} CPUs{% endblocktrans %}
+
{% blocktrans with running=given.vcpus.running %}{{ running }} running{% endblocktrans %}
diff --git a/computes/templates/computes/list.html b/computes/templates/computes/list.html index 2bc49784..95d77942 100644 --- a/computes/templates/computes/list.html +++ b/computes/templates/computes/list.html @@ -35,7 +35,7 @@ {% for compute in computes %} - {% with given=compute.capacity %} + {% with given=compute.shown_capacity %}
@@ -55,41 +55,21 @@ {% if given %} -
{{ given.ratio|floatformat:1 }}:1
-
{% blocktrans count counter=given.vcpus.all %}{{ counter }} vCPU{% plural %}{{ counter }} vCPUs{% endblocktrans %} {% blocktrans count counter=given.cpus %}on {{ counter }} CPU{% plural %}on {{ counter }} CPUs{% endblocktrans %}
-
{% blocktrans with running=given.vcpus.running %}{{ running }} running{% endblocktrans %}
+ {% include "computes/capacity_vcpus.html" %} -
{% blocktrans with all=given.assigned.all|filesizeformat memory=given.memory|filesizeformat %}{{ all }} given of {{ memory }}{% endblocktrans %}
- -
{% blocktrans with running=given.assigned.running|filesizeformat %}{{ running }} of it running{% endblocktrans %}
+ {% include "computes/capacity_memory.html" %} - {% if given.pool %}{% include "computes/pool_bar.html" with pool=given.pool %}{% endif %} - {% if given.more_pools %} -
- {% blocktrans count counter=given.more_pools|length %}{{ counter }} more pool{% plural %}{{ counter }} more pools{% endblocktrans %} - {% for pool in given.more_pools %} -
{% include "computes/pool_bar.html" %}
- {% endfor %} -
- {% endif %} + {% include "computes/capacity_storage.html" %} {% else %} {% if compute.status is True %}{% trans "The capacity could not be read." %}{% else %}{% trans "No capacity while the host cannot be reached." %}{% endif %} {% endif %}
- {% if compute.status is True %} - {% bs_icon 'eye-fill' %} - {% else %} - {% bs_icon 'eye' %} - {% endif %} + {# a host the panel cannot reach opens too: its page says why #} + {% bs_icon 'eye-fill' %} {% bs_icon 'pencil-fill' %} {% bs_icon 'x-circle-fill' %}
diff --git a/computes/templates/overview.html b/computes/templates/overview.html index efc5a40e..fda5fcb3 100644 --- a/computes/templates/overview.html +++ b/computes/templates/overview.html @@ -5,97 +5,154 @@ {% block title %}{% trans "Overview" %} - {{ compute.name }}{% endblock %} +{% block page_heading_extra %} +{% bs_icon 'pencil' %} {% trans "Edit connection" %} +{% bs_icon 'x-circle' %} {% trans "Remove from panel" %} +{% endblock page_heading_extra %} + {% block page_heading %}{{ compute.name }}{% endblock page_heading %} {% block content %} -
- {% include "computes/host_nav.html" with section="overview" %} -
+

+ + {% if connected %} + {% trans "Connected" %} + {% else %} + {% trans "Not connected" %} + {% endif %} + + {{ compute.hostname }} + {% if compute.details %}{{ compute.details }}{% endif %} +

