Peter Memishian wrote:
> I've updated the block comment above check_testconfig() to cover this.

Looks good.

> 
> BTW, I've uncovered another issue during my testing: when an interface is
> brought back online, in.mpathd can end up logging a warning about missing
> test addresses before if_mpadm has had a chance to bring the test
> addresses back IFF_UP.  In the IPMP Rearchitecture workspace, the fix is
> simple: have if_mpadm bring the addresses back IFF_UP before bringing the
> interface back online.  In Nevada, this is not so easy because (a) the
> if_mpadm code is a mess and (b) in.mpathd tries to detect this condition
> (since it shouldn't be generally be probing offline interfaces) and emits
> a warning.  So probably the simplest fix for Nevada is to add another
> timer to prevent the warnings from being emitted when an interface comes
> back online :-(

Are you going to implement this now, or defer to another bug fix?  Is 
this the TESTADDR_CONF_TIME check at 885?

-Seb

Reply via email to