[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: dropping back-ldbm

>>If LDBM had been removed, we might also have found the incentive to work
>>around the BDB bugs we encountered, or to live with the ways BDB was a
>>poorer solution than LDBM at those times.
>>In any case, I guess my main point is that BDB still seems rather more
>>fragile than LDBM.
> That seems to me to be a faulty perception. In terms of data integrity,
> back-ldbm is far more fragile - it's just that the software would never
> detect when damage had occurred. back-bdb may require more attention to
> set up properly and maintain, but that is because the design offers more
> visibility into the actual condition of the data.

I only want to comment this last point, since I rose it earlier in this
thread.  I don't think back-bdb is more fragile than back-ldbm; in general
I do trust BDB (in fact, i would only compile back-ldbm with the same BDB
that's required by back-bdb).  Others (significantly some customers, but
also one of my bosses) have this perception, so I feel more comfortable if
I can have back-ldbm around while suggesting to use back-bdb (or
back-hdb).  All the new projects we start are based on back-bdb, and we
always suggest to use back-bdb to those who feel comfortable about
upgrading.  For my personal use, I've been using back-bdb since it came
out or so (and I suffered from some of its growth problems, of course). 
Note that for my personal use I don't even need to configure DB_CONFIG,
because BDB's defaults are fine ;).


Pierangelo Masarati

    SysNet - via Dossi,8 27100 Pavia Tel: +390382573859 Fax: +390382476497