[Edit] ____ thanks _ Pappa _ I just read your post over at the NC beta message boards ____ [/Edit]


[Edit] ____ Oops _ forgot to memtion _ I'm running _ Single _ user Instalation _ BOINC 5.8.15 _ on all three of my computers -- WXPSP2 ____ [/Edit]


Hi everyone!

I need to sk for assistance from SETI@home community :)


settings which affect caching for BOINC 5.8.15 any ideas ?


my three computers ...............

host_id_5302_

host_id_5303_

host_id_5304_

are crunching over at SETI@home/AstroPulse Beta testing _here_

four or five days ago I changed my connect to network to 0.5 days updated and then host 5303 and 5304 my two slowest of the three computers almost immediately DL six or seven WU for a small work cache but host 5302 my fastest of the three computer would not DL any cache even when I changed connect to network to 1 day and manually updated still no joy I looked in the message log of of host 5302 and it would say not requesting work .. all three host are set for Location home.

stdoutdae

2007-03-26 16:52:56 [SETI@home Beta Test] Sending scheduler request: Requested by user
2007-03-26 16:52:56 [SETI@home Beta Test] (not requesting new work or reporting completed tasks)
2007-03-26 16:53:01 [SETI@home Beta Test] Scheduler RPC succeeded [server version 509]
2007-03-26 16:53:01 [SETI@home Beta Test] Deferring communication for 7 sec
2007-03-26 16:53:01 [SETI@home Beta Test] Reason: requested by project



Josef W. Segur posted the following coments _here_

You're only running SETI Beta, so it can't be debt, I guess. If you've used the preference settings from the "simple GUI" mode of 5.8.15 there may be settings in global_prefs_override.xml which are keeping the web settings from working. And the cc_config.xml also has settings which affect caching. Perhaps someone else who is using 5.8.15 can suggest other possibilities.

 
                                                               Joe

"the cc_config.xml also has settings which affect cahing."

"there may be settings in global_prefs_override.xml which are keeping the web settings from working"

any ideas ?



here's a screen shot of my General preferences for all three of my computers ..... host 5302 ... host 5303 ... host 5304 ... all three set for home.




Byron

[Edit]

to my post below

that should read:

Hi everyone!

I need to _ask_ for assistance from the SETI@home community :)

Best Wishes
Byron


[/Edit]

Hi everyone!

I need to ask for assistance from SETI@home community :)

because my computer will not DL any cache


could I get your thoughts and Ideas on the following ..............

on the BOINC 5.8.15 GUI manager

1. I open on Advanced

2. scrolled down to and click __Read config file__

3. stdoutdar.txt ...... reads as follows ...

2007-03-26 21:07:40 [---] 4 active tasks
2007-03-26 21:07:40 [---] 13fe07af.21450.2935.3.5.237_0
2007-03-26 21:07:40 [---] 13fe07af.21480.3344.6.5.32_2
2007-03-26 21:07:40 [---] 13fe07af.21452.3344.4.5.26_0
2007-03-26 21:07:40 [---] 13fe07af.21471.3344.5.5.49_1
2007-03-26 21:07:45 [---] Unrecognized tag in cc_config.xml: <measurement_debug/>
2007-03-26 21:07:45 [---] [state_debug] set dirty: schedule_cpus
2007-03-26 21:07:45 [---] [status_debug] CLIENT_STATE::write_state_file(): Writing state file
2007-03-26 21:07:45 [---] [status_debug] CLIENT_STATE::write_state_file(): Done writing state file
2007-03-26 21:08:06 [SETI@home Beta Test] [task_debug] result 13fe07af.21471.3344.5.5.49_1 checkpointed

4. opened BOINC 5.8.15 folder

5. right click on cc_config.xml file

6. open with notepad

7. copied and paste the following ..... dose it look right to you ?


<cc_config>
<log_flags>
<task/>
<state_debug/>
<task_debug/>
<measurement_debug/>
</log_flags>
</cc_config>



dose that mean I need to edit out <measurement_debug/> ... in the cc_config.xml file ?


