Abort the transfers on those that don't D/L. Means essentially manually getting work. Slow but it works.

Here is another method of handling this problem. I have selected 'no new tasks' under 'projects' tab. This will prevent getting any new work units. In the meantime, my queue of currently available workunits were successfully completed and the results manually returned to the project server. Right now, I do not have any workunits in the pipeline for this project to crunch. I am waiting for the people who run this project to fix it. When this is fixed (hopefully soon) I will select 'allow new tasks' to get all downloaded workunits without any download problems. If this problem takes longer to resolve, then I will have to manually get work.

The best is to have a spare project to run if the WUs stuck while your at work or so. My queue runs out of work in less than 2 hours and I can't run home from work to manually get more work, so I run Cosmology at home simultaneously on my second core and it also takes over this core if I run out of work on M@H.


Hm, but I wonder what happened with Travis, he promised to be back new years eve, but we still haven't heard a word...

He could at least let us know that he have acknowledged the problem so we know they are working on it ...

---

snip...

Hm, but I wonder what happened with Travis, he promised to be back new years eve, but we still haven't heard a word...

Promise ? What promise? I never heard any promises made.....whats a few days among friends during the holidays in an alpha project? How many projects(all of them) give dates they don't achieve...LHC has made me patient :)

---

snip...

Hm, but I wonder what happened with Travis, he promised to be back new years eve, but we still haven't heard a word...

Promise ? What promise? I never heard any promises made.....whats a few days among friends during the holidays in an alpha project? How many projects(all of them) give dates they don't achieve...LHC has made me patient :)

HAve you checked the front page?

December 26, 2007 Out of Town
I'm going to be out of town until new years eve, so if i don't get back to any questions on the forum or email that's why. I've set up the assimilator and validator to run as daemons in the config.xml now, so if the machine goes down and comes back up, these should start back automatically even if i'm still out of town -- so hopefully things will run smoothly while I'm gone. I hope everyone is having a happy holidays!
--Travis

:P

---

The best is to have a spare project to run if the WUs stuck while your at work or so.

I don’t think that helps in this case: the stuck WUs seem to block everything else. My G5 (which is attached to several currently active projects) was idling and ‘cold’ the other night—at about 55°C, an unfamiliar sight!—with several MW@h tasks stuck at 30%; it wouldn’t ask for new work from anywhere until I aborted those transfers. However, a newer version of BOINC than mine (v5.4.9) might not have the same problem.

---

Hi,

....it wouldn’t ask for new work from anywhere until I aborted those transfers. However, a newer version of BOINC than mine (v5.4.9) might not have the same problem.

I'm running BOINC v5.10.13 (64-bit build) and can confirm it doesn't exhibit this problem; it'll carry on requesting work from and crunching other projects just fine when "Milkyway" transfers hang.


TTFN - Pete.

---

I'm using 5.10.13, and I've had a lot of them stick.

I'm using 6.1.0 on some machines, 5.10.30 on others and Boinc will switch to switch to another project when the Milky Way app hangs.

---

I'm using 6.1.0 on some machines, 5.10.30 on others and Boinc will switch to switch to another project when the Milky Way app hangs.

Just to clarify: I haven’t seen the app hanging, only the stuck downloads. My BOINC v5.4.9 doesn’t hang, either: apparently it just stops looking for work when it thinks there are downloads in progress.

The theory that they’re really HTML error messages sounds plausible …

Odysseus: Any reason you're sticking with a Boinc Manager as old as 5.4.9? Wouldn't that grossly underclaim credits at SETI and screw things for your 'wingperson'? My G5 is happy enough with 5.8.17. (I'm avoiding Manager 5.10 onwards because the reduced benchmark count lessens my credits claim at SZTAKI.)

I think 5.8.17 will download other projects past any Milkyways stuck in the doorway. Sorry I can't check at the moment because I'm committed with SIMAP.

Odysseus: Any reason you're sticking with a Boinc Manager as old as 5.4.9? Wouldn't that grossly underclaim credits at SETI and screw things for your 'wingperson'? My G5 is happy enough with 5.8.17. (I'm avoiding Manager 5.10 onwards because the reduced benchmark count lessens my credits claim at SZTAKI.)

I think 5.8.17 will download other projects past any Milkyways stuck in the doorway. Sorry I can't check at the moment because I'm committed with SIMAP.

are people still getting workunits that aren't working?

It looks like while i was gone the permissions to our download directory got borked (read write and execute privileges were removed), so that would explain why transfers weren't working. This should be fixed now... let me know how it's going.

are people still getting workunits that aren't working?

It looks like while i was gone the permissions to our download directory got borked (read write and execute privileges were removed), so that would explain why transfers weren't working. This should be fixed now... let me know how it's going.

Absolutely, if you're talking about the parameters file refusing to d/l with an unspecified http error. I have a steady percentage of these on two pcs; I manually abort the transfer to clear them.

are people still getting workunits that aren't working?

It looks like while i was gone the permissions to our download directory got borked (read write and execute privileges were removed), so that would explain why transfers weren't working. This should be fixed now... let me know how it's going.

