[Knightrider]

I'm using BOINC Gui V 5.9.10 and I tried changing the app_info file but the only result was permission errors and having to re-install BOINC over the top. When I sorted that out, the error message was still there. The only cure, for me anyway, was to detach from sah and re-attach. This of course, clears the project folder and new is installed, along with the standard seti worker.

SAH downloads are very slow, but at least its started for me. I notice that the run time for the majority of wu's is over 14 hours.

I will not try the enhanced 7.1 yet. viz the permissions problem.

K.

[zombie67]

Odd. I added the following line:

<version_num>513</version_num> ( <---- right after this line)
<platform>i686-apple-darwin</platform>

And it worked just fine. No problemo.

Now, if only the downloads could please keep up with my dual quad....

[UL1]

Looks like it's not running as it should over there at SETI...
I was running 5.9.10 (GUI) too and my iMac had finally made it to download some WUs. I was really happy and then started to install Alex' 7.2...with the same result: permission errors. After I fixed that one without re-installing BOINC I realized that all running calculations and all formerly downloaded and ready-to-use WUs had gone away.
So I re-installed the 5.4.9 (GUI), waited for new WUs and then installed the optimized client again. This time it did work without any probs. But I can't upload any results...
And my Pro hasn't made it to download any received WUs yet...

Edit: one (!) result has made it to SETI now...

[zombie67]

Don't confuse your permission problems with S@H upload/download issues.

The former can be solved with this:

(quit BOINC)
(in a terminal copy/paste the following two lines)
cd /Library/Application\ Support/BOINC\ Data/projects/setiathome.berkeley.edu/
sudo chown boinc_master:boinc_project *
(restart BOINC)

The latter is a real problem. They have a DL limit of 100/thread or 400, whichever comes first....and I've hit the wall.

That is not going to help with the challenge from who?

[Echrei]

Now I'm not getting work from the server because of another error, this time it's saying "incomplete request"

[zombie67]

Originally Posted by Echrei 

Now I'm not getting work from the server because of another error, this time it's saying "incomplete request"

No to worry. No one is downloading anything from S@H now. Give it another day or two.

[Lycanthrope]

Hi Chaps - my first post...

I came over here via the message boards on S@H - I've downloaded the Xeon optimized client i386v7 and have copied the files into the correct directory under Library/Application Support etc.

But I can't find anywhere the instructions on how to activate this client from the BOINC manager - surely there's some preference, parameter or config that should be changed?

Cheers
Dave

[zombie67]

Nothing else to do. The app_onfo file tells BOINC to use the optimized binary instead.

[Lycanthrope]

Originally Posted by zombie67 

Nothing else to do. The app_onfo file tells BOINC to use the optimized binary instead.

Ah well, that would explain quite a lot then, cheers!

And it used this new app also for Einstein or just for SETI?

Sorry for the dumb questions...

[zombie67]

Just for SETI. Each project has its own folder, and its own application.

[Lycanthrope]

Originally Posted by zombie67 

Just for SETI. Each project has its own folder, and its own application.

Yes, of course, I also realized this myself about 1 minute ago also.

I told you it was a dumb question

Now I just need some WU's to test out the SPEEEEED - I also updated the client on my G5 iMac so there's double the anticipation.

Thanks for your help

[zombie67]

Someone discovered a work-around to get WUs downloaded.

1) quit BOINC

2) rename app_info.xml to something else

3) start BOINC. Downloads will work now. Download as much as possible.

4) quit BOINC

5) rename app_info.xml back to app_info.xml

6) start BOINC

This probably means you will need to go through this every time you drain your queue. At least until S@H fixes the problem.

[arkayn]

So what is it looking at when it sees that file?

[zombie67]

Fri May 18 08:44:29 2007|SETI@home|Found app_info.xml; using anonymous platform

When there is an app_info.xml file, BOINC sends "anonymous platform" to the server. Or something like that. Apparently, S@H servers are having problems with "anonymous platform".

