It constantly get the following error message:

Deferring communication with project for 23 hours, 55 minutes, and 24 seconds



I have reset the project many times. I have detached and attached many times.

This is on a Multiprocessor PPC64 system running Linux 2.6.11-pseries

I wanted to get the basic package running before I make any alterations to the SRC to make it run more efficently. I also wanted to use the SRC to aide in porting it over to my MainFrame.

Any assistance would be VERY grateful.

Thanks!

tj at dreamvisionweb_dot_com

could you please post the full output of the boinc client ?
That would help for further assistance.

regards
Crunch3r

---

Join BOINC United now!

Sure! I am sorry that I did not think of that in the first place:



2005-09-25 19:13:05 [---] Starting BOINC client version 4.72 for powerpc64-unknown-linux-gnu
2005-09-25 19:13:05 [---] Data directory: /home/contest/twalker/test/boinc_public/client
2005-09-25 19:13:05 [---] Processor Inventory: 8 Processor(s)
2005-09-25 19:13:05 [---] Memory Inventory: Memory total - 7.00 GB, Swap total - 1.86 GB
2005-09-25 19:13:05 [---] Disk Inventory: Disk total - 110.27 GB, Disk available - 30.33 GB
2005-09-25 19:13:05 [SETI@home] Computer ID: 1442882; location: work; project prefs: default
2005-09-25 19:13:05 [---] No general preferences found - using BOINC defaults
2005-09-25 19:13:05 [---] Remote control not allowed; using loopback address
2005-09-25 19:13:05 [SETI@home] Deferring communication with project for 3 hours, 5 minutes, and 48 seconds
2005-09-25 19:13:05 [SETI@home] Deferring communication with project for 3 hours, 5 minutes, and 48 seconds

Sure! I am sorry that I did not think of that in the first place:



2005-09-25 19:13:05 [---] Starting BOINC client version 4.72 for powerpc64-unknown-linux-gnu
2005-09-25 19:13:05 [---] Data directory: /home/contest/twalker/test/boinc_public/client
2005-09-25 19:13:05 [---] Processor Inventory: 8 Processor(s)
2005-09-25 19:13:05 [---] Memory Inventory: Memory total - 7.00 GB, Swap total - 1.86 GB
2005-09-25 19:13:05 [---] Disk Inventory: Disk total - 110.27 GB, Disk available - 30.33 GB
2005-09-25 19:13:05 [SETI@home] Computer ID: 1442882; location: work; project prefs: default
2005-09-25 19:13:05 [---] No general preferences found - using BOINC defaults
2005-09-25 19:13:05 [---] Remote control not allowed; using loopback address
2005-09-25 19:13:05 [SETI@home] Deferring communication with project for 3 hours, 5 minutes, and 48 seconds
2005-09-25 19:13:05 [SETI@home] Deferring communication with project for 3 hours, 5 minutes, and 48 seconds

Looks normal to me. You should stop reseting the project because there are allready two hostids asigned to your machine.
You should wait until the time passed and see what message comes next.

Have you allready compiled the seti app. ?

---

Join BOINC United now!

2005-09-25 19:13:05 [SETI@home] Deferring communication with project for 3 hours, 5 minutes, and 48 seconds
2005-09-25 19:13:05 [SETI@home] Deferring communication with project for 3 hours, 5 minutes, and 48 seconds

Is the name resolution OK ? - Can you nslookup setiboinc.ssl.berkeley.edu ?

Can you ping/traceroute to setiboinc.ssl.berkeley.edu ?

Is the connection to the servers via port 80 OK ? - Do you get a prompt from the command "telnet setiboinc.ssl.berkeley.edu 80" ?

---

Looks normal to me. You should stop reseting the project because there are allready two hostids asigned to your machine.

It is not neccesary, you can try a lot of detaches and attaches. You could later merge your computers via the account site into one.

If the networktests are OK, can you try to remove all the files, except boinc_client in the directory /home/contest/twalker/test/boinc_public/client, and try to attach to the project.

You can also try to re attach, and if it not works to make an update (boinc_client -update http://setiathome.berkeley.edu).

---

2005-09-25 19:13:05 [SETI@home] Deferring communication with project for 3 hours, 5 minutes, and 48 seconds
2005-09-25 19:13:05 [SETI@home] Deferring communication with project for 3 hours, 5 minutes, and 48 seconds

Is the name resolution OK ? - Can you nslookup setiboinc.ssl.berkeley.edu ?

