tests: reduce maxmem of HVM when running in nested virt#842
Conversation
Nested EPT is super slow, which makes the initial balloon down very long
- sometimes over a minute. PVH are not affected, as those are running
with memory hotplug instead of populate-on-demand on boot, and balloon
down doesn't happen. But HVM are incompatible with memory hotplug
(looks like QEMU limitation), so initial balloon down does happen
there. And before that, Linux thinks it has up to maxmem memory and will
try to clear it. This may look like this for example:
[ 0.879054] mem auto-init: stack:all(zero), heap alloc:on, heap free:on
[ 0.879060] mem auto-init: clearing system memory may take some time...
[ 86.287478] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=2, Nodes=1
Reduce the impact when running tests in nested virt by reducing maxmem
to 1G in those cases.
QubesOS/qubes-issues#10723
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #842 +/- ##
==========================================
- Coverage 70.35% 70.34% -0.01%
==========================================
Files 61 61
Lines 14157 14131 -26
==========================================
- Hits 9960 9941 -19
+ Misses 4197 4190 -7
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
OpenQA test summaryComplete test suite and dependencies: https://openqa.qubes-os.org/tests/overview?distri=qubesos&version=4.3&build=2026071518-devel&flavor=pull-requests Test run included the following:
New failures, excluding unstableCompared to: https://openqa.qubes-os.org/tests/overview?distri=qubesos&version=4.3&build=2026050504-devel&flavor=update
Failed tests13 failures
Fixed failuresCompared to: https://openqa.qubes-os.org/tests/176874#dependencies 32 fixed
Unstable testsDetails
Performance TestsPerformance degradation:16 performance degradations
Remaining performance tests:95 tests
|
|
It works! Those seems to be related to this PR:
|
Nested EPT is super slow, which makes the initial balloon down very long
sometimes over a minute. PVH are not affected, as those are running
with memory hotplug instead of populate-on-demand on boot, and balloon
down doesn't happen. But HVM are incompatible with memory hotplug
(looks like QEMU limitation), so initial balloon down does happen
there. And before that, Linux thinks it has up to maxmem memory and will
try to clear it. This may look like this for example:
[ 0.879054] mem auto-init: stack:all(zero), heap alloc:on, heap free:on
[ 0.879060] mem auto-init: clearing system memory may take some time...
[ 86.287478] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=2, Nodes=1
Reduce the impact when running tests in nested virt by reducing maxmem
to 1G in those cases.
QubesOS/qubes-issues#10723