But we must have an app_info.xml in order to use alexkan's optimized app. It's what tells BOINC to use his app instead of the stock app.

So by renaming app_info.xml, it allows work download, and then by putting it back, allows optimized crunching.

[QSilver]

Originally Posted by zombie67 

Fri May 18 08:44:29 2007|SETI@home|Found app_info.xml; using anonymous platform

So by renaming app_info.xml, it allows work download, and then by putting it back, allows optimized crunching.

Worked for me. Thanks for the tip, zombie!

QS

[kennet]

I followed the advice of zombie67 and made an extraordinary discovery!

I renamed app_info.xml as 'Alex' so that it would not be identified as per
zombie's instructions. This was done after midnight Friday Pacific time
(Canada's West Coast here): so there would have been no one at SETI
working. It downloaded a couple of workunits and I went to bed with
out reverting 'Alex' back to app_info.xml.

In the morning I had my normal cache of work units and completed
work units were properly uploaded.

I doubled checked and the "Show graphics" icon is faded and not useable.

I checked the Work Unit times and they are as they were pre-thumper crash.

Kenn

Originally Posted by zombie67 

Fri May 18 08:44:29 2007|SETI@home|Found app_info.xml; using anonymous platform

When there is an app_info.xml file, BOINC sends "anonymous platform" to the server. Or something like that. Apparently, S@H servers are having problems with "anonymous platform".

But we must have an app_info.xml in order to use alexkan's optimized app. It's what tells BOINC to use his app instead of the stock app.

So by renaming app_info.xml, it allows work download, and then by putting it back, allows optimized crunching.

[kennet]

Just to be clear my message queue does not have (as it did previously)
"anonymous platform" line inserted as shown per below:

Sat 19 May 04:34:32 2007||Starting BOINC client version 5.8.17 for i686-apple-darwin
Sat 19 May 04:34:32 2007||log flags: task, file_xfer, sched_ops
Sat 19 May 04:34:32 2007||Libraries: libcurl/7.15.5 OpenSSL/0.9.7l zlib/1.2.3
Sat 19 May 04:34:32 2007||Data directory: /Library/Application Support/BOINC Data
Sat 19 May 04:34:32 2007||Processor: 2 GenuineIntel Intel(R) Core(TM)2 CPU T7200 @ 2.00GHz [x86 Family 6 Model 15 Stepping 6] [FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM SSE3 MON DSCPL VMX EST TM2 MNI CX16 TPR]
Sat 19 May 04:34:32 2007||Memory: 1.00 GB physical, 127.41 GB virtual
Sat 19 May 04:34:32 2007||Disk: 148.73 GB total, 127.17 GB free
Sat 19 May 04:34:32 2007|SETI@home|URL: http://setiathome.berkeley.edu/; Computer ID: 3378085; location: home; project prefs: default
Sat 19 May 04:34:32 2007||General prefs: from SETI@home (last modified 2007-05-19 00:54:47)

[zombie67]

You may be crunching with the default application, not the optimized application. Make sure your app_info.xml file name is spelled correctly.

[arkayn]

I have found that even though I don't have my app_info named correctly, it is still using Alex's cruncher instead of the default application. It never downloaded one from the server.


Now the 2 PC's did download the default cruncher. I think part of that is because ours is named closer to the default app.

[kennet]

Originally Posted by zombie67 

You may be crunching with the default application, not the optimized application. Make sure your app_info.xml file name is spelled correctly.

Please read the originating email, you will find that I am not crunching with the default.