Can you ping/traceroute to setiboinc.ssl.berkeley.edu ?

Is the connection to the servers via port 80 OK ? - Do you get a prompt from the command "telnet setiboinc.ssl.berkeley.edu 80" ?

It is obviously working he allready got 2 hostids assigned to his machine.
Reseting will make it even worse the time will increase.
I think there is no problem with the boinc client at all.

the next logical step would be to compile the seti app.
I would sugest using the fftw3 patched source from ned slider
(http://www.pperry.f2s.com/files/seti_boinc-client-fftw3-2004-12-01.tar.gz)

But first of all get the fftw3 source and configure it with
./configure --enable-fma --enable-float ( and --enable-altivec if your cpu supports it)
the --enable-fma awitch is important for ppc cpus.

regards
Crunch3r

---

Join BOINC United now!

Reseting will make it even worse the time will increase.

No.

---

Alright. I will try the src that you mentioned below. Yes, I can connect. I even tried to disconnect from the setiathome.berkeley.edu address and attach to the setiboinc.ssl.berkeley.edu address and this is what I got:

--------------------------- SNIP HERE -----------------------------------
2005-09-25 23:46:08 [---] Starting BOINC client version 4.72 for powerpc64-unknown-linux-gnu
2005-09-25 23:46:08 [---] Data directory: /home/contest/twalker/test/boinc_public/client
2005-09-25 23:46:08 [---] Processor Inventory: 8 Processor(s)
2005-09-25 23:46:08 [---] Memory Inventory: Memory total - 7.00 GB, Swap total - 1.86 GB
2005-09-25 23:46:08 [---] Disk Inventory: Disk total - 110.27 GB, Disk available - 30.29 GB
2005-09-25 23:46:08 [http://setiboinc.ssl.berkeley.edu/] Computer ID: not assigned yet; location: ; project prefs: default
2005-09-25 23:46:08 [---] No general preferences found - using BOINC defaults
2005-09-25 23:46:08 [---] Remote control not allowed; using loopback address
2005-09-25 23:46:09 [http://setiboinc.ssl.berkeley.edu/] Couldn't access URL http://setiboinc.ssl.berkeley.edu/.
The project's servers may be down,
in which case please try again later.
2005-09-25 23:46:09 [http://setiboinc.ssl.berkeley.edu/] Couldn't access URL http://setiboinc.ssl.berkeley.edu/.
The project's servers may be down,
in which case please try again later.
2005-09-25 23:46:09 [http://setiboinc.ssl.berkeley.edu/] Resetting project
2005-09-25 23:46:09 [---] request_reschedule_cpus: exit_tasks
2005-09-25 23:46:09 [http://setiboinc.ssl.berkeley.edu/] Detaching from project

----------------------------- END HERE -----------------------------------------

This is interesting. I am going to assume that this is not how it is supposed to work.

I appreciate all of the help.

Peace!
-TJ_

2005-09-25 19:13:05 [SETI@home] Deferring communication with project for 3 hours, 5 minutes, and 48 seconds
2005-09-25 19:13:05 [SETI@home] Deferring communication with project for 3 hours, 5 minutes, and 48 seconds

Is the name resolution OK ? - Can you nslookup setiboinc.ssl.berkeley.edu ?

Can you ping/traceroute to setiboinc.ssl.berkeley.edu ?

Is the connection to the servers via port 80 OK ? - Do you get a prompt from the command "telnet setiboinc.ssl.berkeley.edu 80" ?

It is obviously working he allready got 2 hostids assigned to his machine.
Reseting will make it even worse the time will increase.
I think there is no problem with the boinc client at all.

the next logical step would be to compile the seti app.
I would sugest using the fftw3 patched source from ned slider
(http://www.pperry.f2s.com/files/seti_boinc-client-fftw3-2004-12-01.tar.gz)

But first of all get the fftw3 source and configure it with
./configure --enable-fma --enable-float ( and --enable-altivec if your cpu supports it)
the --enable-fma awitch is important for ppc cpus.

regards
Crunch3r

---

Reseting will make it even worse the time will increase.

No.

Then i dont know what happened today with my alpha .. i´ve reseted the project
and in result of that it´s defering comunication with the project for two days now :-).

---

Join BOINC United now!

