|
Larry, okay, okay, you are right. I thought that, when you start with a RAID set, of lets say 4 drives à 8.58 gb, then you add up to three, which are still "under protection" of the first 4 forming the RAID set, then you have full capacity with those three additional disks. No you carry on adding disks and get another set of four (reducing capacity ?) until you add more disks than 8, 12, 16 .... (What I was guessing when looking at AS/400 RAID was a modulo 4 scheme) But I missed the fact, that there is a change in handling, when you add all disks at once. So there must be two different "algorithms". As Neil Palmer stated somewhere else on this list, you should stop parity protection when adding more than 3 disks of equal size to a RAID set, after adding to your config, you start it again and get RAID info spreaded in a new fashion. OK. Lets perhaps discuss a new theme: What do you prefer: RAID-5 or good old mirroring ? I had several disk failures at my customers machines and RAID-5 was awfully slow when one disk was failing ... about 10 - 20% of normal performance. Okay we did not loose any data, but mirroring is still the better way to keep on with your business like nothing happened .. except you need to call a technician to replace a disk. Regards, Philipp Larry Bolhuis schrieb: > Phillip, > > I maintain that I AM correct: > > In your new 170 there are 10 drives. ALL 10 drives are in the SAME > RAID set. Thus 8 drives of 7.51G and 2 of 8.58G in the set. Total > usable storage of 77.22G. > > Where you are mislead is that on your 620 the drives were of different > sizes and so must be in different sets - ALL drives in a set must be of > identical capacity. On your 170 the drives are all the same size so > they can all be in one set. From 4 to 10 drives can be in a single raid > set. > > Hence the New Math! > > HTH - Larry > > Philipp Rusch wrote: > > > > Hello Larry, > > > > Obviously, youre right ! > > > > But you are only near the complete truth ... > > > > The 620 we were moving from had 5 disks 4.1 gb and 5 disks 8.5 gb, > > like this, 4 disks of each size form a RAID set and the fifth disk of each > > set is protected by the others of same size - agreed. > > But, the new 170 is equipped with 10x 8.5 gb and is also having 2 disks > > with complete capacity and 8 disks with reduced capacity, which I would > > expect to have the same size as with the 620 because every 4 disks form > > a set ... (?) > > > > Still no clue what is this new math now, > > > > regards, Philipp > > > > Larry Bolhuis schrieb: > > > > > Philipp, > > > > > > It's the new math! > > > > > > Actually you have fallen into the trap of picking a percentage which > > > is not correct. To calculate storage lost to RAID you simply subtract > > > the capacity of 1 disk from the set. This will range from a high of 25% > > > on the smallest set (1 disk of 4) to a low of 10% on the largest sets (1 > > > disk of 10). > > > > > > Note that the amount lost per drive varies as the RAID data is kept on > > > either 4 or 8 drives depending on the total number of drives in the set. > > > (4 drives for sets of 4 to 7 drives and 8 drives for sets of 8 to 10 > > > drives). Thus WRKDSKSTS will report different sizes for the drives > > > carrying RAID data than those that are not. > > > > > > Likely your 170 has 10 drives (1 of 10 lost to RAID) and your 620 had > > > 5 (1 of 5 lost). > > > > > > HTH - Larry > > > > > > Philipp Rusch wrote: > > > > > > > > Hi all, > > > > > > > > I noticed while being at a transition job from one AS/400 to another, > > > > that there is a difference in how V4R4 and V4R5 handles disk sets on > > > > a 2740 controller. > > > > I used to calculate the resulting capacity when "raiding" a set of disks > > > > on a 274x controller as a loss of about 20% of total capacity. > > > > When working on a 620-2179 with V4R4 and 6713 disks (8.58 GB)we > > > > got a resulting capacity of about 6400 MB each, the same disks gave > > > > me round about 7512 MB on a system 170-2385 with V4R5 and both systems > > > > were using a 2740 RAID controller. > > > > Looks to me as if we have a better algorithm as before, because only > > > > about 10% is used up for running RAID-5 on that set. > > -- > Larry Bolhuis > Arbor Solutions, Inc. > (616) 451-2500 > (616) 451-2571 -fax > lbolhuis@arbsol.com > +--- > | This is the Midrange System Mailing List! > | To submit a new message, send your mail to MIDRANGE-L@midrange.com. > | To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com. > | To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com. > | Questions should be directed to the list owner/operator: david@midrange.com > +--- +--- | This is the Midrange System Mailing List! | To submit a new message, send your mail to MIDRANGE-L@midrange.com. | To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com. | To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com. | Questions should be directed to the list owner/operator: david@midrange.com +---
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.