I will try more tomorrow, but I think so far I cannot connect back to around 4 days worth of work on this puter. I may have been overriding old competed tasks, or they may be sitting there finished and I just cannot connect to upload.
You can see the server's view of your client instances at
My Contribution: Results Status. It's awkward to use though, among else because all client instances on the same computer most likely have the same computer name.
Further, to gain some info on what is left and what not, you can look into the client data directories and what files they contain in the projects\www
.worldcommunitygrid
.org\ subdirectory. It may be difficult to tell task files apart from results files though. (There are also files of the applications and their screensavers in there.) And of course the file client_state.xml in the client's top data directory contains information about any tasks and results which are left in the client.
To get this cleaned up, my suggestion would be to
- first make sure that boinc is not automatically started at reboot
(if in doubt, Windows users need to chime in whether or not this is possible to configure, and if so, how),
- reboot as an easy means to shut down all clients,
- edit all gui_rpc_auth.cfg files to contain a password which you can easily memorize and enter (or no password at all, if you know that the boinc RPC ports cannot be accessed remotely),
- review all cc_config.xml files to contain <allow_multiple_clients>1</allow_multiple_clients>,
- then start one client instance, — only one —, taking care of the right data directory and right port number in the command line of boinc.exe,
- attach boincmgr to it,
- check what's up with it, switch the WCG project state to "no new tasks", switch on network transfers, complete any left tasks before their deadline or abort tasks which are already past the deadline, upload all results, report all results = trigger an update of the WCG project,
- shut the client down,
- check in taskmanager that the client really was shut down,
- repeat with one next client instance.
After that, proceed working with one of the instances, e.g. the primary one.
Later, look into BoincTasks as an additional means to control several client instances.
You will get the hang of running several instances safely at the same time too.
PS, use the "advanced view" of boincmgr, not the default simplified view.
Perhaps more than one client was started using the same data directory.
For Linux, I have a startup script which makes a good attempt to avoid this situation:
Step 5.1 in
my Howto.
Maybe somebody experienced with Windows scripting can come up with an equivalent for Windows.
I once wrote one or another Powershell script, but I would have to re-learn the features and syntax of this scripting language every time, and I find it icky.