Hi all,

I got a message on rosetta with boinc osx, I can't return more than 2 results per day for each computer ?
Is that true ? They ask to calculate as much as possible for 2 weeks and add a limite for 2 results per day ?

Isn't there something wrong here ?
Phil

Ven 21 jul 12:18:31 2006|rosetta@home|Sending scheduler request to https://boinc.bakerlab.org/rosetta_cgi/cgi
Ven 21 jul 12:18:31 2006|rosetta@home|Reason: Requested by user
Ven 21 jul 12:18:31 2006|rosetta@home|Requesting 172800 seconds of new work, and reporting 1 completed tasks
Ven 21 jul 12:18:36 2006|rosetta@home|Scheduler request succeeded
Ven 21 jul 12:18:36 2006|rosetta@home|Message from server: No work sent
Ven 21 jul 12:18:36 2006|rosetta@home|Message from server: (reached daily quota of 2 results)

Hi Phil,

it looks like the machine in question does not work properly and thus does not get too many WUs/day any more. You should have a look at it or post some more details if you have trouble understanding what happened (no offence meant, it just often happens that you have a networking problem but do not have the slightest notion where it comes from).

So Rosetta is of course interest in any result you can get, but BOINC has some measures to prevent unnecessary network loads. Here are some details the "daily quota" that may help you to understand what is going on:

The "daily quota means the maximum number of results a specific computer will get per day. The default is 100 and it only changes if your machine does not return valid results or lets them time out before they are sent back. In other words your daily quota is reduced by one WU/day as soon as a results errors out or is not returned in time. If you return valid rersults during the time that happens, the quota is increased again.

This is done to prevent PCs with severe errors from downloading multiple results which then all error out. Many such PCs would consume a lot of bandwidth an resources for Rosetta AND the user without doing any valuable work which both will not want.

This also accounts for machines that can download WUs but (probably by some port or firewall setting) not upload the results. If the result times out before it is received the quota is reduced. The machine will finally receive no more work (or rather just 1 WU/day).

With each result resent the quota doubles, so you can quickly recover after a problem (after returning 5 valid results your quota is back to 32 per day).

(By the way, if you hide your computers in order not to show your IP addresses and computer names: this info is visible only to you, anybody else will only get the host numbers. So there is no threat from that whatever. But it is of course up to you to do so.)

Regards,

Christoph

Hi Christoph,

you're right, all last results on this computer exit with client error 5x05 and the quota is result 1/day

I'll check why on the computer itself, but maybe you can help me to find why ?

Thanks in advance
Phil

ps: after quick check the file /usr/lib/libstdc++.6.dylib doesn't exist on my OS 10.3 server...



Result ID 29407379
Name FRA_t374_CASP7_hom001_4_t374_4_1vhsA_IGNORE_THE_REST_110_1015_9_0
Workunit 25231544
Created 21 Jul 2006 7:57:25 UTC
Sent 21 Jul 2006 7:59:12 UTC
Received 21 Jul 2006 10:18:33 UTC
Server state Over
Outcome Client error
Client state Computing
Exit status 5 (0x5)
Computer ID 93938
Report deadline 28 Jul 2006 7:59:12 UTC
CPU time 0
stderr out

<core_client_version>5.4.9</core_client_version>
<message>
process got signal 5
</message>
<stderr_txt>
dyld: rosetta_5.25_powerpc-apple-darwin can't open library: /usr/lib/libstdc++.6.dylib (No such file or directory, errno = 2)

</stderr_txt>

Validate state Invalid
Claimed credit 0
Granted credit 0
application version 5.25

---

Hi Phil,

I am sorry, but I am not familiar with Macs, so I do not know if a damaged or missing lib file is the cause or if the path to the lib file is just assigned in the wrong way:

can't open library: /usr/lib/libstdc++.6.dylib

But I would interpret "libstdc++" as a "standard C++ library" that is missing, faulty, or misassigned on your machine.

You might try to reinstall that library. And maybe you should also run a diskcheck to see if the sector in question has been damaged in some way.

If this does not help, maybe somebody here familiar with Macs knows what to do(?). Any comments?

Regards,

Christoph

I would expect the science client, (and the BOINC core of course), to be statically linked, and thus have no need for external libraries. Consider, not everyone has a C++ compiler on their machine.

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

I would expect the science client, (and the BOINC core of course), to be statically linked, and thus have no need for external libraries. Consider, not everyone has a C++ compiler on their machine.

So would I, but apparently, it is not true that everything is built with static libs for all platforms.

---

BOINC WIKI

Boinc crashed (all 60 qeued WU's suddenly were on 100%)
After restarting I got the same message:
Maximum daily WU quota per CPU 1/day

With a average of 35 WU's a day I don't think I'm glad with that ;)

But I'll see what happens if he is done with that 1 WU :)

---

With a average of 35 WU's a day I don't think I'm glad with that ;)

Brabant, it looks like one of your hosts racked up a pile of WU errors (due to that crash I suppose), here is a link. This may be why it locked down the one host to 1 per day, and leaves the other at 100.

Hopefully as you successfully crunch WUs on that host, it will increase your quota.

You might consider increasing your WU runtime preference to 24hrs to make that one last you for the day. Then you'd only need one per day to keep you fully busy. Well, 2 I guess, because you have a dual core on that host. That would be a way to remain fully busy crunching Rosetta while your wait for your quota to get back to normal. The runtime preference will take effect for any WUs on deck at the time you update to the project.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

Boinc crashed (all 60 qeued WU's suddenly were on 100%)
After restarting I got the same message:
Maximum daily WU quota per CPU 1/day

With a average of 35 WU's a day I don't think I'm glad with that ;)

But I'll see what happens if he is done with that 1 WU :)

Is it warm where you are? Might want to looking into overheating. Something has caused a failure of those Wus. To stop runaway hosts a quota of 100/day was introduced. It reduces by ONE for each error reported. When you start returning "good" results the quota will double for each good WU you return, up to 100/day. If you return nothing but good WUS you'll see your quota change from 1 to 2 to 4 to 8 to 16 to 32 to 64 and to 100. If you return 8 successful WUs, you'll be back at 100.

does this help? Need help troubleshooting the machine?

tony

yes it is it very hot in the netherlands but my systems are well cooled (the P4 CPU's are never above de 53 degrees)

After some research.
Rosetta gives me more WU's then I can handle with my preferences, so rosetta calculated the wrong numbers of WU's I need to overcome 3 days (=old preferences, now 2 days)
So the last 60 WU's or so where "out of date", I think that's the reason the were suddenly all 100% done.

And no, normally I don't use a manager or monitorprogram, not even boincmanager (boinc is installed as service), but for this week I will, just to be sure everything is normal again.
And for now, no quota anymore, so crunching again with no restrictions :)

---