qBT compiled with libtorrent 0.16.16

[ciaobaby]

Since recompiling qBT master build 3.2.0a with libtorrent 16.16 on Linux Mint 16 I have noticed that all activity will cease for no apparent reason.

This not the odd task or three slowing down, but ALL running tasks (downloading and seeding) will simply come to a stop pretty much all at the same time.

Nothing is logged,
I have no queueing enabled,
no peer blocking of any kind,
no throttling (because the stuff I run is not likely to max out my connection),
no software firewall,
router SPI is off,
uTP is on,
nothing untoward appears in the "blocked IP" log,
'Half open' connection set at 10.


Taking qBT down (File -> Exit)  and restarting a few minutes later or "Pause All" and "Resume All" will bring  everything is back to normal. There is no particular time scale that I can discern.

DHT node count stays around the same. I am loathe to say it is definitely a libtorrent issue but that is where the symptoms seem to point. So before I load up libtorrent 1.0 to test with that, are there any issues or 'gotchas'  that anyone knows of?

[sledgehammer_999]

Nope, apart for other users complaining for stalled torrents I don't know the real reason. You could build 0.16.17 (or even RC_0_16 from svn) and see what happens.
Apart from building libtorrent did you also package+install it? If not are you sure that qbt picks up your new lib and not mint's one?

[ciaobaby]

Yep, output from dpkg-query -l libtorr*

ii  libtorrent-rasterbar-dbg        0.16.16-0ubuntu1~sau amd64                Debug symbols for libtorrent-rasterbar
ii  libtorrent-rasterbar-dev        0.16.16-0ubuntu1~sau amd64                Development files for libtorrent-rasterbar
ii  libtorrent-rasterbar-doc        0.16.16-0ubuntu1~sau all                  Documentation for libtorrent-rasterbar
ii  libtorrent-rasterbar7          0.16.16-0ubuntu1~sau amd64                C++ bittorrent library by Rasterbar Software

[ciaobaby]

Bit more info;

I set up "Alternative speed limits" (several different setting over a couple of days) and when the tasks stop, switching to alternative limits for a few minutes then switching back can trigger all the tasks into restarting activity.
It appears to matter not what the alternative limits are, just that they are 'different'.
Still nothing logged.