VP22 Upgrade from 24.04-1.0 to 24.04-1.1 failed
-
maybe (I did not check the code on that) but you installed a new version successfully right ? Even if you checked 'don't erase user data', if a new version can be not compatible with the cached data, the cached data should go (after all, it should not hold any interesting data if it conforms to its name)
-
Where should I report the bug?
-
Logic would say the installer on Github. However the issues are not actively followed, it's obvious that there is not a lot of people monitoring or even janitoring the list. So maybe give a first try to the port for your device, after all we don't know if there are special handling per device for this.
-
Got hit by the failed OTA on my VP22 (stable channel). How long does the ubports installer usually take to flash from recovery? 3 attempts with the appimage have resulted in 15 minutes+ of the installer stuck at "Activating slot a" with the phone showing the fastboot menu along with the incredibly unhelpful "Text" as the output log.
-
3600 seconds limit for edits? Really?!?
Anyway, another attempt on a Mint Debian Edition computer instead of my usual minimal Devuan install gets further along. Hopefully it's ready within an hour so I can update my post.
EDIT: flashing worked fine, also VoLTE is active now, which is nice.
-
Since last upgrade I seem to experience occasionally system crashes, or rather freezes: the system does not respond and I have to reboot. I did not have such problems before the upgrade.
It seems to happen more when recording a video, when terminating a call, but It also happened in random moments like, just after starting Signal UT.
Anyone else noticed that?
-
While I don't have a VP22, I also did the same upgrade on my FP5, I have triggered a similar looking crash by doing more or less cheesy things with the dialer (I did not manage to repro it unfortunately).
I could connect to the phone via ssh and to launch top, all the signs of an oom were presents (no more free swap, swap process taking all the cpu time). I did not have the time to do any tests, at this point a Linux system is all but finished.
Another symptom was that the phone was unnaturally hot. -
I've tried to access the shell when the phone freezed, but It does not reply.
Honestly the frequency of freezes is quite alarming, it almost never happened before 24.04.1.1 .
-
maybe this problem could be related ?
-
I almost do not use Waydroid, the crashes and freezes happen when waydroid is not started and have not been started since boot.
Well I've investigated and it seems to be an out of memory problem.
Because I realize that now if I start a few apps (like the Morph browser + whatsweb + Signal UT ) then I rapidly run out of Ram AND Swap. (I can see it in top)
Then my question is why did it not happen before? Is the memory usage is higher in 24.04-1.1 compared to 24.04-1.0 and if so why ?
And secondly why do we have so few swap (2 Go) if there are Out of Memory problems? Or maybe we had more swap in 24.04-1.0 and it was reduced ?
Is it possible to increase the swap?
Edit: By the way I've reported the problem here. https://gitlab.com/ubports/porting/reference-device-ports/android11/volla-phone-22/volla-mimameid/-/issues/24
-
Ps: I also just experimented a black screen when receiving a call. In Adb I could see the phone was not in out of memory at that time. The phone was vibrating and I could not turn on the screen to see the call.
Although previously in this boot I had an Out of Memory problem, it freezed temporarilly, but the system recovered, so I wonder if some important system processes may have been killed in the process.
Honestly I'm wondering whether I should downgrade, because this starts to be a real problem.
-
@pparent said in VP22 Upgrade from 24.04-1.0 to 24.04-1.1 failed:
I can see it in top
When using top, did you have enough time to notice if one particular application's Ram use was growing without limit ?
To elaborate, here is the grand total of my experience with OOM.
First generally with Linux (desktop and servers): it has always happened in the context of a 'rogue' application, never because I loaded too much the system. If I was loading too many apps on a given Linux system, I could always unload one of them and get back to a more normal behaviour; when a rogue application was running, I could see this application eating more and more Ram and if I could not kill it fast enough (before I lost control of the console to use kill -9) the system was all but dead, only a hard reset could get it back. Needless to say (it's a pretty well known fact of using Linux) the system OOM handler was never of any help.
Then with UT
First I have found a 'rogue' app that can reliably hang my phone. This was the first experience of it, and I could see the very same behaviour using top (it is hanging progressively, it takes a few minutes to eat all my 4 GB of swap)
Then with the dialer: this has happened very recently, while I was trying to experiment with some advanced particularities of it and suddenly the system became sluggish and the UI unusable, connecting with ssh I could see with top that lomiri-dialer-app was growing more and more and the swap was eaten. The ssh session got soon unresponsive (the phone UI itself was already gone so it was another hard reboot. I am sure that at this point the dialer was the only application active, so it's the dialer or more probably one of its libraries that is the culprit. As the dialer use quite a lot of libraries, it's not helping much, but it's possible that one of the apps you were using has triggered the same bad code path. Unfortunately I could not repro it with the dialer.
I don't know if you have noticed it, but in the system there is something of a special OOM handling, application can submit a specific OOM profile when starting. Unfortunately I have not seen it any more useful than the pretty useless kernel default handling, the rogue applications can do their evil deeds without being bothered by it.
A last detail: on my FP5 (I'd suprised if your Volla was different because it's using half of your Ram like for my FP5, it's a giveaway), the swap handling is not the 'classical' sort, it's the more advanced type of 'compressed Ram'. I don't think it's a novelty of 24.04.1.1.
-
About the dialer, I cannot exclude that the app was opened (or had been used) at the time of every crash.
It's true also that I now use VoLTE, that I did not use before 24.04.1 , so it could have a impact on the RAM(?)
It's also true that I cannot reproduce the bug by opening a lot of apps straight after a reboot.
