Some days ago, I suddenly found the following error message on my screen:

diskfull

At first, I simply restarted the computation from the last checkpoint to verify that the problem is reproducible. After three days I encountered the same problem again. So I really had to care about the problem, which already delayed the computation by one week.

One way to get more disk space is to move some of the checkpoint files to my spare disks and to create soft links pointing to them, such that y-cruncher still finds them in the same place. With four 6 TB spare disks, this gives me about 1 TB more space on each of the 20 disks used for the computation.

Today, the computation has been running for four days since this reorganisation of the checkpoint files. Hopefully there is now sufficient disk space to finish the computation, which is currently in the final steps of the series summation.

More Posts