Today my computer began an new project hadcm3iozn_cplf_2000_80_155898940_7.
The expiration date is 13. 12. 1901.
I get the following messages:

08.12.2007 15:58:39||Starting BOINC client version 5.2.13 for windows_intelx86
08.12.2007 15:58:39||libcurl/7.14.0 OpenSSL/0.9.8 zlib/1.2.3
08.12.2007 15:58:39||Data directory: C:\\Programme\\BOINC
08.12.2007 15:58:39||Processor: 1 AuthenticAMD AMD Athlon(tm) 64 Processor 3800+
08.12.2007 15:58:39||Memory: 2.00 GB physical, 3.85 GB virtual
08.12.2007 15:58:39||Disk: 37.28 GB total, 23.85 GB free
08.12.2007 15:58:39|climateprediction.net|Computer ID: 430806; location: home; project prefs: home
08.12.2007 15:58:39||General prefs: from climateprediction.net (last modified 2005-05-17 11:10:39)
08.12.2007 15:58:39||General prefs: no separate prefs for home; using your defaults
08.12.2007 15:58:39|climateprediction.net|Result hadcm3iozn_cplf_2000_80_155898940_7 is 133038.10 days overdue.
08.12.2007 15:58:39|climateprediction.net|You may not get credit for it. Consider aborting it.
08.12.2007 15:58:41||Remote control not allowed; using loopback address
08.12.2007 15:58:41|climateprediction.net|Resuming computation for result hadcm3iozn_cplf_2000_80_155898940_7 using hadcm3i version 544
08.12.2007 15:58:42||Suspending work fetch because computer is overcommitted.
08.12.2007 15:58:42||Using earliest-deadline-first scheduling because computer is overcommitted.

What would you propose to do?

---

The 1901 problem is well known. It\'s caused by
1) A large change in the time manually made on the computer
2) A failure of the cmos backup battery in the computer, so that the computer no longer remembers it\'s settings when turned off.
The main setting on most computers that is not a default, is the \'real\' time.

BOINC keeps a record of the current time, so that it can decide which work units are in danger of missing their deadline. So when the time goes way back to the default, (early 1980\'s), a bug in early versions of BOINC sets it\'s value to 1901, and then keeps this.

If you\'re running other projects as well as cpdn, then the WUs from these projects will be rejected as being too late.
But if you\'re only running climate models, the only problem will be constant nagging messages, as this project ignores the deadline message.

It\'s possible to edit the client_state.xml file to fix things.
I think the instructions are somewhere in the Number crunching section of these boards.

---

Backups: Here

One further thing you can do is to upgrade your Boinc manager to version 5.8.16 or more recent, since the newest boinc managers don\'t have the bug which causes the deadline to reset.

It won\'t fix the deadline for that particular model, but following the instructions in the \'README - crashes and other problems / how to fix a 1901 deadline\' will solve that.

---

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

An additional cause to the ones that Les posted (and probably the one that hit you).

There\'s a bug in BOINC versions earlier than 5.8.16 which causes any deadlines that are more than 1000 days in the future to be forced back to 1901 before you\'ve even downloaded any of the files for the task. Unfortunately the new HadCM3 tasks all have their deadlines set in 2012, which is more than 1000 days in the future ...

---

"The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer