LHC@home

since last night, all my machines which crunch ATLAS show tons of download errors:
-186 (0xFFFFFF46) ERR_RESULT_DOWNLOAD

the computers which crunch THEORY are not affected.

does anyone else make the same experience?

... does anyone else make the same experience?

All your wingmen do.
Just look at col 2 of your tasklist.
It appears to be a faulty batch (that needs to run until all tasks are automatically sorted out).

... does anyone else make the same experience?

All your wingmen do.
Just look at col 2 of your tasklist.

oh, yes, that's right; I should have looked up col 2 to begin with :-)

Seeing also HTTP-Error, but also atm 200 KBps instead of 7.000 for download!
15 min. 50% of download of 242,86 MByte.

There must be two Atlas-Download-Server.
One PC(Win11pro): First download with 7.000 KB/s and a few minutes later a Second with 45 KB/s atm.

There must be two Atlas-Download-Server.

3 (at least)

dig +short lhcathome-upload.cern.ch.
188.185.89.232
188.185.86.60
188.185.125.226

Sorry,
Download, not upload. Load-Balance Problem?

Seeing Downloadspeed between 50 and 100 KBps for Atlas since three days now. (ok it's weekend).

The servers behind the FQDN "lhcathome-upload.cern.ch" handle (ATLAS) uploads as well as (ATLAS) downloads.
It looks like they worked fine during the whole last week including this weekend.

FYO, or for all of us?
63k Cobblestones today, for this year the highest ever.
Seeing this slow download not for every download, but see it!

381 MByte Atlas-Tasks have this small Data-Transfer between 25 and 50 KBps.

Two downloads for Atlas-Tasks at the same moment on one PC,
the first have a full speed, the second for example 100 KBps.
When the first is finished, the second don't make a correction of the speed,
so this second have a longer downloadtime.

Do they come from same download server? If they come from different ones, that may explain different speed. The other server may have some connection issues that the LHC team should take a look at.

---

For this, a deeper look into networking is needed.
Don't know, if they come from different DownloadServer.
Had made a bandwidth-check. max. Speed is possible on this PC.
Checked during download with Boincmanager.
Have found this message from 2020:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=5533&postid=43523#43523
Edit: atm one Atlas-Download only, with 140 KBps on this Win11pro:
https://lhcathome.cern.ch/lhcathome/show_host_detail.php?hostid=10631979
WAN miniport Installation from OS was in Devicemanager.
Will this checking on all Win11pro.

Have WAN miniport SSTP deactivated.
Download with less than 150 KBps is always active, but
only for a Atlas-Task, which is not running (only two are limited for running),
so, no problem to have a download of more than one hour!
No Solution found with Google or Microsoft-Support.

Have Windows PC's changed from CMS to Atlas, seeing there also low Downloadspeed (70KBps).

63k Cobblestones today, for this year the highest ever.

Slow download is not going away, but this heavy Atlas-Tasks get now 88k Cobblestones today.
Would be nice to see this 1k Collisions (Longrunner) once again in the Future.

Atlas-native in VM (CentOS7 and CentOS8) and Atlas-Tasks under Win10pro and Win11pro,
have per day between 55k and 97k Cobblestones,
but, the slow download is seen on native and Windows! Not every download, but....

CentOS8-VM 50KBps 46 min for 120 MByte from 240 MByte downloaded so long!
2. Win11pro 50 KBps for the next 1,5 hours!

I can't say that I monitor the downloads very often, but I am running 22 cores of a Ryzen 3900X on native ATLAS and CMS, and have not noticed any problems yet.
I never see any stuck in transfers when I look.
https://lhcathome.cern.ch/lhcathome/results.php?hostid=10750296&offset=0&show_names=0&state=4&appid=

I do have a good cable modem with 110 Mbps download though.