Best Wishes
Byron
:)

http://boinc.berkeley.edu/client_msgs.php

http://boincfaq.mundayweb.com/index.php?language=1&view=91

---

me@rescam.org

One of my computers stopped downloading any units and worked itself out completely. I finally detached and then reattached and everything is fine now. It was on one of my 2.8's.

---

Official Abuser of Boinc Buttons...
And no good credit hound!

Hi everyone!

I need to ask for assistance from SETI@home community :)

because my computer will not DL any cache


could I get your thoughts and Ideas on the following ..............

on the BOINC 5.8.15 GUI manager

1. I open on Advanced

2. scrolled down to and click __Read config file__

3. stdoutdar.txt ...... reads as follows ...

2007-03-26 21:07:40 [---] 4 active tasks
2007-03-26 21:07:40 [---] 13fe07af.21450.2935.3.5.237_0
2007-03-26 21:07:40 [---] 13fe07af.21480.3344.6.5.32_2
2007-03-26 21:07:40 [---] 13fe07af.21452.3344.4.5.26_0
2007-03-26 21:07:40 [---] 13fe07af.21471.3344.5.5.49_1
2007-03-26 21:07:45 [---] Unrecognized tag in cc_config.xml: <measurement_debug/>
2007-03-26 21:07:45 [---] [state_debug] set dirty: schedule_cpus
2007-03-26 21:07:45 [---] [status_debug] CLIENT_STATE::write_state_file(): Writing state file
2007-03-26 21:07:45 [---] [status_debug] CLIENT_STATE::write_state_file(): Done writing state file
2007-03-26 21:08:06 [SETI@home Beta Test] [task_debug] result 13fe07af.21471.3344.5.5.49_1 checkpointed

4. opened BOINC 5.8.15 folder

5. right click on cc_config.xml file

6. open with notepad

7. copied and paste the following ..... dose it look right to you ?


<cc_config>
<log_flags>
<task/>
<state_debug/>
<task_debug/>
<measurement_debug/>
</log_flags>
</cc_config>



dose that mean I need to edit out <measurement_debug/> ... in the cc_config.xml file ?


Best Wishes
Byron
:)

Yes, you should edit that line out of the cc_config.xml file, however, that is unrelated to the problem.

It is most likely that the host is either in EDF with sufficient work to meet the Connect Every X setting from other projects, OR S@H has a sufficiently negative long_term_debt to prevent work fetch, OR (and I believe that you have provided information to rule this one out) S@H is in EDF and therefore blocking work fetch from S@H.

---

BOINC WIKI

Eric J Korpela Project scientist posted the following:

Hi Byron,

I'll look for your machines in the logs to see if everything is OK on the server side.

Eric

Josef W. Segur posted the following:

You're only running SETI Beta, so it can't be debt, I guess. If you've used the preference settings from the "simple GUI" mode of 5.8.15 there may be settings in global_prefs_override.xml which are keeping the web settings from working. And the cc_config.xml also has settings which affect caching. Perhaps someone else who is using 5.8.15 can suggest other possibilities.

                                                                Joe

Ingleside posted the following:

measurement_debug was deprecated in v5.8.11, use benchmark_debug instead.

Usable for work-request-decisions are:
1
1

Some of the most common reasons for not asking for work are:
1; Already enough work on host, based on on_frac, active_frac, cpu_efficiency and estimated run-time based on wu's fpops_est, clients benchmark, and projects duration_correction_factor.
2; Obviously, if project set to "no new work". :)
3; Or, if network suspended or suspended due to time-of-day...

Also, some things blocks work-request from a project. Note, these are per project, one project being blocked doesn't mean other projects can't ask for work:
4; If cache-size is so large that project is in permanent deadline-trouble, blocked until project out of work, or idle cpu.
5; If atleast one result is suspended.
6; If atleast one download is being deferred.
7; If there's more uploads than 2 * ncpus.

