I\'m running boinc on Archlinux (in a OpenSolaris zone) and I have a problem where climateprediction creates no output files.

I\'ve checked that both boinc and the binaries in projects/climateprediction.net/ are all 32 bit. eg:

[boinc@arch services]$ file /usr/bin/boinc
/usr/bin/boinc: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.18, stripped

Yet I see this a frequently in the stdout file:

06-May-2010 07:36:55 [climateprediction.net] Task hadsm3dhet2_k2o9_006613115_4 exited with zero status but no \'finished\' file
06-May-2010 07:36:55 [climateprediction.net] If this happens repeatedly you may need to reset the project.

There is nothing in the stderr file. it\'s size is zero.

I have reset the project, but it still appears that there is no processing happening.

[boinc@arch ~]$ boinccmd --get_host_info
  timezone: -14400
  domain name: arch
  IP addr: 
  #CPUS: 4
  CPU vendor: GenuineIntel
  CPU model: Intel(r) Atom(tm) CPU D510   @ 1.66GHz [Family 6 Model 28 Stepping 10]
  CPU FP OPS: 410739332.961561
  CPU int OPS: 1282749429.406051
  CPU mem BW: 1000000000.000000
  OS name: Linux
  OS version: 2.6.18
  mem size: 2121326592.000000
  cache size: 524288.000000
  swap size: 1064300544.000000
  disk size: 49815768576.000000
  disk free: 48937944576.000000
[boinc@arch ~]$ 

Anyone have any ideas about how I might resolve this?

That message is mentioned in our Readme files, in the one labelled Crashes and other problems. (Top section, 4th one down.)
It\'s the same as the description of the Earth, in The Hitcher\'s Guide to the Universe: Mostly harmless.
It\'s also a very old message; it was changed in more recent versions of BOINC.

What to do:
1) Ignore it. Usually nothing will happen to the model.
2) Follow the advice, and you WILL lose the model.

As for \"no output files\": you\'ll only get an output file when the model has produced 24 trickles. If it\'s not running, then you\'ll never get a trickle.

As for what to do, there\'s always problems when being adventurous, so good luck.

(I don\'t run Linux, but how about just running the computer on it\'s own, and not as part of an OpenSolaris zone?)

---

Backups: Here

If you can get a gui started, you\'ll be able to see the status of the models, and also their amount of progress.

---

Backups: Here

I do run Linux, but not in a virtual machine from OpenSolaris on an Atom processor. :)

The stderr on the result pages for the failed tasks is unfortunately not helpful in troubleshooting either with exit code 22 and a long listing of \"Model crashed: 7R\". If some type of library was missing, I would think it would be listed here.

Ok, so I launch \"boincmgr\" to see the gui,

It shows \"Error\" above \"retrieving current status\". The messages button blinks and shows a window formatted view of the stdout file:

Sat May 8 13:23:41 2010|climateprediction.net|If this happens repeatedly you may need to reset the project.
Sat May 8 13:23:41 2010|climateprediction.net|Restarting task hadsm3dhet2_k2a8_006612610_4 using hadsm3 version 0
Sat May 8 13:26:08 2010|climateprediction.net|Task hadsm3dhet2_k2a8_006612610_4 exited with zero status but no \'finished\' file
Sat May 8 13:26:08 2010|climateprediction.net|If this happens repeatedly you may need to reset the project.
Sat May 8 13:26:08 2010|climateprediction.net|Restarting task hadsm3dhet2_k2a8_006612610_4 using hadsm3 version 0
Sat May 8 13:28:04 2010|climateprediction.net|Task hadsm3dhet2_k2a8_006612610_4 exited with zero status but no \'finished\' file
Sat May 8 13:28:04 2010|climateprediction.net|If this happens repeatedly you may need to reset the project.
Sat May 8 13:28:04 2010|climateprediction.net|Restarting task hadsm3dhet2_k2a8_006612610_4 using hadsm3 version 0
Sat May 8 13:29:40 2010|climateprediction.net|Computation for task hadsm3dhet2_k2a8_006612610_4 finished
Sat May 8 13:29:40 2010|climateprediction.net|Output file hadsm3dhet2_k2a8_006612610_4_1.zip for task hadsm3dhet2_k2a8_006612610_4 absent
Sat May 8 13:29:40 2010|climateprediction.net|Output file hadsm3dhet2_k2a8_006612610_4_2.zip for task hadsm3dhet2_k2a8_006612610_4 absent
Sat May 8 13:29:40 2010|climateprediction.net|Output file hadsm3dhet2_k2a8_006612610_4_3.zip for task hadsm3dhet2_k2a8_006612610_4 absent


