[4.1.0.23] Heavy power drain due to connected WLAN

Continuing the discussion from [release notes] Kvarken 4.1.0:

As reported (and confirmed by other) my X suffers from a heavy power drain.
I tested a lot and it is absolutely only happening when WLAN is connected.

It is a more or less plain vanilla device (storeman, sfos-upgrade, harbour-defender) for testing.
Upgraded from 3.3 up to 4.1.

One mail account, set to sync manual.

Tested with all radio on, only network, network and cell data, WLAN enabled (but not connected) and other combinations. But it boils down to WLAN connected.

Top does not reveal anything.

Any more info I can give?

3 Likes

In my case, the device was updated from 4.0.1 (no power drain) to 4.1.0 (heavy power drain), so it must be related to a change between 4.0.1 and 4.1.0. And indeed, the only factor that has any effect is whether or not WLAN is enabled. The power drain is the same, regardless of whether there are any accounts syncing or not, whether Android support is running or not, whether applications are running or not, whether background services are running or not, etc. Only enabling WLAN causes the power drain. I think it occurs on both 2.4GHz and 5GHz Wi-Fi networks.

WLAN enabled → heavy power drain
WLAN disabled → no power drain

The fact that the regression occurred in between 4.0.1.48 and 4.1.0.23, and that the WLAN connection is causing (or is a primary factor in) the power drain, should hopefully narrow down which of the changes between the versions led to the regression.

1 Like

Where is that a settable option for the jolla email client? I don’t see it in the app or from the System-App-Mail settings page

You’re configuring this under account settings.

1 Like

Thanks. It’s so long ago I set it up that I had forgotten those options were there.

And this one is just another (bad) example why
I asked for a better bug handling (tracking/tagging) that needs to be done by @Jolla as long as they keep hiding their internal bug tracking system (which is okay!) but do not open up a public one or start doing work on this forum’s bug reports.

I have no hard data yet to back it up, but I’m experimenting with a self-built connmand playing around with the stats gathering.

connmand keeps transfer statistics and writes them to /home every couple of seconds.

Increasing those intervals, and adding one backport fix about logind connection seems to make wifi power drain a little better.

But I’ve been running that only since yesterday so I might be wrong.

When you say connected, are you logged in via ssh?
I have seen differences in power usage with active ssh sessions, and without.

Yes, might be the release 4.3.0 decreased power hungryness (? or just hunger?).

Right now it is down from 3-4%/h to ca. 2%/h.

Nope. No ssh.
Just plain and simple laying around connected to WiFi AP, no usage and no connection at all (maybe one standard set up mail IMAP account, sync all 39 min).