<core_client_version>7.4.53</core_client_version>
<![CDATA[
<message>
finish file present too long
</message>
<stderr_txt>
command: ../../projects/boinc.bakerlab.org_rosetta/rosetta_4.15_arm-android-linux-gnu -abinitio::fastrelax 1 -ex2aro 1 -frag3 00001.200.3mers.index -in:file:native 00001.pdb -silent_gz 1 -frag9 00001.200.9mers.index -out:file:silent default.out -ex1 1 -abinitio::rsd_wt_loop 0.5 -relax::default_repeats 5 -abinitio::use_filters false -abinitio::increase_cycles 10 -abinitio::rsd_wt_helix 0.5 -beta 1 -abinitio::rg_reweight 0.5 -in:file:boinc_wu_zip pass_build3gbaab+1.bp_20200416211333_-3_0001_0001_fragments_data.zip -out:file:silent default.out -silent_gz -mute all -nstruct 10000 -cpu_run_time 28800 -boinc:max_nstruct 5000 -checkpoint_interval 120 -database minirosetta_database -in::file::zip minirosetta_database.zip -boinc::watchdog -boinc::cpu_run_timeout 36000 -run::rng mt19937 -constant_seed -jran 2159818
======================================================
DONE :: 1 starting structures 3851.34 cpu seconds
This process generated 1 decoys from 1 attempts
======================================================
BOINC :: WS_max 0
called boinc_finish(0)

</stderr_txt>
]]>

What does it mean? How to avoid if possible?

Heavy I/O at the time of a Task finishing. The after-the-task-ends-cleanup doesn't happen fast enough for the Manager's liking, so it gets clobbered. Even if it was a valid result, it ends up as an error.
From a BOINC thread back in 2015 (and it references a report of the error in May 2014).

In looking through the client code it looks like this condition occurs when the client finds that the boinc finish file has been written to disk but the science application process is still running.

The fix is meant to have been (finally) included with the latest BOINC version 7.16.5

---

Grant
Darwin NT

Heavy I/O at the time of a Task finishing. The after-the-task-ends-cleanup doesn't happen fast enough for the Manager's liking, so it gets clobbered. Even if it was a valid result, it ends up as an error.
From a BOINC thread back in 2015 (and it references a report of the error in May 2014).

In looking through the client code it looks like this condition occurs when the client finds that the boinc finish file has been written to disk but the science application process is still running.

The fix is meant to have been (finally) included with the latest BOINC version 7.16.5

Thanks. But 7.16.5 caused crashes on similar phone... looks like stalemate... but I'l try to upgrade.

Some errors on my Android (Galaxy S8) phone - running 4 cores out of 8
All ended after just 1 decoy completed

These three in the form
jgHE_b02_COVID-19_SAVE_ALL_OUT_IGNORE_THE_REST_6du1pq5e_924710_1_0
Junior_HalfRoid_design5_COVID-19_SAVE_ALL_OUT_IGNORE_THE_REST_1ld2iu3x_924647_2_0
Junior_HalfRoid_design5_COVID-19_SAVE_ALL_OUT_IGNORE_THE_REST_1qu3aj2y_924612_2_0

Too many restarts with no progress. Keep application in memory while preempted.
======================================================
DONE :: 1 starting structures 15519.2 cpu seconds
This process generated 1 decoys from 1 attempts
======================================================
BOINC :: WS_max 0
called boinc_finish(0)

</stderr_txt>
<message>
upload failure: <file_xfer_error>
<file_name>blah-di-blah</file_name>
<error_code>-161 (not found)</error_code>
</file_xfer_error>

This one containing an extra few lines
Junior_HalfRoid_design5_COVID-19_SAVE_ALL_OUT_IGNORE_THE_REST_7gy0qz5g_924612_3_0

BOINC client no longer exists - exiting
timer handler: client dead, exiting
Too many restarts with no progress. Keep application in memory while preempted.
======================================================
DONE :: 1 starting structures 24965.3 cpu seconds
This process generated 1 decoys from 1 attempts
======================================================
BOINC :: WS_max 0
called boinc_finish(0)

</stderr_txt>
<message>
upload failure: <file_xfer_error>
<file_name>Junior_HalfRoid_design5_COVID-19_SAVE_ALL_OUT_IGNORE_THE_REST_7gy0qz5g_924612_3_0_r1564564761_0</file_name>
<error_code>-161 (not found)</error_code>
</file_xfer_error>

How do I keep application in memory in the Boinc app?

---

How do I keep application in memory in the Boinc app?

That's a big problem on Android.
As I said in other thread, there is no such option in BOINC client for Android.
Moreover, the nature of Android memory management makes it much worse. Android feels free to kill background app if it decides to do so. Not just swap it out, kill the process.
Hence, BOINc client itself (not only project's science app) got killed time to time.
Maybe second of your logs reflects case when BOINc client was killed but science project app still running - it founds no BOINc client to communicate with and reports exit.

