HPE Storage Users Group

Hello,
it seems that until version 3.1.1 there were the rcopy options no_fail_wrt_on_err (default) and no_fail_wrt_on_err that determined the behaviour if the rcopy target was unavaibale or failed: Either to continue I/O or to fence the I/O.

Are these options are still available in 3.1.3? What is the default behaviour? Is there a replacement in order to specify the behaviour?

Cheers,
Thomas

An earlier version of this command provided fail_wrt_on_err and no_fail_wrt_on_err policies. These policies were deprecated in HP 3PAR OS release 3.1.1.

Basically the "fail_wrt_on_err" was deprecated since exactly zero people actually made use of this mode, my experience on other platforms also is that when implemented, the first time this actually kicked in they'd think better of it and immediately remove the policy. The "fail_wrt_on_err" effectively stopped/fenced/failsafed all I/O to the local array if the remote array (target) could no longer accept replicated I/O. May sound like a good idea but it isn't for 99.999% of people, without the need for a "fail_wrt_on_err" there was no longer a need for the alternate mode "no_fail_wrt_on_err".

The default is that even if the remote array can't accept replicated data, the local array will continue to service I/O with the changes captured in a snapshot for later replay. I'm sure if there were demand for this then it would be quickly reinstated, but I have yet to meet anyone that would actually use it outside of an attempted lockout in a vendor influenced RFP.