Severely increased battery drain after 1.0.4.20 update

The "no account" bug is not the main problem. With one or more active accounts power drain varies wildly depending on the type of network connection. I've got the impression that (any) sync + no WLAN + a flaky 3G connection can trigger battery drain problems.

Seems related to the syncing indeed, on one or more of my accounts. See my previous comment under "At least mail sync..." answer, but I tried last night by having Sync settings disabled for FB, Twitter and Google accounts. The result is a drop of 9% of battery instead of the previous night's 46%, in what I understand identical conditions otherwise.

seriously, wake up to check if there is anything to sync!? I have only 2 email-accounts (without auto-sync) and the jolla account active and it is pretty dumb to be hit by this anyways! The "is there something to sync"-check should be deactivated if there are no accounts to sync... some like "I am going to sleep now, is there anything to wake up for?" would be much better.

Just to keep on updating, I've not seen battery problems since my Mar 21 update, even though I've nowadays re-enabled Google and Facebook sync options, leaving only Twitter disabled. So it might be either Twitter only or then some more random problem. Regardless, I don't need the Twitter sync so I'll keep that disabled.

Fix still needed and fast. Currently cannot get it through the day. 10% battery drain in less than one hour with little use. Only account configured is Jolla. T

@karlefr It is definitely fantastic result! Unbelievable for me to be honest. I can not get 15% per 24hrs even with device idle Wi-Fi or 2G. My usual idle drain rate over night is 1% per hour. That is without even touching the phone.

I use the aeroplane mode at nights to save battery, but my Jolla still drains battery just like it would when there were connections :(

how do you do such regular measurements? do you have a scheduler doing this for you?

Miska has made an app to monitor the battery: Hunger Meter

Yes, Hunger Meter. Currently no UI yet to display all information it collects (in my TODO), but you can get data from commandline via sqlite3 /tmp/harbour-hungermeter/measurements.sqlite .dump where energy column is mWh left in battery

I don't use mail sync at all and I still feel the battery drain is faster than before update.

This being said, the effect of the syncing should not be as drastic as has been reported in the first post, so if you find some other things that seem to affect this, please keep on reporting.

We're gonna need a new battery drain test apparently :)

From my own non-scientific test, battery drain is as usual (about 2% per hour) with mail sync + WLAN enabled, but a lot more (4-5% per hour) with 3G (sync every 30 min).

I don't have a Mailbox set up on my jolla.

@PhilippeDeSwert It seems to be few random problems, on which I try to make sense out of chaos (never mind 'style' or my english). As longtimer in SW and Linux, I see it pretty hard to validate or pinpoint without help of logging/tracing tools. I have collected more of my findings as diary in answers below my original question, and continue as long I get some indication that putting my (also valuable) time on this is of any help to Jolla grew and other users and help to get rid of these bugs and clitches.

So far as Black Box approach my todays hunch is twofold.

1. Having both connection methods enabled (WLAN and 3G) in settings and causing a situation where Jolla is losing connection to WLAN and switches to 3G automatically (possibly user double tapped at the same time->App pings server or some other network activity occurs). This has happened to me now several times with mitäkuuluu (see my latest 'answers' in my original question you linked ). Again I consider mitakuuluu merely as victim not the cause of problem.

2. Some (propably other than 'mitakkuuluu') process/system service gets in runaway/race condition or whatever bad state, eating resources memory/irq/d-bus younameit, causing pixelations led blinkings and majority of srange random problems reported here during last week (as seen in my list).

   I have not enough tools or guidance how to track those more deeply, that's why I urge Jolla to step in and try to reproduce problems with the preconditions I try to point out. Or give me info how I can enable logging so that I could for example see system messages after the clitch occurs next time etc.. Before 1.4.20 I newer ever had any clitches in camera, gallery and switching between them. Neither had I any pixelation problems before. Led blinking (red I think) I have seen before only in November (unboxing->had to factory reset) before 1.4.20.

I give _zero_ value on any new features before these are ironed out for good. IMO don't even consider to bring in 4G before things like handover WLAN->3G data works rock solid also in poor conditions. I have enough features already, I want reliability, so that I can trust Jolla without worrying, what contacts I might have missed, because phone happened to be in pocket in poor field.

I love my Jolla anyway, so keep up the good work, but keep cool on priorities.

@targon : Thanks for your feedback. I will give you some more answers in your thread for logging etc.

But it seems here that mitäkuuluu is potentially causing a number of your issues. If somehow connectivity gets bad, it is its own fault if it goes beserk. So imho it is not a victim but part of your problem. Also system logging might not show you anything when some app goes crazy as it is not part of the system. I have raised the point though that we need some app/system/tool to find and denounce those run-away apps.

@PhilippeDeSwert Thanks for your comment, I agree about mitäkuuluu, just didn't want to point to one app direction before other things are ruled out. What I am now testing is start all other apps, except mitäkuuluu after latest crash i reported in 'my' thread some 15 minutes ago. What i'm still wondering is connectivity logic typically what'sMyIp app is showing only WLAN IP but now after latest crash i reported it shows both IP:s (=also 3G ip)

Mitakuuluu has been know to sometimes run amok and use 90% cpu constantly.

Anyway can anybody check with top if some process is running amok when the power drain is noticed?

Mitakuuluu is not the problem here. This battery drain is there even if MK is not installed.

F.ex I left my Jolla to idle on my work desk today with Wifi on, 2 exchange accounts, 1 fb account, twitter, jolla account.

After 5 hours of pretty much idle state (some little usage now and then) I was down from 100% to 58%. And that is not normal in anyway.

@inffy: Which is why it would be good to identify other potential run-away processes. As it is indeed not normal.

Just run top in a terminal, wait a few seconds and see what stuff keeps popping up on top of the list with much cpu usage.

@Philippe De Swert yep, i will do that today evening or tomorrow when I have some more spare time to tinker with the device :)

whoever has the spare-time earlier - i'd like to know the cpu load of the background-mail check at slow internet connection ... maybe setting the device to 2g could help here... cheers

If you are not using persistent storage (or you don't have anything there yet) and you just started the application, estimated time left is how long your battery will last given average (in this case current) power consumption. Jolla uses systemd and that at least on desktop will run GUI as fast as it can and then silently finishes the boot in the background. Right after boot I usually have pretty busy Jolla eating up something between 2-3W. It usually settles after a little while... Would suggest check the consumption part as well.

@miska, thx for answering. I'm aware of the background boot process eating up some W, that's why I generally wait about a minute or more before disconnecting the charger. And the method has worked well for my 2 previous test charges. But now the indicator just fluctuates, so I can't trust the displayed output.

@miska, thank you for pointing out I should've monitored the actual consumption, now I see the percentage calculations were not what I thought. And although I'm still not sure about the occasional "power surge for seemingly no reason", the battery consumption hasn't really been that bad.