#4, due to some wu's having only 4.3 days deadline, cache-sizes larger than 1.54 days can lead to project being blocked... For a Beta-project, it's not recommended to have a large cache anyway so...
#6, so a project there download-server is down won't block other projects from delivering work.
#7, to not get into a scenario there work is crunched faster than can be uploaded. Side-effect, also blocks work-requests when upload-server is down...

John McLeod VII posted the following:

Yes, you should edit that line out of the cc_config.xml file, however, that is unrelated to the problem.

It is most likely that the host is either in EDF with sufficient work to meet the Connect Every X setting from other projects, OR S@H has a sufficiently negative long_term_debt to prevent work fetch, OR (and I believe that you have provided information to rule this one out) S@H is in EDF and therefore blocking work fetch from S@H.

John , Ingleside , Eric , Joe or anyone :-)

Oops sorry my bad I think I may have found the problem why I can't DL any cache

% of time BOINC client is running 99.2201 %

While BOINC running, % of time work is allowed 0.0725 %

I don't know how that 0.0725 % happened ?

that 0.0725 % hasn't moved in four (4) days and I have been running that _computer 24/7_

I only run one project SETI@home/AstroPulse Beta

and I allow work to run 24/7 While BOINC running

any Ideas on how to fix _While BOINC running, % of time work is allowed 0.0725 %___ ?


for what it is worth:

Average CPU efficiency 0.982737

Result duration correction factor 0.655195

Eric J Korpela Project scientist posted the following:

Hi Byron,

I'll look for your machines in the logs to see if everything is OK on the server side.

Eric

Josef W. Segur posted the following:

You're only running SETI Beta, so it can't be debt, I guess. If you've used the preference settings from the "simple GUI" mode of 5.8.15 there may be settings in global_prefs_override.xml which are keeping the web settings from working. And the cc_config.xml also has settings which affect caching. Perhaps someone else who is using 5.8.15 can suggest other possibilities.

                                                                Joe

Ingleside posted the following:

measurement_debug was deprecated in v5.8.11, use benchmark_debug instead.

Usable for work-request-decisions are:
1
1

Some of the most common reasons for not asking for work are:
1; Already enough work on host, based on on_frac, active_frac, cpu_efficiency and estimated run-time based on wu's fpops_est, clients benchmark, and projects duration_correction_factor.
2; Obviously, if project set to "no new work". :)
3; Or, if network suspended or suspended due to time-of-day...

Also, some things blocks work-request from a project. Note, these are per project, one project being blocked doesn't mean other projects can't ask for work:
4; If cache-size is so large that project is in permanent deadline-trouble, blocked until project out of work, or idle cpu.
5; If atleast one result is suspended.
6; If atleast one download is being deferred.
7; If there's more uploads than 2 * ncpus.

#4, due to some wu's having only 4.3 days deadline, cache-sizes larger than 1.54 days can lead to project being blocked... For a Beta-project, it's not recommended to have a large cache anyway so...
#6, so a project there download-server is down won't block other projects from delivering work.
#7, to not get into a scenario there work is crunched faster than can be uploaded. Side-effect, also blocks work-requests when upload-server is down...

John McLeod VII posted the following:

Yes, you should edit that line out of the cc_config.xml file, however, that is unrelated to the problem.

It is most likely that the host is either in EDF with sufficient work to meet the Connect Every X setting from other projects, OR S@H has a sufficiently negative long_term_debt to prevent work fetch, OR (and I believe that you have provided information to rule this one out) S@H is in EDF and therefore blocking work fetch from S@H.

John , Ingleside , Eric , Joe or anyone :-)

Oops sorry my bad I think I may have found the problem why I can't DL any cache

% of time BOINC client is running 99.2201 %

While BOINC running, % of time work is allowed 0.0725 %

I don't know how that 0.0725 % happened ?

that 0.0725 % hasn't moved in four (4) days and I have been running that _computer 24/7_

I only run one project SETI@home/AstroPulse Beta

and I allow work to run 24/7 While BOINC running

any Ideas on how to fix _While BOINC running, % of time work is allowed 0.0725 %___ ?


for what it is worth:

Average CPU efficiency 0.982737

Result duration correction factor 0.655195

Yes, that would prevent you from getting much work.

