I've been having upload/download problems for the last couple of days.
While work has been getting through in both directions it seems to be affecting throughput by about 20%, based on credit scores.
It happens on multiple machines with different Internet connections, so I do believe it is on the server side.
I'm just wondering if others are seeing the same thing and if we know why.
The event logs contain entries like:
Sat 01 Oct 2011 12:43:34 AM PDT | | Project communication failed: attempting access to reference site Sat 01 Oct 2011 12:43:34 AM PDT | Einstein@Home | Temporarily failed download of p2030.20100513.G48.07-00.16.C.b3s0g0.00000_1840.binary: HTTP error Sat 01 Oct 2011 12:43:34 AM PDT | Einstein@Home | Backing off 2 hr 12 min 11 sec on download of p2030.20100513.G48.07-00.16.C.b3s0g0.00000_1840.binary Sat 01 Oct 2011 12:43:34 AM PDT | Einstein@Home | Started download of p2030.20100513.G49.08-00.78.C.b0s0g0.00000_121.binary Sat 01 Oct 2011 12:43:35 AM PDT | | Internet access OK - project servers may be temporarily down. Sat 01 Oct 2011 12:44:20 AM PDT | Einstein@Home | Finished download of p2030.20100513.G49.08-00.78.C.b0s0g0.00000_120.binary
Joe
Copyright © 2024 Einstein@Home. All rights reserved.
Internet access OK - project servers may be temporarily down.
)
I'm having the same effect. Oldest "stuck" download started 32 hours ago. Five work units in download state, no Einstein jobs running for hours now.
Same here. Detached and
)
Same here. Detached and rejoined. Exe and GPU files downloaded, but as soon as it started to download actual workfiles........ Nothing.
This also happened a couple of days ago, but seemed to right itself. Then again, maybe not.
Following Furlozza, I
)
Following Furlozza, I detached and reattached. Work units are now completing downloads, but excruciatingly slowly - mostly under 10 kbps - and with some failures and retries - "connect() failed" or "HTTP error". So all seems not wholly well at the server end. However once more running wus.
I realise this is probably the same problem as reported in an adjacent thread.
I notice on the server status
)
I notice on the server status page http://einstein.phys.uwm.edu/server_status.html that the Glasgow mirror is going up and down a lot.
Is there a way to determine which mirror one is using for downloads or can we specify one?
RE: I notice on the server
)
I believe changing your time zone location, or time will change which download server you'll use,
You can always download the files manually from one of the other URL's in your client_state.xml (With Network Use Suspended), drop the files into your project folder, then re-enabling Network Use, Boinc will find the files and skip downloading them.
Claggy