On Tue, Oct 09, 2001 at 04:40:11PM -0700, Andrew Ryan wrote: > On Tuesday 09 October 2001 03:06 pm, Jim Trocki wrote: > > the dtlog records are only written when the host stops failing. has > > it ever gone through the state transition of failing->not failing? i'd > > imagine not if you haven't seen any entries in dtlog. > > On a somewhat related note, is there a better way to get dtlog to record > failures which never recover before the server is killed or reset? It's not > worth a lot of effort, but if there's something easy which could help, that > would be nice. > > Even nicer would be authoritative state maps that persist across server > resets/restarts, and some form of downtime logging that is complete enough to > build an extremely accurate picture of up/downtime and service output. > Especially if it is available on a per-host basis. All of which I suppose is > on the list for 1.2...
On the state map note, we could tie a hash to a dbm for persistant state across restarts. As for complete downtime logging, I'd like to see that too. When things calm down a bit a work (yeah right) I plan to look at this. My boss loves Mon and is the boss of at least three very adept perl hackers, and is a perl wiz himself. I'll see if our group can contribute some useful patches to Jim in the coming months. -- Nate Campi <[EMAIL PROTECTED]> GnuPG key: 0xC17AEF79 http://www.campin.net So so is good, very good, very excellent good: and yet it is not; it is but so so. -- William Shakespeare, "As You Like It"