-
-

{% trans "Basic details" %}

-
-
{% trans "Hostname" %}
-
{{ hostname }}
-
{% trans "Hypervisors" %}
-
-
- {% for arch, hpv in hypervisor.items|slice:":4" %} - -
- {% for h in hpv %} - {{ h }} - {% endfor %} -
- {% endfor %} -
-
- {% if hypervisor.items|length > 4 %} - - - {% endif %} -
-
-
{% trans "Emulator" %}
-
{{ emulator }}
-
{% trans "Version" %}
-
- {% trans 'Qemu' %} - {{ version }}   - {% trans 'Libvirt' %} - {{ lib_version }}   -
-
{% trans "Memory" %}
-
{{ host_memory|filesizeformat }}
-
{% trans "Architecture" %}
-
{{ host_arch }}
-
{% trans "Logical CPUs" %}
-
{{ logical_cpu }}
-
{% trans "Processor" %}
-
{{ model_cpu }}
-
{% trans "Connection" %}
-
{{ uri_conn }}
-
{% trans "Details" %}
-
{{ compute.details }}
+{% include "computes/host_nav.html" with section="overview" %} + +{% if not connected %} + +{% else %} + +{% if read_error %} + +{% endif %} + +

{% trans "VMs and what they were given" %}

+{% if given %} +
+
+
+
{% trans "VMs" %}
+
{{ given.vm_count }}
+
+ {% blocktrans with running=given.vms.running shutoff=given.vms.shutoff %}{{ running }} running, {{ shutoff }} shut off{% endblocktrans %}{% if given.vms.paused %}, {% blocktrans with paused=given.vms.paused %}{{ paused }} paused{% endblocktrans %}{% endif %}{% if given.vms.other %}, {% blocktrans with other=given.vms.other %}{{ other }} other{% endblocktrans %}{% endif %} +
+ {% trans "Instances" %} → +
+
+
+
+
{% trans "vCPUs" %}
+ {% include "computes/capacity_vcpus.html" %} +
+
+
+
+
{% trans "Memory" %}
+ {% include "computes/capacity_memory.html" %} +
+
+
+
+
{% trans "Storage" %}
+ {% include "computes/capacity_storage.html" %} + {% trans "Storages" %} → +
+
+
+{% else %} +

{% trans "The capacity could not be read." %}

+{% endif %} + +{# the load is read on its own: shown also when the details could not be #} +

{% trans "Load" %}

+
+
+
+
{% trans "CPU in use" %} –
+
+
+
+
+
+
{% trans "Memory in use" %} –
+
+
+
+
+

+ {% trans "The last 10 readings, one every 5 seconds. Memory in use leaves out the host's buffers and cache." %} + {% trans "Waiting for the first reading." %} +

+ +{% if host %} +
+
+

{% trans "Hardware" %}

+
+
{% trans "Hostname" %}
+
{{ host.hostname }}
+
{% trans "Architecture" %}
+
{{ host.arch }}
+
{% trans "Processor" %}
+
{{ host.processor|default:"–" }}
+
{% trans "Logical CPUs" %}
+
{{ host.topology.cpus }}
+
{% trans "Physical cores" %}
+
{{ host.topology.cores|default:"–" }}
+
{% trans "Sockets" %}
+
{{ host.topology.sockets|default:"–" }}
+
{% trans "NUMA nodes" %}
+
{{ host.topology.nodes|default:"–" }}
+
{% trans "Clock" %}
+
{{ host.topology.mhz }} MHz
+
{% trans "Memory" %}
+
{{ host.memory|filesizeformat }}
- -

{% trans "Performance" %}

-
-
-
-
- {% bs_icon 'arrow-right' %} - {% trans "CPU Utilization" %} -
-
-
-
- -
-
-
- {% bs_icon 'arrow-right'%} {% trans "RAM Utilization" %} -
-
-
-
-
-
+
+
+

