Skip to content

fix: do not try to use put_dimensions to merge dimension sets #5075

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.

Sign up for GitHub

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

Merged
merged 2 commits into from
Mar 10, 2025
Merged

fix: do not try to use put_dimensions to merge dimension sets #5075

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b700886
fix: do not try to use put_dimensions to merge dimension sets
roypat Mar 10, 2025
f45b2f5
Merge branch 'main' into put-dimensions-is-unintuitive
roypat Mar 10, 2025
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
74 changes: 44 additions & 30 deletions tests/integration_tests/performance/test_snapshot_ab.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,7 +44,7 @@ def id(self):
"""Computes a unique id for this test instance"""
return "all_dev" if self.all_devices else f"{self.vcpus}vcpu_{self.mem}mb"

def boot_vm(self, microvm_factory, guest_kernel, rootfs, metrics) -> Microvm:
def boot_vm(self, microvm_factory, guest_kernel, rootfs) -> Microvm:
"""Creates the initial snapshot that will be loaded repeatedly to sample latencies"""
vm = microvm_factory.build(
guest_kernel,
Expand Down Expand Up @@ -74,16 +74,6 @@ def boot_vm(self, microvm_factory, guest_kernel, rootfs, metrics) -> Microvm:
)
vm.api.vsock.put(vsock_id="vsock0", guest_cid=3, uds_path="/v.sock")

metrics.set_dimensions(
{
"net_devices": str(self.nets),
"block_devices": str(self.blocks),
"vsock_devices": str(int(self.all_devices)),
"balloon_devices": str(int(self.all_devices)),
"huge_pages_config": str(self.huge_pages),
**vm.dimensions,
}
)
vm.start()

return vm
Expand Down Expand Up @@ -115,15 +105,23 @@ def test_restore_latency(

We only test a single guest kernel, as the guest kernel does not "participate" in snapshot restore.
"""
vm = test_setup.boot_vm(microvm_factory, guest_kernel_linux_5_10, rootfs, metrics)

snapshot = vm.snapshot_full()
vm.kill()
vm = test_setup.boot_vm(microvm_factory, guest_kernel_linux_5_10, rootfs)

metrics.put_dimensions(
{"performance_test": "test_restore_latency", "uffd_handler": "None"}
metrics.set_dimensions(
{
"net_devices": str(test_setup.nets),
"block_devices": str(test_setup.blocks),
"vsock_devices": str(int(test_setup.all_devices)),
"balloon_devices": str(int(test_setup.all_devices)),
"huge_pages_config": str(test_setup.huge_pages),
"performance_test": "test_restore_latency",
"uffd_handler": "None",
**vm.dimensions,
}
)

snapshot = vm.snapshot_full()
vm.kill()
for microvm in microvm_factory.build_n_from_snapshot(snapshot, ITERATIONS):
value = 0
# Parse all metric data points in search of load_snapshot time.
Expand Down Expand Up @@ -151,7 +149,20 @@ def test_post_restore_latency(
pytest.skip("huge page snapshots can only be restored using uffd")

test_setup = SnapshotRestoreTest(mem=1024, vcpus=2, huge_pages=huge_pages)
vm = test_setup.boot_vm(microvm_factory, guest_kernel_linux_5_10, rootfs, metrics)
vm = test_setup.boot_vm(microvm_factory, guest_kernel_linux_5_10, rootfs)

metrics.set_dimensions(
{
"net_devices": str(test_setup.nets),
"block_devices": str(test_setup.blocks),
"vsock_devices": str(int(test_setup.all_devices)),
"balloon_devices": str(int(test_setup.all_devices)),
"huge_pages_config": str(test_setup.huge_pages),
"performance_test": "test_post_restore_latency",
"uffd_handler": str(uffd_handler),
**vm.dimensions,
}
)

vm.ssh.check_output(
"nohup /usr/local/bin/fast_page_fault_helper >/dev/null 2>&1 </dev/null &"
Expand All @@ -163,13 +174,6 @@ def test_post_restore_latency(
snapshot = vm.snapshot_full()
vm.kill()

metrics.put_dimensions(
{
"performance_test": "test_post_restore_latency",
"uffd_handler": str(uffd_handler),
}
)

for microvm in microvm_factory.build_n_from_snapshot(
snapshot, ITERATIONS, uffd_handler_name=uffd_handler
):
Expand All @@ -191,14 +195,24 @@ def test_population_latency(
):
"""Collects population latency metrics (e.g. how long it takes UFFD handler to fault in all memory)"""
test_setup = SnapshotRestoreTest(mem=128, vcpus=1, huge_pages=huge_pages)
vm = test_setup.boot_vm(microvm_factory, guest_kernel_linux_5_10, rootfs, metrics)
snapshot = vm.snapshot_full()
vm.kill()
vm = test_setup.boot_vm(microvm_factory, guest_kernel_linux_5_10, rootfs)

metrics.put_dimensions(
{"performance_test": "test_population_latency", "uffd_handler": "fault_all"}
metrics.set_dimensions(
{
"net_devices": str(test_setup.nets),
"block_devices": str(test_setup.blocks),
"vsock_devices": str(int(test_setup.all_devices)),
"balloon_devices": str(int(test_setup.all_devices)),
"huge_pages_config": str(test_setup.huge_pages),
"performance_test": "test_population_latency",
"uffd_handler": "fault_all",
**vm.dimensions,
}
)

snapshot = vm.snapshot_full()
vm.kill()

for microvm in microvm_factory.build_n_from_snapshot(
snapshot, ITERATIONS, uffd_handler_name="fault_all"
):
Expand Down