Lustre stripe error noise when resuming from checkpoint

[davebiffuk]

If a pcp job with -l (preserve Lustre striping) is resumed from a checkpoint, there are a lot of error messages of the form:

error on ioctl 0x4008669a for '/lustre/file/name' (10): stripe already set

which obscure other problems. It would be nice if these messages didn't appear.

(I understand that these are printed directly by the Lustre library)

[guycoates]

Provisional fix code is in 5df8852. That silences the messages. I'm not sure of the best way to pass lustre errors up to the calling code to allow the caller to do something useful...

[davebiffuk]

Thanks, that looks safe for kernel 2.6.11+ because pipe size is 64K from then on, and max path length is 4K. I guess there's an edge case for the combination of earlier kernels and path names approaching 4K long, the stderr content would be bigger than the pipe - I don't know how the library would react to that...