Re: EMC vs. ordinary RAID 5
From: Stefan Weideneder <stefan@weideneder.de> > >just because some guys of this wonderfull NG told us >that we should avoid RAID 5: > >Here's the result of an I/O throughput comparison >unfortunately made on two different platforms: Unlike Art, I don't have much of a beef about the performance of RAID5, because the caching often masks the performance hit. Like Art, I have considerable worries about the integrity of the data on RAID5, because the checksum isn't ever verified. I've personally seen a RAID5 let go like this, and it was a frightening thing to behold. So, performance isn't my problem with RAID 5, it's the lack of data integrity. If we hadn't had good backups...<shudder> ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com
Obnoxio The Clown wrote: > > From: Stefan Weideneder <stefan@weideneder.de> > > > >just because some guys of this wonderfull NG told us > >that we should avoid RAID 5: > > > >Here's the result of an I/O throughput comparison > >unfortunately made on two different platforms: > > Unlike Art, I don't have much of a beef about the performance of RAID5, > because the caching often masks the performance hit. Like Art, I have > considerable worries about the integrity of the data on RAID5, because the > checksum isn't ever verified. I've personally seen a RAID5 let go like this, > and it was a frightening thing to behold. So, performance isn't my problem > with RAID 5, it's the lack of data integrity. If we hadn't had good > backups...<shudder> > ______________________________________________________ > Get Your Private, Free Email at http://www.hotmail.com Hey you, today I discussed with some Informix customers, what you might reply. Now you know, you are well known in the Informix world :-) I know on some hardware platforms we had troubles with RAID 5. I never found the reason why. But we had a few problems with Kernel-AIO, restore failures, some server bugs, too. I agree, from time to time we should archive our data. Best regards, -- Stefan