On Mon, Nov 19, 2007 at 05:50:48PM -0800, Darren Reed wrote: > What I would like to have seen is the ability of a service to specify > how to transition from maintenance to offline (which may or may not > be the same as "disable") is made, and for the service to be parked > in that state until there's an external event (enable/boot.) ...
You half describe the maintenance state itself. There is no fix-yourself service method. Adding one might not be hard at all, but adding useful fix-yourself methods to complex services may be quite hard. I wonder too if this isn't something for FMA to grow, rather than SMF, but either way, it's new functionality that will need a lot of work to truly be useful. E.g., if sshd fails to start, and say it's because its host keys are missing and there are no suitable Kerberos V keytab entries, should its putative fix-yourself method generate ssh host keys? Is that really the right thing to do or should a sysadmin be asked? Or is it the right thing to do in some sites but not others? That's a simple case. It gets more complicated. Slightly more complicated: someone disabled the ssh service, started sshd by hand on port 22 and left it that way, then along comes another sysadmin who wants to start the service again, but it fails to start -- should that fix-yourself method find and kill that daemon? I bet things get awful complicated awful quick with, say, databases, filesystems, ... Which isn't to mean that we shouldn't try, but that this needs more thought. Nico --