What is your setting for "Do work while computer is active."? If this is no, try yes.

You can edit the client_state.xml to change this number directly (the max is 1.0). To do so:
1) Stop BOINC
2) Open client_state.xml.
3) Change the relevant value in <time_stats> which is, I believe <active_frac> to 0.99 or 1.00.
4) Save the change.
5) Restart BOINC.

[EDIT]

Also check the settings to make certain that nothing is suspended.

---

BOINC WIKI

Eric J Korpela Project scientist posted the following:

Hi Byron,

I'll look for your machines in the logs to see if everything is OK on the server side.

Eric

Josef W. Segur posted the following:

You're only running SETI Beta, so it can't be debt, I guess. If you've used the preference settings from the "simple GUI" mode of 5.8.15 there may be settings in global_prefs_override.xml which are keeping the web settings from working. And the cc_config.xml also has settings which affect caching. Perhaps someone else who is using 5.8.15 can suggest other possibilities.

                                                                Joe

Ingleside posted the following:

measurement_debug was deprecated in v5.8.11, use benchmark_debug instead.

Usable for work-request-decisions are:
1
1

Some of the most common reasons for not asking for work are:
1; Already enough work on host, based on on_frac, active_frac, cpu_efficiency and estimated run-time based on wu's fpops_est, clients benchmark, and projects duration_correction_factor.
2; Obviously, if project set to "no new work". :)
3; Or, if network suspended or suspended due to time-of-day...

Also, some things blocks work-request from a project. Note, these are per project, one project being blocked doesn't mean other projects can't ask for work:
4; If cache-size is so large that project is in permanent deadline-trouble, blocked until project out of work, or idle cpu.
5; If atleast one result is suspended.
6; If atleast one download is being deferred.
7; If there's more uploads than 2 * ncpus.

#4, due to some wu's having only 4.3 days deadline, cache-sizes larger than 1.54 days can lead to project being blocked... For a Beta-project, it's not recommended to have a large cache anyway so...
#6, so a project there download-server is down won't block other projects from delivering work.
#7, to not get into a scenario there work is crunched faster than can be uploaded. Side-effect, also blocks work-requests when upload-server is down...

John McLeod VII posted the following:

Yes, you should edit that line out of the cc_config.xml file, however, that is unrelated to the problem.

It is most likely that the host is either in EDF with sufficient work to meet the Connect Every X setting from other projects, OR S@H has a sufficiently negative long_term_debt to prevent work fetch, OR (and I believe that you have provided information to rule this one out) S@H is in EDF and therefore blocking work fetch from S@H.

John , Ingleside , Eric , Joe or anyone :-)

Oops sorry my bad I think I may have found the problem why I can't DL any cache

% of time BOINC client is running 99.2201 %

While BOINC running, % of time work is allowed 0.0725 %

I don't know how that 0.0725 % happened ?

that 0.0725 % hasn't moved in four (4) days and I have been running that _computer 24/7_

I only run one project SETI@home/AstroPulse Beta

and I allow work to run 24/7 While BOINC running

any Ideas on how to fix _While BOINC running, % of time work is allowed 0.0725 %___ ?


for what it is worth:

Average CPU efficiency 0.982737

Result duration correction factor 0.655195

Yes, that would prevent you from getting much work.

What is your setting for "Do work while computer is active."? If this is no, try yes.

You can edit the client_state.xml to change this number directly (the max is 1.0). To do so:
1) Stop BOINC
2) Open client_state.xml.
3) Change the relevant value in <time_stats> which is, I believe <active_frac> to 0.99 or 1.00.
4) Save the change.
5) Restart BOINC.

[EDIT]

Also check the settings to make certain that nothing is suspended.

Hi John

thanks a million that work beautifully ! I Just Dl four (4) Beta WUs
I now have a Cache of Four Beta WUS to to see me through tomorrow's regular scheduled outage :-)

Best Wishes
Byron