List:General Discussion« Previous MessageNext Message »
From:Adrian Phillips Date:March 19 1999 7:42pm
Subject:Re: fastCPU vs moreRAM
View as plain text  
>>>>> "Wade" == Wade Maxfield <maxfield@stripped> writes:

    Wade>   Well, I understand your opinion and your point.

    Wade>   However, raid 5 can usually be had in a hot swap
    Wade> environment, and it can consume as little of 20% of the disk
    Wade> resources to implement.

    Wade>   If my memory serves, raid 0 is striping and raid 1 is
    Wade> mirroring.  Mirroring uses 50% of the disk resources to
    Wade> implement.  The mirroring solutions I have seen requires you
    Wade> to take the system down to recover. I have not seen a hot
    Wade> swap mirror environment, though one may exist.  I have seen
    Wade> several raid 5 hot swap environments.

For PCs ? I don't know - AIX has had mirroring for some years as part
of the volume manager - hot swapping has functioned for a while IF the
hardware allowed it. We have a couple of SSA (serial storage
architecture) based systems that can either do RAID in hardware or
mirroring as part of AIX - SSA allows hot swapping. Both of these
systems have 2 machines attached to the SSA cabinett - SSA allows many
machines/disk cabinets connected in a loop - if one end of the loop
disappears all systems will automatically go the other way to get to
the disks they "own". Very nice solution I think (although it costs
more that off-the-self PC Raid systems).

To keep this Mysql related - we run Mysql on all of these systems !!

In fact there is a high availability option that allows more than one
AIX system to use the disk/filesystem at the same time - a much faster
NFS I believe, although I haven't personally tried it.

    Wade>   It really is up to the final customer.  I prefer raid 5
    Wade> over mirroring for maximum up time in the commercial
    Wade> products I've seen.

    Wade> wade

Adrian
Thread
fastCPU vs moreRAMScott Moseman18 Mar
  • Re: fastCPU vs moreRAMWade Maxfield18 Mar
  • Re: fastCPU vs moreRAMSasha Pachev19 Mar
  • Re: fastCPU vs moreRAMHironori Sato19 Mar
Re: fastCPU vs moreRAMDerick H Siddoway19 Mar
Re: fastCPU vs moreRAMDerick H Siddoway19 Mar
Re: fastCPU vs moreRAMKarl Pielorz19 Mar
Re: fastCPU vs moreRAMWade Maxfield19 Mar
  • Re: fastCPU vs moreRAMAdrian Phillips19 Mar
Re: fastCPU vs moreRAMDerick H Siddoway19 Mar
Re: fastCPU vs moreRAMDerick H Siddoway19 Mar