[Linux-ha-dev] Future of /proc/ha

[Volker Wiegand]

On Thu, 28 Oct 1999, Alan Robertson wrote:

> My only concern about the Berkley DB code is that most databases are *highly*
> prone to corruption during crashes.  I would recommend that no permanent state
> be kept in a database.  When a node comes up after a crash, its old idea of the
> cluster topology is now invalid anyway, so I don't envision this being a
> problem.
> 
Oh yes, I agree. Using a file based DBMS was _not at all_ for persistency
reasons, and of course a _clean_ cluster startup would initialize a fresh
empty set of tables.  With Berkeley DB I'm after two benefits:

(1) You get free local support for transactions.  I want this because my
model is a layered one.  The lowest layer is a reliable message exchange
within the cluster.  All other services build on top of this layer.

(2) You can at least try to get post-mortem analysis on crashed nodes.  It
can also be used as an audit trail if you keep the last two or three logs.

> A good API for accessing the data is essential.
> 
Great words from a great soul :-)

> Go for it.
> 
> 	-- Alan Robertson
> 	   alanr@bell-labs.com
> 
Volker

--
Volker Wiegand               Phone: +49 (0) 6196 / 50951-24
SuSE Rhein/Main AG             Fax: +49 (0) 6196 / 40 96 07
Mergenthalerallee 45-47     Mobile: +49 (0) 179 / 292 66 76
D-65760 Eschborn            E-Mail:  Volker.Wiegand@suse.de
++ Only users lose drugs. Or was it the other way round? ++