The telling evidence is the time it takes crunching the Work Units. I am back to the times
pre-thumper crash and the no show graphics button is inactive. As well you can see the
contents of my projects file below:

16ja05ab.25838.592.647150.3.251
16ja05ab.25838.1152.940902.3.39
16ja05ab.25838.1394.304840.3.151
16ja05ab.25838.2401.642322.3.190
18mr05aa.11342.273.792316.3.213
18mr05aa.11342.321.454834.3.193
18mr05aa.11342.1808.603396.3.12
18mr05aa.11342.3042.404844.3.30
18mr05aa.11342.3280.840898.3.187
18mr05aa.11342.3426.878414.3.103
18mr05aa.11342.4192.797144.3.57
18mr05aa.11342.4386.498588.3.218
27fe05aa.20439.4752.1028384.3.133
27fe05aa.20439.4848.372156.3.50
27fe05aa.20439.4848.372156.3.50_1_0
27fe05aa.20439.5042.79836.3.10
27fe05aa.20439.5042.79836.3.10_2_0
27fe05aa.20439.6114.1028384.3.10
27fe05aa.20439.7122.323574.3.31
27fe05aa.20439.7456.115896.3.9
27fe05aa.20439.8081.1015886.3.124
27fe05aa.20439.8800.228398.3.141
27fe05aa.20439.9521.454834.3.25
27fe05aa.20439.10433.379824.3.212
27fe05aa.20439.11297.642322.3.113
27fe05aa.20439.12304.965922.3.153
27fe05aa.20439.13648.1028410.3.119
27fe05aa.20439.13696.703400.3.76
alex
arecibo_181.png
AUTHORS
COPYING
COPYRIGHT
EXTRA_NOTE
README
sah_40.png
sah_banner_290.png
sah_ss_290.png
seti_enhanced-i386-v7.2-core2-nographics
slideshow_setiathome_enhanced_00
slideshow_setiathome_enhanced_01
slideshow_setiathome_enhanced_02
stat_icon

[Echrei]

Now I'm getting:
Scheduler request failed: server returned nothing (no headers, no data)

[zombie67]

Originally Posted by Echrei 

Now I'm getting:
Scheduler request failed: server returned nothing (no headers, no data)

I believe that is message just means that the server is overloaded.

[zombie67]

Originally Posted by kennet 

Please read the originating email, you will find that I am not crunching with the default.

The telling evidence is the time it takes crunching the Work Units. I am back to the times
pre-thumper crash and the no show graphics button is inactive. As well you can see the
contents of my projects file below:

I'm not quite sure I understand your question then. What are you trying to solve for?

Yes, even if you don't put back the app_info.xml, the WUs you had already downloaded will continue to crunch with the optimized application. Those WUs are already associated with the optimized application in the "client_state.xml" file. But the new WUs you download won't know about the optimized application until you put back the app_info.xml.

[arkayn]

Mine is downloading new units and crunching with seti_enhanced-i386-v7.2-core2-nographics even without the app_info.xml loading.

[kennet]

There is no question: it is a statement. All workunits that are downloaded are being crunched by the optimized client, even the newly downloaded Work Units. I thought that I was being very clear. I will try to be clearer.

Uploads and downloads are proceeding normally and are being processed normally with Alex's client.

Over eighty Work Units have been or are being processed by the optimized client. Rest assured that
I did not start with eighty or more work units downloaded. After renaming app_info.xml to Alex, my
client downloaded an initial total of eight work units. All other work units were downloaded after the
first two initially downloaded work units were uploaded.

I have not reverted back to the app_info.xml name and the optimized client made by Alex is still
processing all downloaded work units, no matter when they were downloaded. All crunch times
are normal as per pre-thumper crash.

In other words, everything is proceeding as normal, with no delays or refusals to work.

Even arkayn has confirmed the same process with his work units.

[zombie67]

Yes, some people are finding that you don't have to switch back. Others have had to, in order to force usage of the optimized app.

You can read about it here: Using optimized apps without app_info.xml

I am putting it back after downloading, just in case.

[arkayn]

Now if we can only figure out what is wrong with the server this time!!

[zombie67]

)(*&!)(*^#!

[arkayn]

Originally Posted by zombie67 

)(*&!)(*^#!

Exactly, at least I downloaded a lot of work last night and this morning so my iMac will not be idle. The eMac on the other hand should run out sometime tomorrow.