{% trans "Software and connection" %}

+
+
{% trans "Hypervisor" %}
+
{{ host.hypervisor }} {{ host.hypervisor_version }}
+
libvirt
+
{{ host.libvirt_version }}
+
{% trans "Emulator" %}
+
{{ host.emulator|default:"–" }}
+
{% trans "Guest architectures" %}
+
+ {% for arch, types in host.guest_archs.items %} +
{{ arch }} ({{ types|join:", " }})
+ {% endfor %} +
+
{% trans "Connection" %}
+
{{ host.uri }}
+
+
+
+{% endif %} +{% endif %} {% endblock %} + {% block script %} +{% if connected %} +{% endif %} {% endblock %} diff --git a/computes/tests/test_capacity.py b/computes/tests/test_capacity.py index 785a7ad3..d7cb6107 100644 --- a/computes/tests/test_capacity.py +++ b/computes/tests/test_capacity.py @@ -11,6 +11,7 @@ from libvirt import libvirtError from lxml import html GIB = 1024 ** 3 +VMS = {"running": 3, "paused": 1, "shutoff": 2, "other": 0} def pool(name, size_gib, free_gib, status=1): @@ -33,6 +34,14 @@ class CapacityTests(TestCase): "cpus": 16, "vcpus": {"all": 24, "running": 8}, "ratio": 1.5, "pools": [pool("default", 500, 120), pool("iso", 100, 90), pool("nfs", 900, 800)]}) + @patch("computes.models.wvmStorages") + @patch("computes.models.wvmHostDetails") + def test_pools_that_cannot_be_read_leave_the_cpus_and_memory(self, details, storages): + details.return_value.get_capacity.return_value = {"cpus": 16, "vcpus": {"all": 24, "running": 8}} + storages.return_value.get_storages_info.side_effect = libvirtError("cannot refresh pool") + self.assertEqual(self.compute.capacity, { + "cpus": 16, "vcpus": {"all": 24, "running": 8}, "ratio": 1.5, "pools": None}) + @patch("computes.models.wvmHostDetails") def test_a_host_whose_read_fails_has_no_capacity(self, details): # a VM that goes while the host is read, for one @@ -53,7 +62,7 @@ class ComputesPageTests(TestCase): c1 = Compute.objects.create(name="host-one", hostname="192.0.2.1", login="u", password="p", type=1) Compute.objects.create(name="host-two", hostname="192.0.2.2", login="u", password="p", type=1) found = {c1.id: {"cpus": 16, "memory": 64 * GIB, "vcpus": {"all": 40, "running": 24}, - "assigned": {"all": 48 * GIB, "running": 32 * GIB}, "ratio": 2.5, + "assigned": {"all": 48 * GIB, "running": 32 * GIB}, "ratio": 2.5, "vms": VMS, "pools": [pool("default", 500, 120)]}} with patch.object(Compute, "capacity", new=property(lambda c: found.get(c.id))), \ patch.object(Compute, "status", new=True): @@ -77,7 +86,7 @@ class ComputesPageTests(TestCase): Compute.objects.create(name="host-one", hostname="192.0.2.1", login="u", password="p", type=1) pools = [pool("default", 500, 120), pool("backup", 900, 800), pool("iso", 100, 90), pool("nfs", 2000, 1000)] given = {"cpus": 16, "memory": 64 * GIB, "vcpus": {"all": 8, "running": 8}, "ratio": 0.5, - "assigned": {"all": 8 * GIB, "running": 8 * GIB}, "pools": pools} + "assigned": {"all": 8 * GIB, "running": 8 * GIB}, "pools": pools, "vms": VMS} with patch.object(Compute, "capacity", new=given), patch.object(Compute, "status", new=True): page = html.fromstring(self.client.get(reverse("computes")).content) (storage,) = page.xpath("//tbody/tr/td[4]") @@ -95,7 +104,7 @@ class SingularTests(TestCase): self.client.force_login(admin) c = Compute.objects.create(name="host-one", hostname="192.0.2.1", login="u", password="p", type=1) given = {"cpus": 1, "memory": 4 * GIB, "vcpus": {"all": 1, "running": 1}, - "assigned": {"all": GIB, "running": GIB}, "ratio": 1.0, "pools": []} + "assigned": {"all": GIB, "running": GIB}, "ratio": 1.0, "pools": [], "vms": VMS} with patch.object(Compute, "capacity", new=given), patch.object(Compute, "status", new=True): page = html.fromstring(self.client.get(reverse("computes")).content) line = self.client.get(reverse("compute_capacity", args=[c.id])).content.decode() diff --git a/computes/tests/test_cpu_graph.py b/computes/tests/test_cpu_graph.py index bbb509f1..30ed4d57 100644 --- a/computes/tests/test_cpu_graph.py +++ b/computes/tests/test_cpu_graph.py @@ -24,6 +24,7 @@ from vrtManager.util import ConnectionFailed # the page script runs after the shared chart helpers CHARTS_JS = settings.BASE_DIR / "static" / "js" / "charts.js" +TOPOLOGY = {"nodes": 1, "sockets": 1, "cores": 2, "threads": 1, "mhz": 2400, "cpus": 2} def host(*readings): @@ -138,17 +139,20 @@ class GraphScriptTestCase(TestCase): compute = Compute.objects.create(name="gs", hostname="192.0.2.1", login="u", password="p", type=1) with patch("computes.views.wvmHostDetails") as details, patch.object( Compute, "status", new_callable=PropertyMock, return_value=True - ): + ), patch.object(Compute, "capacity", new=None): details.return_value.get_node_info.return_value = ("h", "x86_64", 1024, 2, "cpu", "qemu+tcp://h/system") - details.return_value.get_memory_usage.return_value = {"total": 1024, "usage": 512, "percent": 50} + details.return_value.get_cpu_topology.return_value = TOPOLOGY page = self.client.get(reverse("overview", args=[compute.id])).content.decode() script = next(s for s in html.fromstring(page).xpath("//script[not(@src)]/text()") if "pollWhileVisible(" in s) js = """ const vm = require("vm"); const charts = [], requests = []; + const element = (dataset) => ({dataset, textContent: "", classList: {add: () => {}, remove: () => {}}}); + const elements = {graph_state: element({updated: "Updated at %(time)s.", stale: "Not updated since %(time)s."}), + cpu_now: element({}), mem_now: element({})}; let poll; const page = { - document: {body: {}, getElementById: () => ({})}, + document: {body: {}, getElementById: (id) => elements[id] || {}}, // Chart.js adds the lists a config leaves out getComputedStyle: () => ({getPropertyValue: () => ""}), Chart: Object.assign(function (canvas, config) { @@ -158,21 +162,23 @@ class GraphScriptTestCase(TestCase): this.update = () => {}; charts.push(this); }, {defaults: {}}), - $: {getJSON: (url, params) => { requests.push(params); return {}; }}, + $: {getJSON: (url, params) => { requests.push(params); return {fail() { return this; }}; }}, pollWhileVisible: (interval, request, onData) => { poll = {request, onData}; }, }; const context = vm.createContext(page); vm.runInContext(CHARTS, context); vm.runInContext(SCRIPT, context); const answer = (cpu, sample) => ({cpudata: cpu, window: 5.1, cpu_sample: sample, timeline: "t", - memdata: {total: 1048576, usage: 0}}); + memdata: {total: 2 * 1073741824, usage: 1073741824}}); poll.request(); poll.onData(answer(3, "token-1")); + const shown = [elements.cpu_now.textContent, elements.mem_now.textContent, elements.graph_state.textContent]; poll.request(); poll.onData({cpudata: null, memdata: {total: null, usage: null}, timeline: "t"}); poll.request(); const label = charts[0].options.plugins.tooltip.callbacks.label({dataset: {label: "Usage"}, formattedValue: "3", dataIndex: 0}); - console.log(JSON.stringify({requests, label, cpu: charts[0].data.datasets[0].data, + console.log(JSON.stringify({requests, label, shown, state: elements.graph_state.textContent, + cpu: charts[0].data.datasets[0].data, mem: charts[1].data.datasets[0].data, memMax: charts[1].options.scales.y.max})); """.replace("SCRIPT", json.dumps(script)).replace("CHARTS", json.dumps(CHARTS_JS.read_text())) out = subprocess.run(["node", "-e", js], capture_output=True, text=True, timeout=30) @@ -181,5 +187,8 @@ class GraphScriptTestCase(TestCase): self.assertEqual(result["requests"], [{}, {"cpu_sample": "token-1"}, {}]) self.assertEqual(result["cpu"], [3, None]) # a host that could not be read leaves a gap and keeps the memory axis - self.assertEqual((result["mem"], result["memMax"]), ([0, None], 1)) + self.assertEqual((result["mem"], result["memMax"]), ([1, None], 2)) + # the readings in words, and since when they are not updated + self.assertEqual(result["shown"], ["3 %", "1 / 2 GB", "Updated at t."]) + self.assertEqual(result["state"], "Not updated since t.") self.assertTrue(result["label"].endswith(": 3 % (average over 5.1 s)"), result["label"]) diff --git a/computes/tests/test_overview.py b/computes/tests/test_overview.py new file mode 100644 index 00000000..e482efc8 --- /dev/null +++ b/computes/tests/test_overview.py @@ -0,0 +1,94 @@ +"""The host overview page: a host the panel reaches, one it cannot reach, +and one whose details cannot be read.""" + +from unittest.mock import PropertyMock, patch + +from django.contrib.auth import get_user_model +from django.test import TestCase +from django.urls import reverse +from libvirt import libvirtError +from lxml import html + +from computes.models import Compute + +GIB = 1024 ** 3 +GIVEN = {"cpus": 8, "memory": 32 * GIB, "vcpus": {"all": 12, "running": 6}, + "assigned": {"all": 24 * GIB, "running": 12 * GIB}, "ratio": 1.5, + "vms": {"running": 3, "paused": 1, "shutoff": 2, "other": 0}, + "pools": [{"name": "default", "status": 1, "size": 500 * GIB, "free": 120 * GIB}]} + + +def text(page, xpath): + return " ".join(" ".join(page.xpath(xpath)[0].itertext()).split()) + + +class OverviewTestCase(TestCase): + def setUp(self): + self.client.force_login(get_user_model().objects.create_superuser("ov-admin", "ov@example.com", "pw")) + self.compute = Compute.objects.create(name="host-one", hostname="192.0.2.1", login="u", password="p", + type=1, details="rack 4") + self.url = reverse("overview", args=[self.compute.id]) + + def page(self, connected=True, given=GIVEN): + with patch.object(Compute, "status", new_callable=PropertyMock, return_value=connected), \ + patch.object(Compute, "connection_error", new_callable=PropertyMock, + return_value="Connection Failed: 192.0.2.1 did not answer within 5 s"), \ + patch.object(Compute, "capacity", new=given and {**given}): + response = self.client.get(self.url) + self.assertEqual(response.status_code, 200) + return html.fromstring(response.content) + + @patch("computes.views.wvmHostDetails") + def test_a_host_shows_its_vms_capacity_hardware_and_software(self, details): + conn = details.return_value + conn.get_node_info.return_value = ("wvc-host", "x86_64", 32 * GIB, 8, "Xeon", "qemu+tcp://192.0.2.1/system") + conn.get_cpu_topology.return_value = {"nodes": 2, "sockets": 2, "cores": 4, "mhz": 2400, "cpus": 8} + conn.wvm.getType.return_value = "QEMU" + conn.get_version.return_value = "10.1.0" + conn.get_lib_version.return_value = "11.10.0" + conn.get_emulator.return_value = "/usr/libexec/qemu-kvm" + conn.get_hypervisors_domain_types.return_value = {"i686": ["kvm", "qemu"], "x86_64": ["kvm", "qemu"]} + page = self.page() + self.assertIn("Connected", text(page, "//p[span]")) + self.assertIn("rack 4", text(page, "//p[span]")) + cards = [" ".join(" ".join(card.itertext()).split()) for card in page.xpath("//div[@class='card-body' or contains(@class,'card-body ')]")] + self.assertIn("VMs 6 3 running, 2 shut off, 1 paused Instances →", cards) + self.assertTrue(any(card.startswith("vCPUs 1.5:1 12 vCPUs on 8 CPUs 6 running") for card in cards), cards) + self.assertTrue(any("24.0 GB given of 32.0 GB" in card for card in cards), cards) + self.assertTrue(any("120.0 GB free of 500.0 GB" in card for card in cards), cards) + hardware = text(page, "//h4[.