Some of the v.5.15 Models were dicey, though they typically showed a specific error (NEGATIVE THETA or NEGATIVE PRESSURE created). I\'d be guessing if I tried to diagnose your machine\'s failure. This would have been a good case to restart from a recent backup. Backups are highly recommended; they\'ve save many Runs. (Follow links in one of the Moderator\'s signatures for a README describing how to do it.)

It\'s only necessary to check the model options if you want to run a particular type of Model. If nothing is selected, the Server chooses one.

Your new Model seems to be running okay. Which version of boinc did you have before?

[Edited for typo.]

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

BOINC was 5.8.? whatever would have downloaded on December 26 2006 as I got this computer for Christmas...

I am sad now. Everything I did, my 14,000 credits, that wu has to be done over again by someone else?

And I still haven\'t seen a trickle up message for this new wu. But it seems that server stats says there is a lot to validate so things may be slow right now?

I understand about the backup, but I do run multiple projects, which will complicate things. It will sure be a learning experience for me ;-)

How can I make sure this wu will finish, though?

These forums are extremely informative, but I expect a lot of people don\'t read them regularly. How many CPDN people do backups do you think? Thanks for your advice...



D

...
I am sad now. Everything I did, my 14,000 credits, that wu has to be done over again by someone else?
...

The climate model uploads climate data at intervals:

* A summary every year
* A more detailed summary every 10 years
* A \'restart dump\' every 40 years (1960, 2000, and 2040).

Since you got to 1977, the servers have a copy of the 1960 restart dump, and have a good idea what happened between 1960 and 1977 as well.

It might be worth having a read through the README posts via the link in my signature (basically a collection of links to useful posts in the various forums).

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

To start with, you don\'t really have enough memory. The computer page says: 446.42 MB, so you probably only have 512 Megs, and no graphics card; so the onboard display chips are using up some of this 512 Megs; LOTS of it when the display gets \"busy\" with data.

This is like buying a big powerfull car, and only getting a 1 gallon fuel tank.
Ram is cheap! With a big processor like yours, you should have at LEAST 1 Gig, preferably 2.

Secondly, this is not like other projects; you get credits all the way through the modelling, and you KEEP those credits, no matter what happens to the model.

There are millions of possible combinations of the parameter values, and thus of models, so perhaps the one you had WILL get re-issued, and perhaps it won\'t. Some people crash dozens of them.
Just get more ram so that you don\'t crash the one that you have now, and carry on crunching.

Trickles are created on about the 3rd of December, and are uploaded by the time the model gets to the 4th.
So you only have to look at the globe display to see if you are anywhere near a trickle yet. Although looking at the globe MAY cause the model to crash, because of the lack of sufficient memory. :(

Backups are not that big a hassle. You just need to keep track of what you\'re doing. Perhaps print this, and tick off the steps as you go.

1) Set all of the projects to \"No new tasks\", (in the Projects tab), and allow all of the WUs for the other projects to complete and be returned.
2) Suspend the climate model in Tasks.
3) Make the backup. (See my instructions in my sig.)
4) Resume the climate model, and set the other projects, (but NOT cpdn), to \"Allow new tasks\".

In the event that you need to restore a climate model:

1) Follow the above to run down the WUs of the other projects.
2) Delete the working BOINC folder.
2) Copy the backup BOINC folder, and paste it into the empty BOINC folder location.
3) Reboot.
4) Because BOINC was Suspended when the backup was made, it won\'t start running immediately.
6) Set the other projects to \"Allow new tasks\", so that they start running again.
7) Unsuspend BOINC (in the menu.)
8) Unsuspend the climate model in Tasks, and allow it to continue from the previously saved position.

---

Backups: Here

Thanks, everybody!

I did get credit this morning from my new wu.

I will, of course, take your advice on more memory. I have two teenagers using this machine as well. One always on Desert Collosus and YouTube, the other always on My Space and AIM and The All American Rejects ;-)

Sigh.. what a world, eh?

Again, many thanks. I will now pay more attention ;-)




D

Haydn, since it says \"Fatal crash :-(\" I think the only way you can recover it is if you have a backup, so I hope you do!

Regards,

Had a similar problem to this but much longer error report (38 pages!) before ending with following:

scan: stderr.txt
scan: SULPC_OXIDANTS_19_A2_1990.gz
scan: SULPC_OXIDANTS_19_A2_1990.mod.gz
scan: volc_v00.gz
scan: yafbg.astart.gz
Not a JPEG file: starts with 0x01 0xda
Not a JPEG file: starts with 0x01 0xda
Not a JPEG file: starts with 0x01 0xda
Not a JPEG file: starts with 0x01 0xda
Not a JPEG file: starts with 0x01 0xda
Not a JPEG file: starts with 0x01 0xda
CPDN Monitor - Quit request from BOINC...

</stderr_txt>
]]>

Background info

Shut down BIONIC following year 11 trickle up.
Made backup copy.
Ran full disk defrag (including boot time MFT defrag).
Restarted BIONIC.
Got whisky.
When I then checked BIONIC, found that a new model had been downloaded and started. Interestingly a UK MET Office modele starting in 1810. This has appealed to my curiosity so will continue to run with it.

Question: Could I merge a restore of the crashed model to run with the new one, or is it better to leave for the Oxford team to reallocate to another user?

---

John,

The crashed model looks eminently restorable, since the crash was caused by something on the PC rather than something in the model (-107 error). Don\'t worry about the long list of messages: these are, in fact, only warnings and may have happened at any time, rather than when the model actually crashed.

You won\'t be able to merge the backup with the running model. You\'ll either have to run the new model (which is a shorter \'slab\' model), or restore the old model and lose the new model (though you could, of course, back that up and restore it when the old model has finally finished - but that\'s a long way off!).

