Unfortunately, the log doesn't show you cancelling the download - which is probably the log's fault, not yours.
..............

Will do. But restarting the machine (or presumably just the Boinc client) always clears the problem. My problem is it doesn't clear without restarting.

Same problem here.
Some downloads keeps stalling (= stop flow of work for R@H as BOINC stops getting new work from R@H and switch to backup project - WCG in my case) every few fays (~ one file stuck of few hundreds downloads clearly ).
Some of the cases, retry transfer does not help (is probable a R@H server bug), aborting stuck transfer usually work, but it some cases it does not help either so i need to restart BOINC completely to resume normal work flow.

There are also no any entries in BOINC event log about aborting transfer:

19-Feb-2020 13:32:36 [Einstein@Home] Requesting new tasks for AMD/ATI GPU
19-Feb-2020 13:32:43 [Einstein@Home] Scheduler request completed: got 1 new tasks
-------------here i have aborted stalled download, but it is not logged -----------
19-Feb-2020 13:33:24 [Rosetta@home] update requested by user
19-Feb-2020 13:33:27 [Rosetta@home] Sending scheduler request: Requested by user.
19-Feb-2020 13:33:27 [Rosetta@home] Reporting 1 completed tasks    - note: this task has stalled download and BOINC recognize it as aborted after stale download canceled, but refuse ask for more work   
19-Feb-2020 13:33:27 [Rosetta@home] Not requesting tasks: some download is stalled
19-Feb-2020 13:33:29 [Rosetta@home] Scheduler request completed
-------------------------------- i restart BOINC and getting new work resume  -------------------- 
19-Feb-2020 13:33:51 [---] Exiting
19-Feb-2020 13:34:44 [---] Starting BOINC client version 7.14.2 for windows_x86_64
19-Feb-2020 13:34:44 [---] log flags: file_xfer, sched_ops, task
19-Feb-2020 13:34:44 [---] Libraries: libcurl/7.47.1 OpenSSL/1.0.2g zlib/1.2.8
-------------- cut --------------------
19-Feb-2020 13:35:06 [Rosetta@home] Sending scheduler request: To fetch work.
19-Feb-2020 13:35:06 [Rosetta@home] Requesting new tasks for CPU
19-Feb-2020 13:35:09 [Rosetta@home] Scheduler request completed: got 7 new tasks
19-Feb-2020 13:35:11 [Rosetta@home] Started download of flags_rb_02_12_15913_15773__t000__2_C1_robetta
19-Feb-2020 13:35:11 [Rosetta@home] Started download of input_rb_02_12_15913_15773__t000__2_C1_robetta.zip
19-Feb-2020 13:35:11 [Rosetta@home] Started download of flags_rb_02_13_14003_15805__t000__2_C1_robetta
19-Feb-2020 13:35:13 [Rosetta@home] Finished download of flags_rb_02_12_15913_15773__t000__2_C1_robetta
19-Feb-2020 13:35:13 [Rosetta@home] Finished download of flags_rb_02_13_14003_15805__t000__2_C1_robetta

We won't be able to solve this unless somebody catches one in the act and can display some diagnostic test results. Then we'll at least know which bit of the system is broken - though even that doesn't guarantee we can fix it.

I can dig some data next time it happen.
I already added <http_debug>1</http_debug> in config.
But it will just show why downloads from R@H server is stalling, but it will not show why BOINC refuse to get new fork even after such stale download is already aborted. isn't it?

And i can not catch it. It does not repeats each day and looks like running BOINC with <http_debug>1</http_debug> cause problems to BOINC itself.
After some time it loses international connection (between BOINC and BOINC Manager - red dot on tray icon) and hangs.
Looks like it caused by too many messages in event log due to very high verbose level of <http_debug> and very high BOINC network activity on machine.
I checked stdoutdae.txt and it generate ~ 50 000 log lines per day.

Mat be will try later on less active machine.

Looks like it caused by too many messages in event log due to very high verbose level of <http_debug> and very high BOINC network activity on machine.

It's not the total amount of lines that causes this but generating over 1,000 lines per second as the polling between the client and the manager happens once every second, which is 1,000 milliseconds, which is those 1,000 lines. Don't set multiple debug flags at the same time (like http_debug and work_fetch_debug for instance) as then you run into this problem.

---

Good to know. Thanks.
Although i highly doubt i actually ever hit 1000 lines per second.
I have run only one debug option for this time (http_debug only).
Machine is quite active (16 thread CPU + 2 GPU each running 2 tasks so 20 tasks is running in parallel) but i skim trough stdoutdae.txt and looks like 200-300 lines per second was max logging speed.

May be same can happen if something delay polling between the client and the manager?
So not > 1000 lines per second but > 1000 lines since last polling?

P.S.
Have found a workaround - looks like BOINC could apply http_debug flag without restart (via "read config files" after cc_config edit). So i will try this method next time if i see more stuck downloads from Rosetta @ Home.

Another way of cutting down the log and making it more readable is to set

<dont_contact_ref_site>1</dont_contact_ref_site>

in cc_config.xml - that removes the debug clutter of the check-call to google.com after a failed BOINC connection.

journalctl --boot=-1 --unit=boinc-client
works for me

---

Corona virus is close to being declared a pandemic.
But don't worry about it. it's only fatal to some people.

Sorry for the off-topic post, but there's no other place to answer this.

But don't worry about it. it's only fatal to some people.

The flu is deadlier. If at least the death numbers are given correctly and governments aren't putting out fake numbers to make it look less problematic.

So far, the new coronavirus, dubbed COVID-19, has led to more than 75,000 illnesses and 2,000 deaths, primarily in mainland China. But that's nothing compared with the flu, also called influenza. In the U.S. alone, the flu has already caused an estimated 26 million illnesses, 250,000 hospitalizations and 14,000 deaths this season, according to the Centers for Disease Control and Prevention (CDC).

...

Typical flu symptoms include fever, cough, sore throat, muscle aches, headaches, runny or stuffy nose, fatigue and, sometimes, vomiting and diarrhea, according to the CDC. Flu symptoms often come on suddenly. Most people who get the flu will recover in less than two weeks. But in some people, the flu causes complications, including pneumonia. So far this flu season, about 1% of people in the United States have developed symptoms severe enough to be hospitalized, which is similar to the rate last season, according to data from the CDC.

With COVID-19, doctors are still trying to understand the full picture of disease symptoms and severity. In a small study of about 100 people with the virus, published Jan. 30 in the journal The Lancet, the most common symptoms were fever, cough and shortness of breath. Only about 5% of patients in that study reported sore throat and runny nose, and only 1-2% reported diarrhea, nausea and vomiting.

https://www.livescience.com/new-coronavirus-compare-with-flu.html

How long before some smart *** names a computer virus corona or covid19?

And virus writers aren't what I would categorise as smart. I know the guy who wrote the raid virus. Sure, he can program, but he has no common sense, no sense of decency, etc.

I don't know the etiquette. Are they named by the people who write them or those who find them in the wild?

You can say ass. That's a donkey. Just don't say arse :-)

But you can only say it once in this thread. Both.
Our rules give more leeway than those at project forums, even though the ones on the left are the same all over. I'm not enforcing them so strictly, unless you're a spammer. I can't speak for my moderators though, leave it up to them and their judgment to decide what to do.
Like someone guiding us back to the topic at hand, but maybe that the topic changed. :)

---