I wanted to get the basic package running before I make any alterations to the SRC to make it run more efficently. I also wanted to use the SRC to aide in porting it over to my MainFrame.

On which Mainframe do you wan't to port it ?

---

Hi, same problem here. Running on sparc64 debian 2.4.18. I compile seti client from source and using app_info.xml (no sure if well done). I let machine over a weekend and message repeat continously :-S

TIA

boinc_client -attach_project setiathome.berkeley.edu XXXXXXXXXXXXXXXXXXXXXX

2005-10-02 23:30:55 [---] Starting BOINC client version 4.72 for sparc64-unknown-linux-gnu
2005-10-02 23:30:55 [---] Data directory: /home/newneo/setiathome
2005-10-02 23:30:55 [---] Processor Inventory: 1 Processor(s)
2005-10-02 23:30:55 [---] Memory Inventory: Memory total - 495.59 MB, Swap total - 1.01 GB
2005-10-02 23:30:55 [---] Disk Inventory: Disk total - 443.88 TB, Disk available - 380.16 TB
2005-10-02 23:30:55 [http://setiathome.berkeley.edu/] Computer ID: not assigned yet; location: ; project prefs: default
2005-10-02 23:30:55 [---] No general preferences found - using BOINC defaults
2005-10-02 23:30:55 [---] Remote control not allowed; using loopback address
2005-10-02 23:30:57 [---] Running CPU benchmarks
2005-10-02 23:31:55 [---] Benchmark results:
2005-10-02 23:31:55 [---] Number of CPUs: 1
2005-10-02 23:31:55 [---] 290 double precision MIPS (Whetstone) per CPU
2005-10-02 23:31:55 [---] 599 integer MIPS (Dhrystone) per CPU
2005-10-02 23:31:55 [---] Finished CPU benchmarks
2005-10-02 23:31:55 [---] Resuming computation and network activity
2005-10-02 23:31:55 [---] request_reschedule_cpus: Resuming activities
2005-10-02 23:31:56 [http://setiathome.berkeley.edu/] Master page download succeeded
2005-10-02 23:31:57 [http://setiathome.berkeley.edu/] Sending scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi
2005-10-02 23:31:57 [http://setiathome.berkeley.edu/] Reason: Requested by user
2005-10-02 23:31:57 [http://setiathome.berkeley.edu/] Requesting 8640 seconds of work, returning 0 results
2005-10-02 23:31:58 [http://setiathome.berkeley.edu/] Scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi succeeded
2005-10-02 23:31:58 [SETI@home] Message from server: platform 'sparc64-unknown-linux-gnu' not found
2005-10-02 23:31:58 [SETI@home] New host venue: home
2005-10-02 23:31:59 [SETI@home] Deferring communication with project for 23 hours, 59 minutes, and 58 seconds
2005-10-02 23:31:59 [SETI@home] Deferring communication with project for 23 hours, 59 minutes, and 58 seconds

<app_info>
<app>
<name>setiathome</name>
</app>
<file_info>
<name>boinc_client</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome</app_name>
<version_num>472</version_num>
<file_ref>
<file_name>boinc_client</file_name>
<main_program/>
</file_ref>
</app_version>
</app_info>

---

Hi, same problem here.

It is a other problem....

2005-10-02 23:31:58 [http://setiathome.berkeley.edu/] Scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi succeeded
2005-10-02 23:31:58 [SETI@home] Message from server: platform 'sparc64-unknown-linux-gnu' not found
2005-10-02 23:31:58 [SETI@home] New host venue: home
2005-10-02 23:31:59 [SETI@home] Deferring communication with project for 23 hours, 59 minutes, and 58 seconds
2005-10-02 23:31:59 [SETI@home] Deferring communication with project for 23 hours, 59 minutes, and 58 seconds

Your boinc don't find the seti app. You must add the excatly name of the seti binary in app_info.xml under "file_info/name" and "app_version/app_name" and "file_ref/name"

<app_info>
<app>
<name>setiathome</name>
</app>
<file_info>
<name>boinc_client</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome</app_name>
<version_num>472</version_num>
<file_ref>
<file_name>boinc_client</file_name>
<main_program/>
</file_ref>
</app_version>
</app_info>

[/quote]
For a correct app_info.xml look at http://boinc.berkeley.edu/anonymous_platform.php

---

Thx Dotsch for replying,

I think that in my case maybe it's not the same architecture than TJ but it's exactly the same behaviour.

