On Fri, 2017-12-01 at 11:24 -0500, Phil Carns wrote:
Hi Ed,
I don't think we've seen this particular error before. Is it also the same application/executable every time in addition to being the same user in each case?
Yep, same application, "pcp". Not sure why I missed these cases before, but there are other users running this application, and each of their darshan3 files have the same errors. So it seems to be the application, not the user. Rob Latham:
Is there any chance darshan2 tools are trying to operate on darshan3 files? I hate to offer such an obvious suggestion, but you have generated both log formats so at some point in time you had darshan2 tooling lying around. Any darshan2 libraries or tools are probably long gone by now but I wanted to rule out the obvious.
I don't think so. All the darshan2 files have ".darshan.gz" as their filename suffix and the first few bytes are a gzip header: 1F 8B 08 00 These files have ".darshan" as their filename suffix, and the header starts with "3.10". $ head -c 64 logtree/00/2f/heitz_pcp_id7666317_10-15-53224-12998966924006311962_1.darshan | hexdump 000000: 33 2E 31 30 00 00 00 00 37 35 64 00 00 00 00 00 3.10....75d..... 000010: 00 00 00 00 00 00 00 00 FB 01 00 00 00 00 00 00 ................ 000020: 05 FE FF FF FF FF FF FF 00 00 00 00 00 00 00 00 ................ 000030: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................