From: Paul H. Hargrove (PHHargrove_at_lbl_dot_gov)
Date: Mon Feb 23 2009 - 11:32:22 PST
In "raw" terms: This error message shows that while trying to write, an error code -27 was returned from the kernel's vfs_write() and this same error has been reported by two levels of function calls. In useful terms: A write to the checkpoint context file has failed with errno=27, which is EFBIG. According to the man pages this means that a file is "too big", which can IIRC mean either that a per-process limit or per-filesystem limit has been reached. In the per-process case try "man setrlimit", or your shell's "limit" or "ulimit" builtin. The per-filesystem limit is less likely, but if you are not saving to one of the "normal" Linux filesystems, it is possible that there could be a 2GB or 4GB limit to file size. -Paul Alexandre Strube wrote: > $ cr_checkpoint 16783 > - vfs_write returned -27 > - cr_freeze_threads failed (-27) -- Paul H. Hargrove PHHargrove_at_lbl_dot_gov Future Technologies Group HPC Research Department Tel: +1-510-495-2352 Lawrence Berkeley National Laboratory Fax: +1-510-486-6900