[zombie67]

...and now S@H is completely down, regardless of what the server status page says. Can't even get to the forum pages.

A tip about maximizing queue size to weather outages like this:

The BOINC scheduling logic is messed up. It is ultra conservative. Assuming you crunch for a single project, the formula for the maximum queue is roughly shortest deadline - 3 days. Since S@H has some WUs with a deadline of 4 days, the maximum queue you can get (if you get any of those WUs) is about a day.

With BOINC 5.10.* (5.9.* is the test version of the same thing), there is a new setting you can make at the project level (web page) or client level (menu). It's "Maintain enough work for an additional [1-10] days. So you can set your "connect every" to something like .1 days, and then add on an additional (say) 5 days. This way you will always have at least 5 days of queue. The potential downside is missed deadlines. For example, you might download 5 days worth of 4 day WUs. Unlikely, but the possibility exists. So if you use this method, keep an eye on your queue.

Anyway, I had installed 5.9.11. And using the settings above, I currently still have about 2-3 days of queue (~600 WUs), which was no easy feat. S@H will let you download only 100 WUs/thread/day, with a max of 400/day, regardless of the number of threads your machine has. Mine will churn through about 200-300/day. So it takes at least 3 days just to build up a 3 day queue. (annoyed grunt)

[arkayn]

From the front page.

May 20, 2007
We've been down since about 3pm Pacific because of further server problems.

[arkayn]

Originally Posted by zombie67 

...and now S@H is completely down, regardless of what the server status page says. Can't even get to the forum pages.

A tip about maximizing queue size to weather outages like this:

The BOINC scheduling logic is messed up. It is ultra conservative. Assuming you crunch for a single project, the formula for the maximum queue is roughly shortest deadline - 3 days. Since S@H has some WUs with a deadline of 4 days, the maximum queue you can get (if you get any of those WUs) is about a day.

With BOINC 5.10.* (5.9.* is the test version of the same thing), there is a new setting you can make at the project level (web page) or client level (menu). It's "Maintain enough work for an additional [1-10] days. So you can set your "connect every" to something like .1 days, and then add on an additional (say) 5 days. This way you will always have at least 5 days of queue. The potential downside is missed deadlines. For example, you might download 5 days worth of 4 day WUs. Unlikely, but the possibility exists. So if you use this method, keep an eye on your queue.

Anyway, I had installed 5.9.11. And using the settings above, I currently still have about 2-3 days of queue (~600 WUs), which was no easy feat. S@H will let you download only 100 WUs/thread/day, with a max of 400/day, regardless of the number of threads your machine has. Mine will churn through about 200-300/day. So it takes at least 3 days just to build up a 3 day queue. (annoyed grunt)

I have 5.9.11 installed on 4 machines now, it only took Opera to download the 2 installers and me clicking on continue every time the download died.

[QSilver]

Originally Posted by zombie67 

Yes, some people are finding that you don't have to switch back. Others have had to, in order to force usage of the optimized app.

You can read about it here: Using optimized apps without app_info.xml

I am putting it back after downloading, just in case.

I was able to get this to work with my G4 iMac (10.4.9). I put the app_info.xml file back after getting an initial set of new WUs. I haven't determined yet whether it will get more WUs.

I have not been able to get this to work with my PC. When I put back the app_info.xml file, it erases all of the S@H WUs and can;t get more. I guess my options are to let it run without app_info and see if still uses the optimized app or review the app_info file line-by-line to see if I can determine why it's acting this way.

QS

[arkayn]

Weird one here, one minute the iMac is crunching and the next thing I I get this.

