RE: Re[2]: Point-of-Sale RAID application

1999-02-23 Thread Brothers, John
Actually, I only need one box, with a mirrored disk and ideally with hot swap/auto rebuild capabilities as well. Thanks for all the help guys. > If you mean to be able to recover from a CPU crash in the middle of a > transaction - OK. But it's definitely not what the original poster had in

Re: Re[2]: Point-of-Sale RAID application

1999-02-23 Thread m. allan noah
actually did not save any hard numbers, but on a 3 disk setup i had here, raid1 was faster under bonnie with the test size 3x my ram. less than that, and i found the results pretty variable. but, our boy was talking about a single hotel firewall, and yet this discussion now revolves around beowul

Re[2]: Point-of-Sale RAID application

1999-02-23 Thread Evgeny Stambulchik
Osma Ahvenlampi <[EMAIL PROTECTED]> wrote: > [EMAIL PROTECTED] (Alvin Oga) writes: > > for larger systems/clusters: > > what about the linux High Availability project ? > > and/or the beowolf project ? > > No Linux project I know of can yet provide generic failsafe clustering. If you mean

RE: Re[2]: Point-of-Sale RAID application

1999-02-23 Thread Andy Poling
On Tue, 23 Feb 1999, Brothers, John wrote: > Actually, I only need one box, with a mirrored disk and > ideally with hot swap/auto rebuild capabilities as well. I would look at the hardware RAID SCSI controllers. They provide always-bootable (i.e. they present a boot sector from the RAID array)

Re[2]: Point-of-Sale RAID application

1999-02-23 Thread Evgeny Stambulchik
"m. allan noah" <[EMAIL PROTECTED]> wrote: > 3. raid 5 is uneeded, as raid1 gives better read performance Hmm, did you actually compare? I haven't benchmarked a 2-disk array, but with 3 disks, RAID5 beats RAID1 on BOTH writes and reads. Theoretically, this shouldn't happen, but it's a current s