I already read http://boinc.berkeley.edu/anonymous_platform.php, my app_info.xml came from C+P from there ;-) plus some changes. As I didn't have very clear where to located, I copied it to every existing directory under setiathome folder.

The exact client binary name is there (boinc_client).

-rwxr-xr-x 1 root root 1095108 Sep 28 11:20 /usr/local/bin/boinc_client

Any other advices appreciated,

Thx

---

Thx Dotsch for replying,

I think that in my case maybe it's not the same architecture than TJ but it's exactly the same behaviour.

I already read http://boinc.berkeley.edu/anonymous_platform.php, my app_info.xml came from C+P from there ;-) plus some changes. As I didn't have very clear where to located, I copied it to every existing directory under setiathome folder.

The exact client binary name is there (boinc_client).

-rwxr-xr-x 1 root root 1095108 Sep 28 11:20 /usr/local/bin/boinc_client

Any other advices appreciated,

Thx

You need to put the binary name of the setiathome app. in the app_info.xml NOT the boinc client name!

then copy it into the directory /projects/setiathome.berkeley.edu/ in you boinc folder.

the app info should look similar to this one.
<app_info>
<app>
<name>setiathome</name>
</app>
<file_info>
<name>setiathome_4.7_alphaev67-unknown-linux-gnu</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome</app_name>
<version_num>470</version_num>
<file_ref>
<file_name>setiathome_4.7_alphaev67-unknown-linux-gnu</file_name>
<main_program/>
</file_ref>
</app_version>
</app_info>

exept you the to change the name of the setiathome app.

---

Join BOINC United now!

the next logical step would be to compile the seti app.
I would sugest using the fftw3 patched source from ned slider
(http://www.pperry.f2s.com/files/seti_boinc-client-fftw3-2004-12-01.tar.gz)

Tried this. The SAH_GRX_INCLUDES macro finds the path to X in a non-portable way:

AC_DEFUN([SAH_GRX_INCLUDES],[
AC_LANG_PUSH(C++)
AC_PATH_PROG(XLSFONTS,[xlsfonts])
sah_xpath=`echo $XLSFONTS | sed 's/bin\\/xlsfonts//' `
CFLAGS=`echo $CFLAGS "-I$sah_xpath/include"`
AC_CHECK_HEADERS([gl.h glu.h glut.h glaux.h GL/gl.h GL/glu.h GL/glut.h GL/glaux.h OpenGL/gl.h OpenGL/glu.h OpenGL/glut.h OpenGL/glaux.h GLUT/glut.h MesaGL/gl.h MesaGL/glu.h MesaGL/glut.h MesaGL/glaux.h])
AC_LANG_POP
])

On debian xlsfonts is in /usr/bin/X11/xlsfonts (X11 is a symlink to XFree86/bin, but /usr/bin/X11/ is what's in my path, so that's what XLSFONTS gets). See the problem? Autoconf has AC_PATH_X which might work better for this. Maybe this is fixed in your main branch?

---

the next logical step would be to compile the seti app.
I would sugest using the fftw3 patched source from ned slider
(http://www.pperry.f2s.com/files/seti_boinc-client-fftw3-2004-12-01.tar.gz)

Tried this. The SAH_GRX_INCLUDES macro finds the path to X in a non-portable way:

AC_DEFUN([SAH_GRX_INCLUDES],[
AC_LANG_PUSH(C++)
AC_PATH_PROG(XLSFONTS,[xlsfonts])
sah_xpath=`echo $XLSFONTS | sed 's/bin\\/xlsfonts//' `
CFLAGS=`echo $CFLAGS "-I$sah_xpath/include"`
AC_CHECK_HEADERS([gl.h glu.h glut.h glaux.h GL/gl.h GL/glu.h GL/glut.h GL/glaux.h OpenGL/gl.h OpenGL/glu.h OpenGL/glut.h OpenGL/glaux.h GLUT/glut.h MesaGL/gl.h MesaGL/glu.h MesaGL/glut.h MesaGL/glaux.h])
AC_LANG_POP
])

On debian xlsfonts is in /usr/bin/X11/xlsfonts (X11 is a symlink to XFree86/bin, but /usr/bin/X11/ is what's in my path, so that's what XLSFONTS gets). See the problem? Autoconf has AC_PATH_X which might work better for this. Maybe this is fixed in your main branch?