So there are 2 ways to handle this issue - for BOINC devs to learn if any additional mechanisms exist in Android to prevent background process be killed and enable them.
And for users:
1) to disable any agressive handling of background apps. For example, my Xiaomi Redmi 8 has "Activity control" option. In "hard restriction" it just kills background process.
Default is "clever" mode that leaves this on OS discretion. I set "no restrictions" for BOINC to reduce such kills as possible.
2) to avoid running anything on smartphone while keeping BOINC as foreground process. Of course this would be acceptable only at night or when phone not in use at all.

How do I keep application in memory in the Boinc app?

That's a big problem on Android.
As I said in other thread, there is no such option in BOINC client for Android.

Thanks - I can stop looking for it now. I search once every 2 or 3 weeks to no avail.

Everything else you say makes sense. No option that I know on this phone, unless someone knows better.
I'm not great with phones tbh

---

Hello,
Try looking at battery optimizing. That can also yank Boinc, too.
I'll be giving this a try on my new phone. On a flagship (or reasonably flagship) snapdragon, how many cores would one recommend using? Is 4 a good number?

thanks!

Some new phones have assymmetric cores - 4 more powerful ones and 4 less, for example (as in Redmi Note 8).
If phone restricted by available RAM, makes sence to use only powerful cores and leave less powerful for maintenance tasks.
Don't know how many and what size of cache levels in modern ARM architectures (if they have shared L3/2 cache for example, reducing number of cores could improve memory access patterns) but definitely one need all active BOINC tasks be in RAM, not in SWAP, for better performance and less phone degradation.

It was said that ARM developers have come in to support the project.
This must be a comparatively trivial thing for them to resolve.
Maybe the problem is that it's too trivial to give any attention to.

In the early 90s, when 16-bit computers were the norm, software was 10x more usable and practical than anything that appears in phone apps, except in terms of graphics.
Quite simply, phone apps don't work. Never have. At the rate things are going, probably never will.

---

It was said that ARM developers have come in to support the project.
This must be a comparatively trivial thing for them to resolve.
Maybe the problem is that it's too trivial to give any attention to.

In the early 90s, when 16-bit computers were the norm, software was 10x more usable and practical than anything that appears in phone apps, except in terms of graphics.
Quite simply, phone apps don't work. Never have. At the rate things are going, probably never will.

A really interesting blog post I always like to share when this topic comes up. Some really good points made.
https://tonsky.me/blog/disenchantment/

It was said that ARM developers have come in to support the project.
This must be a comparatively trivial thing for them to resolve.
Maybe the problem is that it's too trivial to give any attention to.

In the early 90s, when 16-bit computers were the norm, software was 10x more usable and practical than anything that appears in phone apps, except in terms of graphics.
Quite simply, phone apps don't work. Never have. At the rate things are going, probably never will.

A really interesting blog post I always like to share when this topic comes up. Some really good points made.
https://tonsky.me/blog/disenchantment/

Quite right too.
I used to use a graphical editor for Fidonet BBS systems called LED on the Atari platform. Best editor I ever used.
Word-wrap within nested quotations within each level of the nest. Outlook can't handle that now, let alone 25yrs ago.
As for forums like this one, I have a word for them - I'd be banned outright for writing the word here.

---

phone manufacturers are more worried about being caught in a bad light for having overheating phones. at one point phones run so hot that it kind of scald the users using it. that soc i think is A72 or a further generation from there. and battery life is the other consideration and Android OS aggressively stop apps and background tasks to curb them running to save on battery juice.

today A72 is running on PI4, it is the appropriate cpu for the appropriate purpose, A72 quad core superscalar runs like a decent performance laptop cpu.
push about 3-5 watts of power i.e. 1 Amp x 5 volts from the usb C. it runs r@h and rival desktop cpus in terms of power efficiency. it takes much longer to produce the same results but that it uses less power than the power hungry desktop cpus

architecture like Big-small cores is of not much use in a context like r@h, when you want to run to run the threads you want to run them fast and produce the most work for the least power. Big-small is sold on the premise that while you are not doing anything serious, shutdown the big cores and run on small cores and hence saving power. but at the same time throttling down on performance.

Well I'm not getting very far on any of my devices. Nothing in the event log to really tell me what's going on. Nothing is being crunched.

I know this is possible - see: Dreamlab. Not sure why Boinc hasn't done something similar yet.
I'll leave it plugged in over night and see how far I get.
Was told to downgrade from 7.16.5 to 7.16.3. It can be found here. https://boinc.berkeley.edu/dl/boinc_7.16.3.apk

Doesn't seem to make much difference. Devices isn't even warm to the touch with 4 cores apparently working. Plenty of ram 8- 8 GB in one and 12 in the other.

If nothing happens by tomorrow I guess I'll either uninstall and try again or remove entirely. 2 GB sitting doing nothing is pretty depressing when all those cores could be doing something useful.