Hi Andres/Nazir,

On Mon, Jun 1, 2026 at 11:57 PM Andres Freund <[email protected]> wrote:
>
> Hi,
>
> Attached is an large incremental patch onto Bilal's version:
[..]

> Thoughts?

The timings I got during morning EMEA hours [1] (my first GHA run) was
with all those commits except last one, and it's LGTM (it's a flawless
experience
at this point as end user; all succeed). Perhaps my only worry would be
success of this and proliferation of using this often, where plenty of people
hit download those packages at runtime (MacOS/Win) way too often (or
download sites being unavailable and causing outages in the pipeline).

Continuing on previous story...:
Windows was still @ 31mins, and whatever I've tried it is was not helping it
(but I cannot measure inside GHA Runner what was happening, so those were blind
shots with fstweaks, etc). One important thing, altough I failed altering
CacheIsPowerProtected (avoid flushing the write cache) as it seems impossible
for me to do so on D:\ (as paging file is there and and altering it also
requires reboot), at least we know stuff is way slower than it could be on
those runners:

"Get-PhysicalDisk | Get-StorageAdvancedProperty" reported:

FriendlyName      SerialNumber IsPowerProtected IsDeviceCacheEnabled
------------      ------------ ---------------- --------------------
Msft Virtual Disk                         False                False
Msft Virtual Disk                         False                False

Perhaps there's way to use some custom image/templ with different settings,
especially for D:\, after all it's just volatile stuff. Thoughts? (not that I
care that much for Win, but waiting half hour for it finish every time is
not going to be nice...)

Also, maybe that's not useful for for GHA/CI, but for folks trying to local test
on win32:

I've run some quick test on my own Windows VM with manual "ninja test".
Immediatley spotted Defener as top#1 CPU. In GHA workflow we already
this thing for disabling Defender AV:
    "Set-MpPreference -DisableRealtimeMonitoring $true
        -SubmitSamplesConsent NeverSend -MAPSReporting Disable"

Yet, when I re-run the verification command it showed still as as enabled (!)
and I've still got Defender as top#1 CPU during next ninja run, so disabling
RealTimeMonitoring didn't kick in? So I went and I've manually disabled
"real-time protection" in "Virus & thread protection settings" in the control
panel as it was still on after above and only then it got a nice boost
(even visually when watching tests). Gemini told me later that there's
"TamperProtection" thing (!) to ignore powershell. see e.g.
    "Get-MpComputerStatus | Select-Object IsTamperProtected"
and it requires manual human steps to disarm... but apparently all of this is
not necessary on GHA...

-J.

[1] - 
https://github.com/jakubwartakEDB/postgres-public-ci/actions/runs/26806653438


Reply via email to