My SSD crashed and had to be replaced. I installed a new copy of Windows 8 rather than use the backup in hopes of solving a hardware problem. After installing the video card drivers, various utilities, and Boinc, I ran Boinc. Here is the log:
05-May-2014 07:27:59  Couldn't parse account file account_setiathome.berkeley.edu.xml
05-May-2014 07:27:59 [---] Couldn't parse statistics_setiathome.berkeley.edu.xml
05-May-2014 07:27:59 [SETI@home] Project SETI@home is in state file but no account file found
05-May-2014 07:27:59 [---] Application setiathome_v7 outside project in state file
05-May-2014 07:27:59 [---] Application setiathome_enhanced outside project in state file
05-May-2014 07:27:59 [---] Application astropulse_v6 outside project in state file
05-May-2014 07:27:59 [---] File info outside project in state file
404 lines of these
05-May-2014 07:27:59 [---] Application version outside project in state file
26 lines of these
05-May-2014 07:27:59 [---] Workunit outside project in state file
200 lines of these
05-May-2014 07:27:59 [---] Task 01ap08ad.26954.22976.438086664205.12.41_0 outside project in state file
200 lines similar to above
05-May-2014 07:27:59 [---] Project files outside project in state file
05-May-2014 07:27:59 [---] State file error: project http://setiathome.berkeley.edu/ not found for task
10 lines of these
Eventually, the server sent 2 new tasks:
05-May-2014 07:49:45 [SETI@home] Scheduler request completed: got 2 new tasks
When one was completed and returned, 115 were sent, and I was back in business:
05-May-2014 07:54:53 [SETI@home] Scheduler request completed: got 115 new tasks
The server made no attempt to resend lost work units. All 200 work units were declared abandoned and sent to the error bin.
I am not sure what caused the error. The computer booted once the morning of May 4, but then the system froze soon after and would not boot again. The SSD exhibited hundreds of bad blocks when tested with chkdsk on another system.
I suppose that it is possible that the client_state.xml file was corrupted when the system froze. However, I did nothing to clean it up; whatever repairs were done were done by Boinc itself.
I used a spare hard drive to make the system run on the morning of May 5. The results you see above. Later, the system on the hard disk was transferred to an new SSD.
As far as the server is concerned, it is a new client system. Sharing results between systems was a way of cheating on the work done in SETI@Home Classic, and was declared to be not tolerated in BOINC...
I agree that the reason that this incident produced over 200 work units marked as abandoned was that I had to re-attach to the S@H project; I forgot about that.
However, that is not the fundamental problem. The question is, Why couldn't Boinc find the S@H account and statistics files? Naturally, I am not sure of the answer to that question, and it could be that they were merely corrupted. But I do know that I was able to make Boinc work after I took ownership (security-wise) of the whole BOINC-DATA directory tree, which is on G: instead of in the default location. When I installed a new Windows 8, the system knew nothing about the owner of that directory and its folders and files. In other words, it may be that the reason Boinc could not read the S@H account and statistics files was due to a security problem. When the Boinc installer sees that the BOINC-DATA directory tree already exists, why can't it make sure that all its folders and files are accessible, and at least tell the user if they are not?