LHC@home

Three units have been returned by a computer running Linux, but the results database still shows them as outstanding.

Is there a particular problem with the Linux client in this regard?

Three units have been returned by a computer running Linux, but the results database still shows them as outstanding.

Is there a particular problem with the Linux client in this regard?

Don't think this is specifc to Linux. It does happen occasionally, including a report of this just two days ago over on Rosetta.

Which version of the client are you using, please?

I think this is a bug in the server code - it should test the database transaction to see if MySQL has safely committed the changes before confirming the report & giving the client permission to delete the work. That way, when this kind of glitch happens work would be double reported rather than lost. The database already has code to ignore a double report.

Anyone reading this who is on the boinc devs boards, please pass this on, unless this suggestion has already been made there.

River~~

---

>Don't think this is specifc to Linux. It does happen occasionally, including a >report of this just two days ago over on Rosetta.

>Which version of the client are you using, please?

Version 4.43 running in console mode.

Boinc Manager starts, then stops due to failure to connect
to the server.

>Don't think this is specifc to Linux. It does happen occasionally, including a >report of this just two days ago over on Rosetta.

>Which version of the client are you using, please?

Version 4.43 running in console mode.

It might be worth an upgrade when you have time - we are now up to 5.4.9 for linux. The upgrade is reasonably straightforward - download the file from BOINC, save it in the directory that contains the BOINC directory, stop BOINC running, do an ls -l on that directory to note the owner and groupd of all the files.
Then run the downloaded file as a shell script. It will update the program files in the BOINC directory but will leave the data files intact.

Check the owner and group of the new files - the one thing the shell script sometimes gets wrong, in my experience, is the ownership. Use chgrp XXX * and chown XXX * if needed. If you forget to check this, the symptom is that BOINC exits immediately complaining of being unable to access files.

Then start BOINC in the same way you usually do(or reboot if you don't know how your box starts BOINC)

I have never had any problems updating the linux version, but obviously it has to be worth taking a copy (or a tarball) of the BOINC folder before the upgrade just in case.

Having said all that, I don't think the problem you mentioned is all that common, and I don't think it will be cured by the upgrade - but you will be gaining the benefit of a lot of other bug fixes.

R~~