='Hardware']/following-sibling::dl") + self.assertIn("Logical CPUs 8 Physical cores 4 Sockets 2", hardware) + self.assertIn("NUMA nodes 2", hardware) + software = text(page, "//h4[.='Software and connection']/following-sibling::dl") + self.assertIn("Hypervisor QEMU 10.1.0", software) + self.assertIn("x86_64 (kvm, qemu)", software) + # the charts poll only for a host that was read + self.assertTrue(page.xpath("//canvas[@id='cpuChart']")) + + @patch("computes.views.wvmHostDetails") + def test_a_host_the_panel_cannot_reach_keeps_its_page_with_the_reason(self, details): + page = self.page(connected=False, given=None) + details.assert_not_called() + self.assertIn("Not connected", text(page, "//p[span]")) + self.assertIn("did not answer within 5 s", text(page, "//div[@role='alert']")) + # the way back: its sections, its connection and its removal + self.assertTrue(page.xpath("//nav[@aria-label='Host sections']")) + self.assertTrue(page.xpath(f"//a[@href='{reverse('compute_update', args=[self.compute.id])}']")) + self.assertTrue(page.xpath(f"//a[@href='{reverse('compute_delete', args=[self.compute.id])}']")) + self.assertFalse(page.xpath("//canvas")) + + @patch("computes.views.wvmHostDetails") + def test_details_that_cannot_be_read_leave_the_capacity_and_the_load(self, details): + details.return_value.get_node_info.return_value = ("h", "x86_64", GIB, 2, None, "qemu+tcp://192.0.2.1/system") + details.return_value.get_lib_version.side_effect = libvirtError("cannot read the version") + page = self.page() + self.assertIn("cannot read the version", text(page, "//div[@role='alert']")) + self.assertTrue(page.xpath("//h5[.='VMs']")) + # the load is read on its own, by the charts + self.assertTrue(page.xpath("//canvas[@id='cpuChart']")) + self.assertFalse(page.xpath("//h4[.='Hardware']")) + + def test_the_list_links_the_overview_of_a_host_it_cannot_reach(self): + with patch.object(Compute, "status", new_callable=PropertyMock, return_value=False): + page = html.fromstring(self.client.get(reverse("computes")).content) + self.assertTrue(page.xpath(f"//a[@title='Overview'][@href='{self.url}']")) diff --git a/computes/tests/test_overview_status.py b/computes/tests/test_overview_status.py index 54074c20..d181cf4e 100644 --- a/computes/tests/test_overview_status.py +++ b/computes/tests/test_overview_status.py @@ -17,17 +17,6 @@ class OverviewStatusTests(TestCase): # an ssh_config alias: not a name a TCP probe can reach self.compute = Compute.objects.create(name="ov-alias", hostname="wvc-alias", login="root", password="", type=2) - @patch("vrtManager.connection.connection_manager.host_is_up", return_value=False) - @patch("computes.views.wvmHostDetails") - def test_overview_state_comes_from_the_connection(self, host_details, host_is_up): - conn = host_details.return_value - conn.get_node_info.return_value = ("host", "x86_64", 1024, 2, "cpu", "qemu+ssh://root@wvc-alias/system") - conn.get_memory_usage.return_value = {"total": 1024, "usage": 512, "percent": 50} - with patch.object(Compute, "status", new_callable=PropertyMock, return_value=True): - response = self.client.get(reverse("overview", args=[self.compute.id])) - self.assertEqual(response.status_code, 200) - self.assertEqual(response.context["status"], "true") - def