Yes, but technically, you will never lose two drives in the same RAID0+1 stripe set as the entire set goes offline after the loss of drive #1
You're right about the RAID10, too, but the likelihood of that 2nd failure being the hazardous one is not 100% like it is with the other solution. ============================================================== ASB - http://www.ultratech-llc.com/KB/?File=~MoreInfo.TXT ============================================================== Automation Leads To Relaxation... -----Original Message----- From: Dennis Depp [mailto:[EMAIL PROTECTED]] Sent: Tuesday, February 05, 2002 4:09 PM To: NT 2000 Discussions Subject: RE: SQL Server and RAID Levels Not completely true. In RAID O+1, I can lose two drives, but they have to be in the same stripe set in order to stay up. In RAID 10 if I lose both drives in a Mirror, then I am toast. Dennis Depp At 09:59 AM 2/5/2002 -0500, Andrew Baker wrote: >They're similar in cost. > >They're not similar in redundancy. In RAID0+1, you can only lose one drive >without issue. > >In RAID10, you could lose one drive from each mirror and remain up. > > > >============================================================== > ASB - http://www.ultratech-llc.com/KB/?File=~MoreInfo.TXT >============================================================== > Automation Leads To Relaxation... > > > >-----Original Message----- >From: Dennis Depp [mailto:[EMAIL PROTECTED]] >Sent: Tuesday, February 05, 2002 8:05 AM >To: NT 2000 Discussions >Subject: RE: SQL Server and RAID Levels > > >Just for the record, RAID 10 is not RAID 0+1. On RAID 10, you mirror the >drives and then stripe the mirrors. On RAID 0+1, you stripe the drives and >mirror the stripes. I think this is correct, but I may have it >reversed. These are both similar in cost and redundancy, but they are not >the same. > >Dennis Depp > >At 03:22 PM 2/4/2002 -0500, Ed Esgro wrote: > >Yes I was referring to RAID 10, which is just RAID 0+1. Although, I >honestly > >do not believe that RAID 5 is faster then 0 or 1 because the writing only > >hits one drive, the other drives mirror off of that write in the >background, > >(a good controller would do most of the work). RAID 5, must write partially > >to each individual drive and then write parity to each individual drive. > >Much more writing, hence less speed. > > > >-----Original Message----- > >From: Szlucha, Chris [mailto:[EMAIL PROTECTED]] > >Sent: Monday, February 04, 2002 3:06 PM > >To: NT 2000 Discussions > >Subject: RE: SQL Server and RAID Levels > > > >Forgive me if I'm wrong, but Ed wasn't saying to use both RAID 0 and RAID >1, > >but rather what is sometimes called RAID 10, or more properly RAID 0+1. >It > >is a mirrored RAID setup. Speed and redundancy, but it's the most costly >of > >the bunch. > > > >And your statement about it being faster on RAID 0 or 1 is incorrect. RAID > >5 is faster, as the write job is split up across the drives and each drive > >writes it's own piece of data at the same time as the others. > > > >-Chris > > > >-----Original Message----- > >From: Paul Timmerman [mailto:[EMAIL PROTECTED]] > >Sent: Monday, February 04, 2002 3:00 PM > >To: NT 2000 Discussions > >Subject: RE: SQL Server and RAID Levels > > > >It would all depend on the importance and size of the data. If there is > >going to be a large amount of data I would stick with thye RAID 5 for > >overall redundancy. In this case, I would recommend using 15K rpm drives >if > > > >that is monetarily feasable. Yes, writing is a lot faster on RAID 0 or 1, > >but does this meet your redundancy needs? > > > >I would be very wary of placing the logs on a RAID 0 drive. That is NO > >redundancy. Again, depending on the importance of the data data, it might > >be okay. However, remeber that RAID 0 means that if that drive goes, so > >goes the transaction log and your ability to do a point in time recovery. > > > > > > >From: Ed Esgro <[EMAIL PROTECTED]> > > >Reply-To: "NT 2000 Discussions" <[EMAIL PROTECTED]> > > >To: "NT 2000 Discussions" <[EMAIL PROTECTED]> > > >Subject: RE: SQL Server and RAID Levels > > >Date: Mon, 4 Feb 2002 14:46:54 -0500 > > > > > >How about using 0+1 on the SQL Database. You get speed and redundancy at > > >the > > >price of space. Much faster then RAID5. > > > > > > > > >-----Original Message----- > > >From: Anthony L. Sollars [mailto:[EMAIL PROTECTED]] > > >Sent: Monday, February 04, 2002 2:40 PM > > >To: NT 2000 Discussions > > >Subject: SQL Server and RAID Levels > > > > > > > > >I am building another production SQL Server for our services team, and >have > > >configured in my default configuration: > > >2 x 18gig SCSI on RAID 1 = OS & Pagefile > > >2 x 73gig SCSI on RAID 0 = Logs & tempDB > > >4 x 73gig SCSI on RAID 5 = SQL Database > > > > > > > > >The problem is the SQL engineers are questioning the performance of RAID5 > > >for their needs. > > > > > >We are using RAID 0 on the logs because this is transactional data that >is > > >not important, and we don't need redundancy here just sheer speed. But >they > > >are saying that RAID 0 should be used isntead of RAID5 on the 4 drive > > >array. > > >The bulk of the work on this RAID5 will be data manipulation, where they > > >willl run sql scripts that compress and organize the tables in the > > >database. > > >In my opinion RAID 5 is good for this also. > > > > > >-TOny > > >Thanks for any advice > ------ You are subscribed as [email protected] Archives: http://www.swynk.com/sitesearch/search.asp To unsubscribe send a blank email to [EMAIL PROTECTED]