2007-05-22 01:36:41 [SETI@home] [error] Application file seti_enhanced-i386-v7.2-core2-nographics missing signature
2007-05-22 01:36:41 [SETI@home] [error] BOINC cannot accept this file
2007-05-22 01:36:41 [SETI@home] Deferring communication for 1 min 0 sec
2007-05-22 01:36:41 [SETI@home] Reason: Unrecoverable error for result 16ja05ab.25838.14656.322166.3.126_0 (Can't create shared memory: system shmat)
2007-05-22 01:36:42 [SETI@home] Deferring communication for 1 min 0 sec
2007-05-22 01:36:42 [SETI@home] Reason: Unrecoverable error for result 11fe05aa.10139.22336.390890.3.204_0 (app_version download error: couldn't get input files:
<file_xfer_error>
<file_name>seti_enhanced-i386-v7.2-core2-nographics</file_name>
<error_code>-123</error_code>
<error_message>missing signature</error_message>
</file_xfer_error>

[kennet]

Originally Posted by QSilver 

I was able to get this to work with my G4 iMac (10.4.9). I put the app_info.xml file back after getting an initial set of new WUs. I haven't determined yet whether it will get more WUs.

I have not been able to get this to work with my PC. When I put back the app_info.xml file, it erases all of the S@H WUs and can;t get more. I guess my options are to let it run without app_info and see if still uses the optimized app or review the app_info file line-by-line to see if I can determine why it's acting this way.

QS

Quit Boinc Manager and then rename the app_info.xml (in my case to Alex). Leave the optimized app and renamed app_info.xml in the projects: setiathome.berkeley.edu file and restart the boinc Manager.

This is what I have done and I am uploading and downloading and have the same crunch times as I had previous to thumper's demise.

Kenn

Here is the list of contents minus the work units in my own setiathome.berkeley.edu file:

Alex
arecibo_181.png
AUTHORS
COPYING
COPYRIGHT
EXTRA_NOTE
README copy
sah_40.png
sah_banner_290.png
sah_ss_290.png
seti_enhanced-i386-v7.2-core2-nographics
slideshow_setiathome_enhanced_00
slideshow_setiathome_enhanced_01
slideshow_setiathome_enhanced_02
stat_icon

[kennet]

Originally Posted by kennet 

Quit Boinc Manager and then rename the app_info.xml (in my case to Alex). Leave the optimized app and renamed app_info.xml in the projects: setiathome.berkeley.edu file and restart the boinc Manager. stat_icon

Here my list of results from today. {Note headers removed for readability}.

537692178 130272665 23 May 2007 1:19:14 UTC 30 May 2007 22:22:37 UTC In Progress Unknown New --- --- ---
537596987 130244456 22 May 2007 16:03:12 UTC 23 May 2007 2:19:45 UTC Over Success Done 21.48 0.08 pending
537436170 130195032 22 May 2007 16:50:42 UTC 22 May 2007 22:44:38 UTC Over Success Done 3,801.50 52.52 pending
537405020 130185255 22 May 2007 16:24:13 UTC 22 May 2007 22:44:38 UTC Over Success Done 4,674.71 62.36 62.36
537395287 130182253 22 May 2007 7:21:24 UTC 22 May 2007 23:19:14 UTC Over Success Done 12,789.72 12.52 12.52
537395285 130182257 22 May 2007 7:21:24 UTC 22 May 2007 23:19:14 UTC Over Success Done 110.29 0.04 0.04
537375076 130175916 22 May 2007 16:03:33 UTC 23 May 2007 1:19:14 UTC Over Success Done 4,797.33 64.48 64.48
537359491 130171124 22 May 2007 5:58:08 UTC 12 Jun 2007 15:12:56 UTC In Progress Unknown New --- --- ---
537329618 130161715 22 May 2007 15:17:08 UTC 22 May 2007 22:44:38 UTC Over Success Done 3,973.59 52.52 pending
537327520 130161033 22 May 2007 15:12:32 UTC 23 May 2007 2:19:45 UTC Over Success Done 3,998.32 52.83 52.83
537321960 130159282 22 May 2007 15:04:49 UTC 22 May 2007 22:44:38 UTC Over Success Done 7.78 0.05 0.05
537321448 130159109 22 May 2007 15:03:48 UTC 23 May 2007 2:19:45 UTC Over Success Done 2,887.85 33.71 pending
537313084 130156538 22 May 2007 14:50:57 UTC 23 May 2007 1:19:14 UTC Over Success Done 19.58 0.08 0.08
537297226 130151388 22 May 2007 14:25:46 UTC 23 May 2007 1:19:14 UTC Over Success Done 2,784.93 31.55 31.55
537289084 130148779 22 May 2007 14:11:33 UTC 23 May 2007 2:19:45 UTC Over Success Done 18.93 0.08 0.08
537286566 130147989 22 May 2007 14:06:45 UTC 23 May 2007 1:19:14 UTC Over Success Done 4,302.63 57.97 57.97
537272256 130143641 22 May 2007 13:42:42 UTC 22 May 2007 22:44:38 UTC Over Success Done 2,717.66 30.59 pending
537256006 130137613 22 May 2007 13:15:02 UTC 22 May 2007 16:24:13 UTC Over Success Done 3,815.97 49.92 pending
537251008 130139067 22 May 2007 13:06:14 UTC 22 May 2007 15:50:30 UTC Over Success Done 933.68 14.09 pending
537248804 130136990 22 May 2007 13:02:31 UTC 22 May 2007 16:23:08 UTC Over Success Done 3,895.11 49.96 49.96

