Posts by Ageless


log in
21) Questions and Answers : Unix/Linux : boinccmd --read_cc_config -> retval 0 -> Proxy settings not loaded! (Message 1671222)
Posted 24 days ago by Profile Ageless
In which directory did you store cc_config.xml?
22) Questions and Answers : Windows : Seti@Home Question (Message 1670924)
Posted 24 days ago by Profile Ageless
Seti Enhanced was v6, the one before the present versions. It's been deprecated, but hasn't yet been removed from the project preferences. The same answers for Astropulse v6.
23) Questions and Answers : Windows : More Computers, Less Credit :( (Message 1670350)
Posted 26 days ago by Profile Ageless
Do some tasks give more credit then others?

Normally:
Seti Multibeam pays about 30 to 120 credit, depending on angle range.
Seti Astropulse pays 510 to 630 credit, depending on amount of radar blanking.

So as you can see, when running lots of Astropulse, your RAC will go bigger faster than running only Multibeam. Astropulse is not available at the time of writing, though. There are only a handful of resends going around.
24) Message boards : Number crunching : Can't reinstall Boinc (Message 1669909)
Posted 26 days ago by Profile Ageless
Anything else you did that you didn't say?
Why not tell everything you already tried, so we don't have to second guess?
25) Message boards : Number crunching : Can't reinstall Boinc (Message 1669870)
Posted 26 days ago by Profile Ageless
Well, you didn't say that.

I read around that this could be due to a corruption problem in the registry.
You may want to delete HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Space Sciences Laboratory, U.C. Berkeley

Then if the problem continues, delete HKEY_CURRENT_USER\Software\Space Sciences Laboratory, U.C. Berkeley as well.

You may want to back them up before you delete them. Right-click the key->Export->give a name and save.

Do know, deleting the HKLM key will delete the installer's 'knowledge' of where your data directory is. If you changed it, you'll have to change it again, as else BOINC starts up clean from C:\Programdata\BOINC\
26) Message boards : Number crunching : Can't reinstall Boinc (Message 1669864)
Posted 26 days ago by Profile Ageless
First try to download the installer anew, saving it to a different location/drive than the last time.
27) Questions and Answers : GPU applications : my results were shown as computing errors (Message 1669321)
Posted 28 days ago by Profile Ageless
Your tasks are failing with error code 0xc0000026B, which in Windows terms means STATUS_DLL_INIT_FAILED_LOGOFF. Which when read in that sense, seems to be a problem with the drivers you use for the Intel device.

Do you know which drivers you use?
If not 10.18.10.3621, try those first. These are old drivers, but sturdy. Here in GPGPU land, the latest drivers don't always mean they're good for use with GPUs doing OpenCL data crunching.

If you want to try newer ones, the other one tested by people here is 15.33.32.64.4061.
28) Questions and Answers : Getting started : Possibly getting new tasks when "no new tasks" is chosen (Message 1668969)
Posted 29 days ago by Profile Ageless
BOINCMGR.EXE communicates with BOINCCMD.EXE via TCP/IP

New to me... :)

You meant BOINCMGR.EXE communicates with BOINC.EXE :)
29) Message boards : Number crunching : Problem with the SSL CA cert (Message 1667700)
Posted 20 Apr 2015 by Profile Ageless
If I'm not mistaken that would also mean that you can only run BOINC as a service on the server, and no longer on any workstation on that network/domain, because only one account with the BOINC user names can be made.

But I'll ask Rom.
30) Message boards : Number crunching : Problem with the SSL CA cert (Message 1667694)
Posted 20 Apr 2015 by Profile Ageless
Could the required local accounts and groups be documented so such a thing can be scripted by SysAdmins?

Hasn't changed (much) since BOINC Six: http://boinc.berkeley.edu/trac/wiki/ClientSetupLogicWinSix

Or is there a way to create global accounts and groups?

No, else it would've been done ages ago.
31) Message boards : Number crunching : Problem with the SSL CA cert (Message 1667688)
Posted 20 Apr 2015 by Profile Ageless
Is there a technical reason that BOINC can't be installed as a service on a DC?

The same reason as to why all of BOINC 6 and part of BOINC 7 cannot be installed on a DC: the making and using of local accounts and groups specifically for BOINC.

Since 7.3.8 there's been a change in BOINC setup that it makes and uses the BOINC accounts and groups only when installed as a service. No longer on the 'normal' or 'user' install.

DCs use global accounts, so when the BOINC installer tries to make local accounts, it'll err on that.
32) Message boards : Number crunching : Problem with the SSL CA cert (Message 1667657)
Posted 20 Apr 2015 by Profile Ageless
I finally got an answer back from Rom.

He'll test with the installer of a next BOINC (7.5.x I guess) if the check for a DC can be disabled. That way BOINC can be installed on a DC, just not as a service. Edit: But the installer will probably grey out the option to install as a service when it detects it's being installed on a DC.