You dont need the X stuff just compile the client with "make -k all" you'll see it'll work. (the complaining about missing gl.h etc is "normal" nothing to worry about)

---

Join BOINC United now!

the next logical step would be to compile the seti app.
I would sugest using the fftw3 patched source from ned slider
(http://www.pperry.f2s.com/files/seti_boinc-client-fftw3-2004-12-01.tar.gz)

Tried this. The SAH_GRX_INCLUDES macro finds the path to X in a non-portable way:
...
On debian xlsfonts is in /usr/bin/X11/xlsfonts (X11 is a symlink to XFree86/bin, but /usr/bin/X11/ is what's in my path, so that's what XLSFONTS gets). See the problem? Autoconf has AC_PATH_X which might work better for this. Maybe this is fixed in your main branch?

You dont need the X stuff just compile the client with "make -k all" you'll see it'll work. (the complaining about missing gl.h etc is "normal" nothing to worry about)

The problem with having that macro was that it makes the configure script fail. I put in a workaround and the tried to build, but then I got this error:

main.cpp: In function `int main(int, char**)':
main.cpp:147: error: parse error before `e'
main.cpp:157: error: must have at least one catch per try block
main.cpp: At global scope:
main.cpp:164: error: parse error before `catch'
main.cpp:166: error: `e' was not declared in this scope
main.cpp:166: error: ISO C++ forbids declaration of `exit' with no type
main.cpp:166: error: `int exit' redeclared as different kind of symbol
/usr/include/stdlib.h:612: error: previous declaration of `void exit(int)'
main.cpp:167: error: parse error before `}' token
make[1]: *** [main.o] Error 1
make[1]: Leaving directory `/opt/magicsack/src/seti_boinc/client'
make: *** [all] Error 2

This is the line involved, seems like the definition of this exception wasn't included...
(main.cpp:147): catch(boinc_runtime_base_exception e) {
SETIERROR(e.ErrorValue(), e.what());
}

Which version of boinc is this supposed to compile against? I'm trying to use version 5.2.1.

the next logical step would be to compile the seti app.
I would sugest using the fftw3 patched source from ned slider
(http://www.pperry.f2s.com/files/seti_boinc-client-fftw3-2004-12-01.tar.gz)

Tried this. The SAH_GRX_INCLUDES macro finds the path to X in a non-portable way:
...
On debian xlsfonts is in /usr/bin/X11/xlsfonts (X11 is a symlink to XFree86/bin, but /usr/bin/X11/ is what's in my path, so that's what XLSFONTS gets). See the problem? Autoconf has AC_PATH_X which might work better for this. Maybe this is fixed in your main branch?

You dont need the X stuff just compile the client with "make -k all" you'll see it'll work. (the complaining about missing gl.h etc is "normal" nothing to worry about)

The problem with having that macro was that it makes the configure script fail. I put in a workaround and the tried to build, but then I got this error:

main.cpp: In function `int main(int, char**)':
main.cpp:147: error: parse error before `e'
main.cpp:157: error: must have at least one catch per try block
main.cpp: At global scope:
main.cpp:164: error: parse error before `catch'
main.cpp:166: error: `e' was not declared in this scope
main.cpp:166: error: ISO C++ forbids declaration of `exit' with no type
main.cpp:166: error: `int exit' redeclared as different kind of symbol
/usr/include/stdlib.h:612: error: previous declaration of `void exit(int)'
main.cpp:167: error: parse error before `}' token
make[1]: *** [main.o] Error 1
make[1]: Leaving directory `/opt/magicsack/src/seti_boinc/client'
make: *** [all] Error 2

This is the line involved, seems like the definition of this exception wasn't included...
(main.cpp:147): catch(boinc_runtime_base_exception e) {
SETIERROR(e.ErrorValue(), e.what());
}

Which version of boinc is this supposed to compile against? I'm trying to use version 5.2.1.

I will compile against 4.19 and 4.11 i think 4.11 is still available for download.

---

Join BOINC United now!

This is on a Multiprocessor PPC64 system running Linux 2.6.11-pseries

The Linux PPC binaries will also run succesfull on every other IBM Power Architecture. I have tested Stefan Urbats PPC binaries on a system with Power5 CPUs successfull.

---