Set less memory to preload before pausing #702
Conversation
ben-grande
changed the title
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #702 +/- ##
==========================================
- Coverage 70.82% 70.53% -0.30%
==========================================
Files 61 61
Lines 13473 13546 +73
==========================================
+ Hits 9542 9554 +12
- Misses 3931 3992 +61
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
qubes/tools/qmemmand.py
Outdated
| resp = "FAIL" | ||
| domid, memory = str(data_args[0]), int(data_args[1]) | ||
| # TODO: ben: create the function to set memory to a domain. | ||
| if system_state.mem_set(domid, memory): |
There was a problem hiding this comment.
what was the issue with using system_state.mem_set? Using under if looks weird, as it always returns None, but otherwise looks correct?
What you might need to add is waiting for the domain to actually balloon down to the requested size - possibly with a simple loop that calls refresh_mem_actual and checks mem_actual, with some timeout.
There was a problem hiding this comment.
I didn't test with mem_set() because I saw that do_balloon was calling it with a loop, so mem_set() was there to give me hint of what to look next, I didn't test with only it.
I reused do_balloon() in order to avoid code duplication, but a lot of functions there use the class object self.dom_dict and that might cause some interference. I will try a new function with a simple loop, do_balloon() might indeed be overkill for a single domain changing values. I thought the xenfree value was necessary to be sure the memory was returned to Xen as well as reusing the waiting time and transfer speed to check again.
ben-grande
force-pushed
the
preload-qmemman
branch
3 times, most recently
from
8d86200 to
c514c51
Compare
|
The simplified loop helped. It is working! Preloaded 10 qubes at once and every qube has low memory assigned to it. Trying to start a "heavy" application such as Firefox doesn't crash. The use of a dictionary is to allow for multiple qubes to be setup at once: "0:0 1:0 2:0 3:0", but I haven't find a use case for that yet as preloads may be on different stages of preloading. If you think it complicates, I will remove it. |
ben-grande
force-pushed
the
preload-qmemman
branch
from
c514c51 to
1032e83
Compare
|
PipelineRetry |
OpenQA test summaryComplete test suite and dependencies: https://openqa.qubes-os.org/tests/overview?distri=qubesos&version=4.3&build=2025071903-4.3&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=2025061004-4.3&flavor=update
Failed tests10 failures
Fixed failuresCompared to: https://openqa.qubes-os.org/tests/142375#dependencies 10 fixed
Unstable testsPerformance TestsPerformance degradation:9 performance degradations
Remaining performance tests:63 tests
|
|
This needs a rebase now |
ben-grande
force-pushed
the
preload-qmemman
branch
from
1032e83 to
0d0d716
Compare
For: QubesOS/qubes-issues#1512
For: QubesOS/qubes-issues#9917