If I were you in this situation, I would run the new slab model to completion (they\'re much quicker), then restore the old coupled model and carry on as before ...

Iain

PS If you want to prevent a new model being downloaded when a crash happens, then press the \'No new tasks\' button on BOINC Manager.

PPS If you do restore the old model, then the warning messages on the results page for that model won\'t go away. However, the server will happily accept any new trickles right up to the end of the run.

John,

The crashed model looks eminently restorable, since the crash was caused by something on the PC rather than something in the model (-107 error). Don\'t worry about the long list of messages: these are, in fact, only warnings and may have happened at any time, rather than when the model actually crashed.

You won\'t be able to merge the backup with the running model. You\'ll either have to run the new model (which is a shorter \'slab\' model), or restore the old model and lose the new model (though you could, of course, back that up and restore it when the old model has finally finished - but that\'s a long way off!).

If I were you in this situation, I would run the new slab model to completion (they\'re much quicker), then restore the old coupled model and carry on as before ...

Iain

PS If you want to prevent a new model being downloaded when a crash happens, then press the \'No new tasks\' button on BOINC Manager.

PPS If you do restore the old model, then the warning messages on the results page for that model won\'t go away. However, the server will happily accept any new trickles right up to the end of the run.

Thanks Iain

Following your advice, will continue with the new slab model until finished and retain the old model and restore when complete. Also will put my digit on the \"No new tasks\" button.

Reagrds John

---

Did the whisky help? ;-)

---

Visit the Scotland team

Did the whisky help? ;-)

Sure did!

Nice to here from my north of the border pyrotechic expert now I\'ve transferred from the BBC experiment!

Kind regards

John

---

Oh, I didn\'t realise you were that John! ;-) Welcome to CPDN and to this forum and well done for continuing to support the project.

Best regards,

---

Visit the Scotland team

Oh, I didn\'t realise you were that John! ;-) Welcome to CPDN and to this forum and well done for continuing to support the project.

Best regards,

Thanks for your continued encouragement.

My offspring are in their early twenties, and my vested interest in this model is to provide evidence to encourage the world’s decision takers and individuals, that they all share a responsibility to pull out all the stops to provide the actions necessary to provide a tolerable global weather environment for future generations.

For this reason, I feel that offering spare computer capacity for this BOINIC project in particular is a small price to pay and well worth the effort and dare I say the most worthy of all the BIONIC projects.

Having said that, I’ve also been actively involved in local politics and have been instrumental in implementing waste recycling – introduction of garden waste, bottle banks, paper, cardboard and metal collections and improved local public transport and, as indicated in my BBC communications reducing my own carbon footprint by minimising flights on annual holidays, walking not driving to local shops, etc. This is where real positive gains can be made by individuals in minimising the adverse effects of climate change.

Regards

John

---

Sorry, last message sounds pompous. It\'s not really for this thread

---

Maybe not in the most relevant section, but certainly not pompous!

---

Cpdn news

I agree Mo! ;-) Good for you, John.

---

Visit the Scotland team

This is the comments thread. Please post any queries, suggestions, and so forth here.

This is the comments thread. Please post any queries, suggestions, and so forth here.

Having difficulty with this model which appears to be hung up at 47.639% completion. Suspend and resume have had no effect. CPU time and time to completion are both increasing! The cycle goes from 47.637 to 47.639% Model time is 2038
04/11/2007 10:20:56||Starting BOINC client version 5.8.11 for windows_intelx86
04/11/2007 10:20:56||log flags: task, file_xfer, sched_ops
04/11/2007 10:20:56||Libraries: libcurl/7.16.0 OpenSSL/0.9.8a zlib/1.2.3
04/11/2007 10:20:56||Data directory: C:\\Program Files\\BOINC
04/11/2007 10:20:57||Processor: 1 GenuineIntel Intel(R) Celeron(R) CPU 2.66GHz [x86 Family 15 Model 4 Stepping 1] [fpu tsc pae nx sse sse2 mmx]
04/11/2007 10:20:57||Memory: 502.08 MB physical, 1.20 GB virtual
04/11/2007 10:20:57||Disk: 71.24 GB total, 17.14 GB free
04/11/2007 10:20:57|climateprediction.net|URL: http://climateprediction.net/; Computer ID: 589201; location: home; project prefs: default
04/11/2007 10:20:57||General prefs: from http://bbc.cpdn.org/ (last modified 2006-04-16 19:16:34)
04/11/2007 10:20:57||Host location: none
04/11/2007 10:20:57||General prefs: using your defaults
04/11/2007 10:20:57||Reading preferences override file
04/11/2007 10:21:37|climateprediction.net|Restarting task hadcm3iozn_cpjr_2000_80_35898880_6 using hadcm3i version 544

---

Here is the model\'s results page on the server:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=6761905

From this, we can see that the last \'trickle\' received from the model was received on the 25th October, and would have been sent on the model date of 4th December 2038.

Could you have a look at the current \'model date\' (you find this by selecting \'show graphics\'), and post the result here?

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

Here is the model\'s results page on the server:

http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=6761905

From this, we can see that the last \'trickle\' received from the model was received on the 25th October, and would have been sent on the model date of 4th December 2038.

Could you have a look at the current \'model date\' (you find this by selecting \'show graphics\'), and post the result here?

Ah, another little problem - I can\'t get the show graphics functon to work. It has been unpredictable for some time in so far as when requested by one user it displays in anothers\' profile. Currently no success at all.

I suspect the information will be buried in some log somewhere, but not sure which one... can anyone shed some light?


Regarding the show graphics problem, what sort of installation do you have, and what sort of system is it installed on?

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