When I compare that with boinc running on my other machine (Mac client) it shows something like this:

07-May-2010 07:33:47 [---] Starting BOINC client version 6.10.17 for i686-apple-darwin
07-May-2010 07:33:47 [---] log flags: file_xfer, sched_ops, task
07-May-2010 07:33:47 [---] Libraries: libcurl/7.19.4 OpenSSL/0.9.7l zlib/1.2.3 c-ares/1.6.0
07-May-2010 07:33:47 [---] Data directory: /Library/Application Support/BOINC Data
07-May-2010 07:33:47 [---] Processor: 2 GenuineIntel Genuine Intel(R) CPU 1400 @ 1.83GHz [x86 Family 6 Model 14 Stepping 8]
07-May-2010 07:33:47 [---] Processor features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM SSE3 MON VMX EST TM2 TPR PDCM
07-May-2010 07:33:47 [---] OS: Darwin: 10.3.0
07-May-2010 07:33:47 [---] Memory: 2.00 GB physical, 15.41 GB virtual
07-May-2010 07:33:47 [---] Disk: 265.88 GB total, 15.17 GB free
07-May-2010 07:33:47 [---] Local time is UTC +10 hours
07-May-2010 07:33:47 [---] No usable GPUs found
07-May-2010 07:33:47 [---] Not using a proxy
07-May-2010 07:33:47 [climateprediction.net] URL http://climateprediction.net/; Computer ID 1024861; resource share 100
07-May-2010 07:33:47 [---] No general preferences found - using BOINC defaults
07-May-2010 07:33:47 [---] Reading preferences override file
07-May-2010 07:33:47 [---] Preferences limit memory usage when active to 1024.00MB
07-May-2010 07:33:47 [---] Preferences limit memory usage when idle to 1843.20MB
07-May-2010 07:33:48 [---] Preferences limit disk usage to 10.00GB
BOINC initialization completed, beginning process execution...
07-May-2010 07:33:48 [climateprediction.net] Restarting task hadsm3dhet2_kh7m_006631956_4 using hadsm3 version 609
07-May-2010 07:33:48 [climateprediction.net] Restarting task hadsm3dhet2_kbpw_006624838_6 using hadsm3 version 609
07-May-2010 11:38:10 [---] Running CPU benchmarks
07-May-2010 11:38:10 [---] Suspending computation - running CPU benchmarks
07-May-2010 11:38:10 [---] Running CPU benchmarks
07-May-2010 11:38:10 [---] Running CPU benchmarks
07-May-2010 11:38:41 [---] Benchmark results:
07-May-2010 11:38:41 [---] Number of CPUs: 2
07-May-2010 11:38:41 [---] 1867 floating point MIPS (Whetstone) per CPU
07-May-2010 11:38:41 [---] 3948 integer MIPS (Dhrystone) per CPU
07-May-2010 11:38:42 [---] Resuming computation


The major difference I notice is that the linux build is trying to start hadsm3 version 0 but the Mac is doing hadsm3 version 609.

Does this help?

Version 0! Not good. Corrupt file?
You may need to Reset the project to kill off all of the files, and then download them again.

Sat May 8 13:29:40 2010|climateprediction.net|Computation for task hadsm3dhet2_k2a8_006612610_4 finished

It\'s crashed.

---

Backups: Here

Another thought.
Version 0 may mean that, even though the setup has downloaded the necessary files, it\'s trying to run them as a different OS.
Perhaps it\'s a problem with the emulator/cross compile for OpenSolaris or Archlinux.

edit
By this, I mean that the OS isn\'t recognizing the part of the code that has the version number.
So it may not recognise other parts of the model code either.

At least I think that\'s what I mean. :)