test_compute_list_shows_why_a_host_is_not_connected(self): with patch.object(Compute, "status", new_callable=PropertyMock, return_value=False), patch.object( Compute, "connection_error", new_callable=PropertyMock, return_value=ERROR diff --git a/computes/views.py b/computes/views.py index 62e246ef..a941875a 100644 --- a/computes/views.py +++ b/computes/views.py @@ -30,28 +30,36 @@ 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: - 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 - for pool in given["pools"]: - pool["used_percent"] = (pool["size"] - pool["free"]) * 100 / pool["size"] if pool["size"] else 0 - # the default pool (else the first) shows; the others (a host may - # have dozens) open on a click - pools = given["pools"] - given["pool"] = pools[0] if pools else None - given["more_pools"] = pools[1:] + compute.shown_capacity = shown_capacity(compute) return render(request, "computes/list.html", {"computes": computes}) @@ -65,33 +73,29 @@ def capacity(request, compute_id): @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 - status = "true" if compute.status else "false" - - conn = wvmHostDetails( - compute.hostname, - compute.login, - compute.password, - compute.type, - ) - ( - hostname, - host_arch, - host_memory, - logical_cpu, - model_cpu, - uri_conn, - ) = conn.get_node_info() - hypervisor = conn.get_hypervisors_domain_types() - mem_usage = conn.get_memory_usage() - emulator = conn.get_emulator(host_arch) - version = conn.get_version() - lib_version = conn.get_lib_version() - conn.close() - - return render(request, "overview.html", locals()) + 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 diff --git a/static/icon_cache/arrow-clockwise_None_None_None.svg b/static/icon_cache/arrow-clockwise_None_None_None.svg new file mode 100644 index 00000000..23dd7e18 --- /dev/null +++ b/static/icon_cache/arrow-clockwise_None_None_None.svg @@ -0,0 +1,11 @@ + + + + + + + + + + + diff --git a/static/icon_cache/x-circle_None_None_None.svg b/static/icon_cache/x-circle_None_None_None.svg new file mode 100644 index 00000000..cb45faf9 --- /dev/null +++ b/static/icon_cache/x-circle_None_None_None.svg @@ -0,0 +1,11 @@ + + + + + + + + + + + diff --git a/vrtManager/hostdetails.py b/vrtManager/hostdetails.py index 3de53687..08f30c65 100644 --- a/vrtManager/hostdetails.py +++ b/vrtManager/hostdetails.py @@ -1,11 +1,26 @@ import time -from libvirt import VIR_DOMAIN_SHUTOFF, VIR_NODE_CPU_STATS_ALL_CPUS +from libvirt import ( + libvirtError, + VIR_DOMAIN_BLOCKED, + VIR_DOMAIN_PAUSED, + VIR_DOMAIN_RUNNING, + VIR_DOMAIN_SHUTOFF, + VIR_NODE_CPU_STATS_ALL_CPUS, +) from vrtManager.connection import wvmConnect from vrtManager.util import get_xml_path +# a VM's state as the host's VM counts group it; others count as "other" +VM_STATES = { + VIR_DOMAIN_RUNNING: "running", + VIR_DOMAIN_BLOCKED: "running", # running, waiting on a resource + VIR_DOMAIN_PAUSED: "paused", + VIR_DOMAIN_SHUTOFF: "shutoff", +} + # the times of virNodeGetCPUStats that add up to the elapsed CPU time; guest is # already part of user and utilization is a percentage CPU_TIMES = ("kernel", "user", "idle", "iowait") @@ -78,23 +93,47 @@ class wvmHostDetails(wvmConnect): info.append(self.host_info[0]) # architecture info.append(self.host_info[1] * 1048576) # memory info.append(self.host_info[2]) # cpu core count - info.append( - get_xml_path(self.wvm.getSysinfo(0), func=cpu_version) - ) # cpu version + try: + info.append(get_xml_path(self.wvm.getSysinfo(0), func=cpu_version)) # cpu version + except libvirtError: # not every driver or