🚀 go-pugleaf

In article <1992Oct5.210010.8108@almserv.uucp> rmcghee@fnma.com(Rich McGhee)  
writes:
> Has anybody spent anytime thinking about or working on a High Availability  
> solution for NeXT machines?  That is, redundant disk drives, two machines  
> that switch over when one goes down?  Any ideas?
> 
> --
> Rich McGhee
> Federal National Mortgage Association
> 3900 Wisconsin Avenue, 4S-2H
> Washington DC 20016
> NeXTMail: rmcghee@fnma.COM

When I talked to next about fault tolerance in their systems,
meaning I was interested in redundant disk drive and down
time they said "Why you asking?".  Finally I was told that RDD
was not supported and would not be in the near future.

I too would like to know if anyone is interested in these
topics and have discussed it with NeXT/3rd parties and
what was said.  Again, NeXT gave me the feeling (about
8 months ago) that they didn't care.  I wonder if this
has changed any since they are focusing in MCCA and
this is a prime location for such issues.

Todd Nathan
Liveware Corporation
-- 
--
Todd Nathan			Voice: (303) 484-7607 (w)
Liveware Corporation		Email: <todd@liveware.com>

In article <1992Oct8.195446.8138@liveware.com> todd@liveware.com writes:
>In article <1992Oct5.210010.8108@almserv.uucp> rmcghee@fnma.com(Rich McGhee)  
>writes:
>> Has anybody spent anytime thinking about or working on a High Availability  
>> solution for NeXT machines?  That is, redundant disk drives, two machines  
>> that switch over when one goes down?  Any ideas?
>
>When I talked to next about fault tolerance in their systems,
>meaning I was interested in redundant disk drive and down
>time they said "Why you asking?".  Finally I was told that RDD
>was not supported and would not be in the near future.
>
>I too would like to know if anyone is interested in these
>topics and have discussed it with NeXT/3rd parties and
>what was said.  Again, NeXT gave me the feeling (about
>8 months ago) that they didn't care.  I wonder if this
>has changed any since they are focusing in MCCA and
>this is a prime location for such issues.
>
>Todd Nathan
>Liveware Corporation
 
  High availability is usually (not always, but usually) associated with
  file or database servers, and occasionally high performance compute 
  servers.  NeXT workstations mostly serve the purpose of front-end
  application or personal workstations, neither of which generally are
  considered requiring high availability since bringing one unit down
  doesn't effect a large number of users or any significant amount of
  persistent data.  

  There are off-the-shelf RAID (redundant arrays of inexpensive disk)
  clusters which are SCSI and can be plugged into a NeXT without any
  trouble which will give you high disk availability.  These clusters
  can usually operate with a disk down (power out or serious electronic
  malfunction) and can even be hot-fixed with a new drive.  The system
  will reconfigure itself when the drive comes back on-line, restriping
  data across the new disk configuration if necessary.  RAID technology
  is clearly the coming thing in database servers and could be easily 
  and effectively used in small-scale file servers as well (as one might
  use a NeXT "server" workstation).  They are not inexpensive compared
  to typical SCSI disks (about $7,000-10,000 for a gigabyte or storage 
  vs. $2,000 for a regular gigabyte) but the benefits of high availability
  may justify the cost in many applications.  The beauty of these systems
  is that they are transparent to the main processor and operating system,
  looking exactly like a normal large SCSI disk device.  Drivers generally
  require little or no modification to work with RAID clusters.

  Dual processor systems with high availability are not likely to be
  seen in the NeXT family for quite a while, IMHO.

  Jon Rosen