[Echrei]

I had to do this to get my G4 and G5 to start downloading work again, else I got an HTTP error.
1. Remove app_info.xml - BOINC downloads the non-optimized worker and work
2. Stop BOINC
3. Put the app_info.xml back in - The non-optimized worker is deleted and BOINC uses the optimized one
4. Stop BOINC
5. Remove the app_info.xml again and start BOINC

This worked for both the CLI and GUI versions of BOINC 5.9.11

[zombie67]

It's back up and working again, even with app_info.xml.

[Billy]

Yes, it is working again, finally. I did notice that I apparently have a second computer listed which is actually the same as the only one I have. Perhaps this is related to the problems there and changing the app.info file? Also, the merge computer command didn't seem to be working.

[zombie67]

Did you upgrade BOINC recently? It changed (again) the way that CPU information is reported. And that can sometimes cause a new host to be created on the projects.

[Billy]

I upgraded about a week or so ago to 5.9.11, nothing this week. Oh, I just got another computer on SETI.

[UL1]

Everything was fine for a couple of days...but today I found this message on my iMac:
"Message from server: Can't find host record Invalid or missing account key. Visit this project's web site to get an account key."
Any idea what's wrong and/or what to do ?

Edit: the MacPro is also affected...

[zombie67]

Yeah, I think I jinxed it. Everyone is having the problem, so it's on the server side.

[arkayn]

Looks like they have everything shut down for maintenance at the moment.

[zombie67]

Wow. Messing with things just before a long weekend. Not smart.

Glad I am using the new 5.9 feature to add an additional 5 days of queue.

[ChillyWilly5280]

Originally Posted by zombie67 

Glad I am using the new 5.9 feature to add an additional 5 days of queue.

Is 5.9.12 stable?

[zombie67]

It is for me. I get a reoccurring error message that seems to have no impact on actual crunching. A bug has been logged.

Fri May 25 18:29:00 2007||Can't rename client_state_next.xml to client_state.xml; check file and directory permissions
Fri May 25 18:29:00 2007||[error] Couldn't write state file: system rename

[arkayn]

Originally Posted by ChillyWilly5280 

Is 5.9.12 stable?

I tried 5.9.12 on my laptop and it just kept locking up the boinc manager. I ended up reinstalling 5.9.11 which is working great.

May 25, 2007
The database and project have been down for a few hours. We're working on the slow feeder problem that was causing work distribution to go slowly.

[Echrei]

I'm running 5.9.12 on my G4 and G5 and both are working fine.

[ChillyWilly5280]

Strange, 5.9.12 seems to have disabled running BOINC as a daemon or system service on both machines I've installed it on so far (iMac G4 800 & 1.25, 10.4.9). Anyone know about this, or have suggestions?