As for XP, BOINC as a service and GPUs, the answer is No, that change will not be reverted.
33) Message boards : Number crunching : Problem with the SSL CA cert (Message 1666133)
Posted 16 Apr 2015 by Profile Ageless
But could we have a version from current sources, built with current tools

A current version of what? OpenSSL? For what?
When building the BOINC client, it's built against a couple of libraries. OpenSSL is one of those libraries. So it's not possible to just update OpenSSL on the older clients, because that then breaks that client, as it's expecting the older OpenSSL library. Then you get what Charlie (Ozzfan) had, that the client just crashes.

So the only option would be to build a new OpenSSL using the older compiling tools and libraries, but with the newer OpenSSL libraries. That's not undoable, but won't be done by the developers. Perhaps that if the community feels it needs to be done, that someone, or some people can do it.
34) Message boards : Number crunching : Problem with the SSL CA cert (Message 1666088)
Posted 16 Apr 2015 by Profile Ageless
David Anderson:
The SSL certificate used for this is issued by a UC Berkeley CA, which I imagine is not included in older CA bundles.

Not worth pursuing this point; we're fine with HTTP.

(read: costly)

Apropos, it's not possible to release new versions of old clients with new OpenSSL, or just an updated OpenSSL for older clients.

Rom:
OpenSSL would have to be re-built using the same version of Visual Studio (with all the patches) used to build that version of BOINC, e.g. with 5.10 we were probably still using Visual Studio 2003.
35) Message boards : Number crunching : Problem with the SSL CA cert (Message 1666070)
Posted 16 Apr 2015 by Profile Ageless
From Rom Walton:
It probably has to do more with what is supported in the specific version of OpenSSL included with BOINC than the CA Bundle.

Backwards compatibility has been in decline on the web with Heartbleed and Freak being discovered. It would not surprise me if older BOINC clients were having problems connecting to up-to-date BOINC servers over SSL.
36) Message boards : Number crunching : Problem with the SSL CA cert (Message 1666028)
Posted 16 Apr 2015 by Profile Ageless
Since no one else seems to have asked the project admin, I did. :)
Seti switched to using HTTPS for scheduler requests a while ago to avoid sending authenticators in cleartext.

For the moment the admins have turned it back to HTTP since this now causes the troubles with older clients. They'll look for a different way to protect the innocent... errr.. :)
37) Message boards : News : Workunit shortage (Message 1663914)
Posted 11 Apr 2015 by Profile Ageless
...when finished you will deserve something a little stronger than coffee.

A Red Bull?
38) Questions and Answers : Windows : No new tasks after update v7.4.42 x64 wx3.0.1 (Message 1663252)
Posted 9 Apr 2015 by Profile Ageless
And even more from Matt:
Due to recent problems we are doing a deep cleaning of one of our larger databases. This is unfortunately progressing slower than expected, and keeping us from generating new work. We might be in this state until next week.
39) Message boards : News : Workunit shortage (Message 1663251)
Posted 9 Apr 2015 by Profile Ageless
Good luck.
40) Questions and Answers : Windows : No new tasks after update v7.4.42 x64 wx3.0.1 (Message 1663149)
Posted 9 Apr 2015 by Profile Ageless
http://setiathome.berkeley.edu/sah_status.html shows Results ready to send 0 and 0.

Matt Lebofsky, project scientists wrote news here:
So! All the recent headaches are due to continuing issues with the master science database. While all the data seems to be intact, there's something fundamentally wrong causing informix to keep hanging up (usually when we are continuing work on reconnecting the fragmented result tables).

During the previous recent crashes I would clean up what informix was complaining about in various error messages (always the result table indexes) but this time I'm in the process of doing a comprehensive check of everything in the database just to be sure. And, in fact, I'm seeing minor problems that I've been able to clean up thus far (once again, no loss in data - just internal bookkeeping and broken index issues).

I thought this full check would be done by now (ha ha) but it's not even close. Meanwhile we *should* be able to do Astropulse work, but the software blanking engine requires the master science database to do some integrity checks, so that is all offline as well.

There are ways to speed up such events in the future. We're working on enacting several improvements. Yes, we here are all beyond tired of our project grinding to a halt and things will change for the better.

- Matt

He also wrote earlier on the 8th of April:
Hey - just so y'all know it's the science database again. This time enough is enough and I'm doing a comprehensive set of integrity checks on everything in that database before starting it up again. So no MB splitting or assimilating. Eventually some work will show up for AP splitting in the meantime...

Might be back up by the end of the work day, if not shortly after that. I did find one problem which was obscured while checking things out after previous crashes, so there's hope.

- Matt


Previous 20 · Next 20

Copyright © 2015 University of California