Absolutely, if you're talking about the parameters file refusing to d/l with an unspecified http error. I have a steady percentage of these on two pcs; I manually abort the transfer to clear them.

i went through our downloads directory, and it looked like a bunch of the subdirectories also had their permissions messed up. i've gone through and hopefully it should be fixed now...

Odysseus: Any reason you're sticking with a Boinc Manager as old as 5.4.9? Wouldn't that grossly underclaim credits at SETI and screw things for your 'wingperson'?

No, any BOINC version later than v5.2.6 or thereabout can handle the SETI@home Enhanced app’s method of claiming credit. I’m reluctant to upgrade because of bad experiences with vv5.8.x & 5.10.x on my G4 at work: it was crashing several times a day until I disabled the screensaver (it’s been mostly OK since I did that, but I plan to downgrade as soon as I find the time, to get the graphics back). Anyway, this behaviour, which has been consistent over several sub-versions now, and which has entailed considerable annoyance and wasted time, quite put me off—despite the improvements to the UI and scheduler.

BTW, old versions of BOINC don’t always underclaim at S@h: sometimes the benchmark-based claims turn out to be higher than those based on the app’s estimates of the size of the computation. People don’t complain about that quite as often, for some reason … ;) However, that’s never been an issue for me, one way or the other: I started out with BOINC Menubar, which IIRC came with a v5.2.13 client, several months before S@h stopped using the time-&-benchmarks method.

i went through our downloads directory, and it looked like a bunch of the subdirectories also had their permissions messed up. i've gone through and hopefully it should be fixed now...

Thank you, no problems so far today :)

I'll be home in 2 hours and will check but it looks like my quad is hung as there has been no changes to credits on my account for 2-3 hours. Travis, you may have fixed it after I got a "bad" D/L though as I last was at my puter 8 hours ago and it appeared fine until about 3 hours ago

---

Home now and all is well, attached 2 more puters

---

Getting these again !

2008-01-10 04:59:50|Milkyway@home|[file_xfer] Giving up on download of parameters_generated_1199517856_137056: file not found
2008-01-10 04:59:50|Milkyway@home|[file_xfer] Giving up on download of parameters_generated_1199517858_137087: file not found
2008-01-10 04:59:50|Milkyway@home|[file_xfer] Started download of file parameters_generated_1199517859_137111
2008-01-10 04:59:50|Milkyway@home|[file_xfer] Started download of file parameters_generated_1199517596_136840
2008-01-10 04:59:51|Milkyway@home|Deferring communication for 58 min 3 sec
2008-01-10 04:59:51|Milkyway@home|Reason: Unrecoverable error for result gs_84_1199517858_137087_1 (WU download error: couldn't get input files:<file_xfer_error> <file_name>parameters_generated_1199517858_137087</file_name> <error_code>-224</error_code> <error_message>file not found</error_message></file_xfer_error>)
2008-01-10 04:59:51|Milkyway@home|[file_xfer] Giving up on download of parameters_generated_1199517859_137111: file not found
2008-01-10 04:59:52|Milkyway@home|Deferring communication for 2 hr 19 min 53 sec
2008-01-10 04:59:52|Milkyway@home|Reason: Unrecoverable error for result gs_94_1199517859_137111_1 (WU download error: couldn't get input files:<file_xfer_error> <file_name>parameters_generated_1199517859_137111</file_name> <error_code>-224</error_code> <error_message>file not found</error_message></file_xfer_error>)
2008-01-10 04:59:52|Milkyway@home|[file_xfer] Finished download of file parameters_generated_1199517596_136840
2008-01-10 04:59:52|Milkyway@home|[file_xfer] Throughput 623 bytes/sec



How ever they seem to be aborting them selves now, I have nothing in the transfere tab...

got these messages


09-Jan-2008 20:04:53 [Milkyway@home] Sending scheduler request: To fetch work. Requesting 50732 seconds of work, reporting 0 completed tasks
09-Jan-2008 20:04:58 [Milkyway@home] Scheduler request succeeded: got 2 new tasks
09-Jan-2008 20:05:00 [Milkyway@home] Started download of parameters_generated_1199517856_137041
09-Jan-2008 20:05:00 [Milkyway@home] Started download of parameters_generated_1199517857_137065
09-Jan-2008 20:05:01 [Milkyway@home] Giving up on download of parameters_generated_1199517856_137041: file not found
09-Jan-2008 20:05:01 [Milkyway@home] Giving up on download of parameters_generated_1199517857_137065: file not found
09-Jan-2008 20:06:04 [Milkyway@home] Sending scheduler request: To fetch work. Requesting 50732 seconds of work, reporting 2 completed tasks
09-Jan-2008 20:06:09 [Milkyway@home] Scheduler request succeeded: got 0 new tasks
09-Jan-2008 20:07:09 [Milkyway@home] Sending scheduler request: To fetch work. Requesting 50732 seconds of work, reporting 0 completed tasks
09-Jan-2008 20:07:14 [Milkyway@home] Scheduler request succeeded: got 0 new tasks

---

i'm hoping these are just because of labstaff taking down the machine for awhile. are these problems still happening?