[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: CVS and database

From: Greg A. Woods
Subject: Re: CVS and database
Date: Fri, 12 Apr 2002 13:41:30 -0400 (EDT)

 Dausy Wim <address@hidden>, Wim_DAUSY/BE/address@hidden wrote:
> Or do we have to provide our own database to work
> with CVS, if we want this kind of security?

You could do passive auditing with redundancy.....

Eg. use CVSup to replicate the repository to one or more backup machines
and then use a "dry-run" replication back to the master after a crash to
check for consistency problems.  CVSup is quite efficient, even for
things the size of *BSD + X11 + 10k or more tiny files in pkgsrc/ports.
As long as you don't have hundreds of developers you can probably run it
frequently enough that it would only leave you with a small number of
changes to manually check if there's a crash at any time.

Of course you'll want very reliable power and environmental security for
those systems, well engineered server-class systems with a proven
history of reliability, and well proven hardware RAID protected (with
battery-backed cache and securely tested cache flush mechanisms if it
has cache) storage for both the main and backup servers; and you'll want
to use the most stable OS platform you can find with the most proven
reliable filesystem implementation (eg. plain simple old BSD FFS with no
fancy new features such as soft dependencies, or maybe a well proven log
structured filesystem such as IBM's AIX LFS).

(i.e. you don't want your code assets sitting on one spindle in a
commodity PC sitting on the dusty carpet under someone's desk!  ;-)

                                                                Greg A. Woods

+1 416 218-0098;  <address@hidden>;  <address@hidden>;  <address@hidden>
Planix, Inc. <address@hidden>; VE3TCP; Secrets of the Weird <address@hidden>

reply via email to

[Prev in Thread] Current Thread [Next in Thread]