My model crashed. Then I restored it from backup and continued crunching. But on server there is state with problems, thus points are granted to me.
http://climateapps2.oucs.ox.ac.uk/cpdnboinc/result.php?resultid=6106654

Can this be changed and how?

---

Hi,

The first reported error will stick on the website, but is ignored by the scientists, who are only interested in the uploaded data. So you can ignore it.

---

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

Hi,

The first reported error will stick on the website, but is ignored by the scientists, who are only interested in the uploaded data. So you can ignore it.

Hmm. I\'ve got one thought about it:
When my model crashes - I restroe it from backup. If this computer is online (the most of my computers, which are in office with LAN) - it says to server, thta model crashed and downloads new model. In the morning I come to work and restore model from backup, which I make every morning, when come to work. So, serve r then gives this model to someone else, with last number in name, increased by one. And I can get model with last number in name not equal 0, which can means, that someone crashed model, reported server, and then restored model from backup and cotinued crunching, dispite everything happened.
So, what is the \"Idea of crunching the same model by some persons simulatiously\"?

Is there any mechanism, which catches this model, when this happens?

And one more qustion: as I understood, if there were not trickles for some days (45, AFAIK), the results are ot accepted by the server. How much time has to be elapsed, when crashed model has to be posted to new person? Are those 45 days is that period?

---

The way to get around that is to set \"No new tasks\" in Projects tab and \"Network activity suspended\" in Activity.

Servers do not refuse data, unless the database is full or otherwise disabled. If they receive duplicates from a computer because of running from a backup, no additional credit awarded and that continues until the last previously received Trickle is passed by the restored Run.

If duplication occurs because of the circumstance you describe, both Runs will be honored and the science team will compare them. That said, I suspect the situation seldom comes up; people who restore typically do it in short order -- and the server waits before reissue.

---

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

A (small number) of duplicates is potentially quite useful, for comparing the results between different platforms. Too many would be a waste of course. If there were none the researchers would probably need to deliberately issue a few in order to do the comparisons.

---

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