host gives its sysinfo + info.append(None) info.append(self.wvm.getURI()) # uri return info def get_capacity(self): """ What the host has given its VMs: vCPUs and memory (bytes) of all VMs - and of the running ones, against its CPUs and memory. A shut-off VM - takes its share again when it starts. + and of the running ones (paused ones too: they keep theirs), against + its CPUs and memory. A shut-off VM takes its share again when it + starts. Also its VMs by state: running, paused, shut off and other + (crashed, suspended, shutting down). """ vcpus = {"all": 0, "running": 0} assigned = {"all": 0, "running": 0} + vms = {"running": 0, "paused": 0, "shutoff": 0, "other": 0} for vm in self.get_host_instances(raw_mem_size=True).values(): for share in ("all",) if vm["status"] == VIR_DOMAIN_SHUTOFF else ("all", "running"): vcpus[share] += int(vm["vcpu"]) assigned[share] += vm["memory"] + vms[VM_STATES.get(vm["status"], "other")] += 1 return {"cpus": self.host_info[2], "memory": self.host_info[1] * 1048576, - "vcpus": vcpus, "assigned": assigned} + "vcpus": vcpus, "assigned": assigned, "vms": vms} + + def get_cpu_topology(self): + """ + The host's logical CPUs and MHz, and from the CPUs of its NUMA cells + in its capabilities, the NUMA nodes, sockets and physical cores (None + when it does not say). Not virNodeGetInfo's topology: it makes one up + for hosts it cannot describe, such as asymmetric NUMA. + """ + + def topology(doc): + cpus = doc.xpath("/capabilities/host/topology/cells/cell/cpus/cpu[@socket_id]") + return { + "nodes": len(doc.xpath("/capabilities/host/topology/cells/cell")) or None, + "sockets": len({cpu.get("socket_id") for cpu in cpus}) or None, + # a core is one within its socket and die + "cores": len({(cpu.get("socket_id"), cpu.get("die_id"), cpu.get("core_id")) for cpu in cpus}) or None, + } + + return {**get_xml_path(self.get_cap_xml(), func=topology), "cpus": self.host_info[2], "mhz": self.host_info[3]} diff --git a/vrtManager/tests/test_host_reads.py b/vrtManager/tests/test_host_reads.py index 9ada59de..f544fb28 100644 --- a/vrtManager/tests/test_host_reads.py +++ b/vrtManager/tests/test_host_reads.py @@ -158,10 +158,35 @@ class CapacityTestCase(unittest.TestCase): "cpus": 4, "memory": 2 * gib, "vcpus": {"all": 14, "running": 6}, "assigned": {"all": 14 * gib, "running": 6 * gib}, + "vms": {"running": 1, "paused": 1, "shutoff": 1, "other": 0}, }) conn.get_host_instances.assert_called_once_with(raw_mem_size=True) +class CpuTopologyTestCase(unittest.TestCase): + """The topology comes from the capabilities' NUMA cells: getInfo makes + one up (one node, one socket, every CPU a core) for hosts it cannot + describe.""" + + def topology(self, cells): + conn = connection(wvmHostDetails) + conn.__dict__["host_info"] = ["x86_64", 65536, 16, 2400, 1, 1, 16, 1] # getInfo's made-up one + conn.get_cap_xml = lambda: f"{cells}" + return conn.get_cpu_topology() + + def test_sockets_and_cores_are_counted_from_the_cpus_of_the_numa_cells(self): + # two nodes, two sockets, four cores of two threads each + cells = "".join( + f"" + + "".join(f"" for i in range(4)) + + "" for node in range(2)) + self.assertEqual(self.topology(cells), {"nodes": 2, "sockets": 2, "cores": 4, "cpus": 16, "mhz": 2400}) + + def test_a_host_that_does_not_say_has_no_sockets_or_cores(self): + cells = "" + self.assertEqual(self.topology(cells), {"nodes": 1, "sockets": None, "cores": None, "cpus": 16, "mhz": 2400}) + + class StoragesInfoTestCase(unittest.TestCase): def test_each_pool_tells_its_free_space(self): conn = connection(wvmStorages)