climateprediction.net home page

Posts by Conan

1) Questions and Answers : Windows : Multiple download errors (Message 51313)
Posted 2 hours ago by Profile Conan
Thanks for the reply Dave.

Unfortunately doing a detach and reattach (which I had done a few days ago) did not make any difference, still can't download a Short model.

BOINC shows that I have over 23 GB available and I can't see any problems with my preferences that would cause a download issue.

It is strange, as all the support files that get downloaded with each work unit, download without an issue, but the main file dies.

Conan
2) Questions and Answers : Windows : Multiple download errors (Message 51311)
Posted 6 hours ago by Profile Conan
Rather than start a new thread I am re-using this one as it is related to what I am getting.
I am getting the following error on nearly all downloads to a Windows 32 bit XP machine, machine Host number = 1352573.


app_version download error: couldn't get input files:
<file_xfer_error>
<file_name>hadam3p_afr_7.22_windows_intelx86.exe</file_name>
<error_code>-200</error_code>
</file_xfer_error>

It is not just on the African models but also Pacific North West and Short models.

I did see a message saying I did not have any memory available? I had over 13 GB but this seems to be not enough.
I have detached a number of other projects from this computer, deleted a number of other files and this has now given me around 16 GB of disk space for BOINC to play with.
I have placed Climateprediction.net in my anti-virus (TrendMicro) exclusion area plus each executable.
Of all the executable files that have tried to download, not counting the graphic files, 3 have consistently failed to download, always just 23.6 kB in size and therefore generating checksum errors due to the wrong size.
It is only this one file that wont download as all the other files download even if the main one fails.

I have the following files that appear to be correct size

hadam3p_anz_6.10_windows_intelx86.exe
hadam3p_anz_um_6.10_windows_intelx86.exe
hadcm3n_6.07_windows_intelx86.exe
hadcm3n_um_6.07_windows_intelx86.exe
hadrm3p_anz_um_6.10_windows_intelx86.exe

The following files all have just 23.6 kB in size and fail every download

hadam3p_afr_7.22_windows_intelx86.exe
hadam3p_pnw_7.22_windows_intelx86.exe
hadcm3s_7.24_windows_intelx86.exe

I have managed to actually start running a few of the Full models and even got to the first trickle on one of them but they all then failed with errors.

My other Windows machine had similar errors for a few work units then downloaded and started running one without further problems, they are both AMD Phenom CPUs both with 4 GB of RAM, the 6 core is running OK the 4 core is the problem and it also only has an 80 GB hard drive.

Any help would be appreciated.

Thanks
Conan
3) Message boards : climateprediction.net Science : Reporting misconfigured computers (Message 51288)
Posted 5 days ago by Profile Conan
Just found this computer owned by Piers.
It has failed to run a single work unit of the 911 that have been sent to it.
Seems to be missing the 32 Bit Linux libraries.

Host 1346377

Conan
4) Message boards : Number crunching : HadCM3 short - errors galore (Message 50665)
Posted 89 days ago by Profile Conan

Also every time I visit the climate site and visit my account I have to re-loggin despite checking the "keep logged in box"... its that a cookie issue?.


This also happens to me when I click the links from Free-DC site to get to Climate.
If I go to the Berkeley list of projects and click on the Climate link from there I don't have to re-log in, as it shows I already am.

I can't see the difference but always happens, no idea why.

Conan
5) Message boards : Number crunching : Extremely high work units done. (Message 50100)
Posted 138 days ago by Profile Conan
The stats have been exported. You should see some of the totals, they have gone past FreeDC credit counter limits.

Conan
6) Message boards : Number crunching : UK Met Office HADAM3P (global only) with MOSES II landsurface scheme v7.03 (Message 49108)
Posted 256 days ago by Profile Conan
I must be one of the lucky ones as I have managed to finish a MOSES without error (as far as I can tell).
See WU 8804573

Just waiting for the validation and credits to catch up.

Took just over 311 hours run time.

Conan
7) Message boards : Number crunching : HTTP gateway timeout (Message 48877)
Posted 277 days ago by Profile Conan
Couldn't find another thread relating to this issue so started another.

I noticed that my account showed I had a WU that was not on my computer.
So I investigated the BOINC Manager log for answers.

Shortly before midnight (Australian time) on the 23/4/14 my computer tried to download a new work unit.

This apparently failed with the message Scheduler Request Failed : HTTP gateway timeout

As I never got this WU it will stay on my list for a year before getting resent.

Why does the system think it downloaded files when nothing connected and just timed out?

Conan
8) Questions and Answers : Preferences : Cannot Update Computer Preferences (Message 48867)
Posted 278 days ago by Profile Conan
If I click on the link for each model type under my preferences I get an error that the Page can't be found, so you can't get any information on the different models to see what is required.

Conan
9) Message boards : Number crunching : Missing Credits from Host Total (Message 43629)
Posted 1118 days ago by Profile Conan
Well I had to fiddle didn't I?
I updated the OS of the Host originally having the credit issues after it had dropped from 491,954 credits to 290,613, I went from 32 bit to 64 bit.
My total has now dropped again to just 184,022 credits.

So HOST 1186021 should have 597,150 Total Points (shows 214,415 points which is the displayed WU total in my account which has yet to be archived)

HOST 1188489 should have 491,954 Total Points (did show 290,613 NOW showing 184,022 points which is the viewable WU total not yet archived)

So for some reason the system is only giving a Host the value of the viewable Work Units accessible in your account, not the overall total of all work processed by the Host (which is Viewable Work Units PLUS Archived Work Units).

Hope is high that Climate Prediction Team will resolve this issue in the near future.

After all it is a new year and stress has taken a back seat over the holiday break.

Thanks
Conan
10) Message boards : Number crunching : Missing Credits from Host Total (Message 43591)
Posted 1131 days ago by Profile Conan
I've certainly had credit totals for a host reduced when that host was merged, which anomalous behaviour was passed onto the project team (failure to carry across archived results?). However, my overall total didn't reduce.


G'Day Iain,
My host total stayed the same after I merged the host. It dropped after the first trickle was received.
My overall total appears OK.
However I would like my Hosts to be reporting the actual amount of work they have performed, at the moment it looks like they have done very little work for the total that I have, also to get the correct placing in host stat rankings, plus I like things to add up. At the moment very little adds up unless you do a lot of maths like I have in this thread.

If something was mucked up in the first place then that same something can be repaired.

If it has been reported then at least project staff know about it and I hope can get round to fixing it up.
At the moment I know they have lots of other more important issues, particularly with data storage problems, so I will continue to wait.

As I said in my previous post, I have waited a year now so I can wait a little longer.

Thanks again, and have a good Christmas.

Conan


Next 10 posts



Copyright © 2002-2014 climateprediction.net