Hi,

I have a hadcm3 WU which is sending trickles as normal but they aren\'t being shown as received.

clicky

Percentage completed is increasing normally as well - currently 50.734% .

Do I have to abort? thanks :)

Hi, Chidge, welcome to the forum.

Hmmm. There was a problem with logging Trickles Friday thru the weekend. It\'s been resolved.

Given that your Trickles are going up normally, they should be in the database and should be credited sooner or later. (After the weekend outage, my Trickles took a day longer to appear than some others who posted.)

No need to abort the Run. These things tend to come out right in the end.

---

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

Hi Chidge

Did you receive something similar to this for your last upload:

29/11/2006 7:58:28 PM|climateprediction.net|Scheduler request to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi succeeded[

If so, then the trickle should be on the upload server, but possibly not processed and transfered to the trickle server yet.

Also, have a look in the folder for the cpdn project, and see if the trickle_up icon is still there, but \"greyed\".
I\'ve got one like this from an upload an hour ago.

The other possibility is a \'looping\' model. To work out if this is happening, keep making a note of the \'model date\' (displayed on the \'show graphics\' screen). If this is cycling between 1st Dec and some arbitrary date in the next year, but never reaching the next year\'s 1st Dec, then it may be a looping model.

---

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

Hi,

thank you for all the replies :)

this is the last trickle up message I have:

29/11/2006 14:59:20|climateprediction.net|Sending scheduler request to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi
29/11/2006 14:59:20|climateprediction.net|Reason: To send trickle-up message
29/11/2006 14:59:20|climateprediction.net|(not requesting new work or reporting completed tasks)
29/11/2006 14:59:25|climateprediction.net|Scheduler request succeeded

I looked for the trickle_up icon in the cpdn folder within boinc but couldn\'t find anything - might not be looking in the right place? :P

As for a \'looping model\' I\'m currently at 24/05/2002 50.92% so I will keep checking to see if that is increasing normally

Finally though if I scroll to the top of the messages I have:
21/11/2006 18:25:37|climateprediction.net|Task hadcm3lbm_b5ed_05289480_1 is 134362.43 days overdue.
21/11/2006 18:25:37|climateprediction.net|You may not get credit for it. Consider aborting it.

Looking at the report deadline it says 1901 - that cant be right?

Incidentally I have been making backups of my cpdn folder - the most recent I have is from the beginning of November

I\'m not 100% sure but I think I\'ve read about a crash or bug that makes BOINC set the deadline to 1901. But apart from annoying messages I don\'t think it hurts...

Don\'t abort it unless it turns out to be a looper! The deadline date won\'t be imposed as long as the computer trickles fairly regularly.

---

Cpdn news

The calendar resets to 1901 when the motherboard battery dies or the BIOS jumper is grounded.

If you\'re not running other boinc clients, it\'s no big deal.

---

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

Hi,

I\'m thinking my model might be looping then

Checking it now it says 03/02/2002 50.728%
Yesterday it was 24/05/2002 50.92%

I have a backup from when the model was at 38% (not sure about the model date)
Would it be worth using this or will it fail at the same point again?

thanks for all your help

Carl posted that it was most important to get past 50% (the \"hindcast\" part).

12% is a lot of rerun (for which you won\'t receive credits) and it might fail at the same place. Some are known to continue on. A friend had the problem and retrys failed. (She had a backup which involved less rerun, so she retried at least twice before performing the mercy-killing.) So, in the end,it\'s your call.

---

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

In general the odds of getting past the looping point is extremely low, so it\'d be best to zap this model and start with another one. On the other hand, if you\'re going to crash anywhere, at 50.x% is the very best place to do it, since it\'s just after the upload!

---

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