Date: Sat, 21 Feb 1998 10:28:11 +0300 (MSK) From: bag@sinbin.demos.su (Alex G. Bulushev) To: jfarmer@goldsword.com (John T. Farmer) Cc: bad@uhf.wireless.net, agdolla@datanet.hu, freebsd-isp@FreeBSD.ORG, jfarmer@goldsword.com Subject: Re: fault tolerant :)) setup Message-ID: <199802210728.KAA11173@sinbin.demos.su> In-Reply-To: <199802201953.OAA23382@sabre.goldsword.com> from "John T. Farmer" at "Feb 20, 98 02:53:08 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
> >> Is there any way you know of to set up a redudant system such that the ISA
> >> cards from the primary sever are automaticaly hot switched upon dead of
> >> the primary over to the secondary?
> >
> >i think it is impossibly using fbsd for now ...
>
> Not necessarily impossible...
>
> One approach would be to use a shared NFS server with a private little
> network. For example:
>
> _________ _____________
> | NFS |-------| Dedicated |
> | Server| | HUB | ____________
> |_______| |___________| | |
> | |______| Server 1 |
> | |__________| ____________
> |____________________________| |
> | Server 2 |
> |__________|
>
> Of course, now you've introduced a failure point within the NFS server
> and the dedicated network (ethernet, FDDI, ATM, etc.).
this is an other case ... we use it on high loaded system,
one NFS server, private 100Mb network and 3 equivalent servers
runing users tasks and share NFS disk ... u can use NAT at the
router for load balansing ...
Internet ...
<---------------+
|
+------+--------+
+--------------+------------+------|router with NAT|
| | | +---------------+
+----------+ +---+----+ +---+----+
|NFS server| |server 1| |server 2| ........
+---+--+---+ +---+----+ +---+----+
| | | |
+--+------------+--------------+
private 100Mb network
(network use hub or so, not showing on the picture)
NFS server is weak point of this scheme
u can replace NFS server by dual servers with disk on RAID shared via
SCSI bus and add second router, and second ethernet (used if first failed)
Alex.
>
> If I was seriously looking at building a new very high availability
> server design, I would probably start with the multi-processor kernel
> work and see if the underlying communication & sync. code could be
> generalized to work across linkages other than the current tightly
> coupled model. (For a look at approaches from the other direction,
> building applications that use networked computers as one, visit:
> http://www.netlib.org.)
>
> John (Will design systems for food.... :^>)
>
> -------------------------------------------------------------------------
> John T. Farmer Proprietor, GoldSword Systems
> jfarmer@goldsword.com Public Internet Access in East Tennessee
> Office: (423)691-6498 for info, e-mail to info@goldsword.com
> Network Design, Internet Services & Servers, Consulting
>
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-isp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199802210728.KAA11173>