[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference ssag::ask_ssag

Title:Ask the Storage Architecture Group
Notice:Check out our web page at http://www-starch.shr.dec.com
Moderator:SSAG::TERZAN
Created:Wed Oct 15 1986
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:6756
Total number of notes:25276

6433.0. "swxcr,raid0,0+1,5 nt swap file" by SIOG::R_TAYLOR (Richard Taylor dtn: 822-4159) Thu Feb 27 1997 12:34

    
    hello,
    
    consider the following config.:
    
    as2100, 3-chnl swxcr, winNT 4.0 server
    
    channel 0 has an external CD tower
    
    Drive group A has 4 rz29b-va's: chnl 1 id 0
    				    chnl 1 id 1
    				    chnl 2 id 0
    				    chnl 2 id 1
    	these are defined as logical drive 0, raid 5, write thru.
    
    Drive group B has 3 rz28-va's & 1 rz28m-va: chnl 1 id 2
    						chnl 1 id 3
    						chnl 2 id 2
    						chnl 2 id 3
    	these are defined as logical drive 1, raid 0+1, write thru.
    				"	"  2, raid 0, write back
    				"	"  3, raid 5, write thru
    
    log drive 0 contains data
    "	"     1 is the NT boot disk
    "   "     2 is the NT swap file
    "	"     3 contains data
    
    Now......, my question!
    
    Am I right in assuming that any of the bits in logical drives 1,2 or 3
    could exist/reside on any of the four disks in drive group B; hence if
    any of the drives in group B gets faulted, some of the data in logical 
    drive 2 will be corrupt/missing (ie the NT swapfile) causeing NT to
    keel over thus makeing NONE of the drives available to the users!
    
    If this is true I will advise the customer (in-house) to make a change.
    
    Any suggestions, clarifications anyone?
    
    rgds
    Richard
    
    	p.s.
    
    	I also have a problem on this configuration where all the drives in
    group B get faulted when power is cycled on the Alpha, but I intend to
    upgrade the revisions of the four drives to try and fix this!
T.RTitleUserPersonal
Name
DateLines
6433.1Your CorrectSUBSYS::CARAKATSANEWed Mar 05 1997 07:3013
Hi Rich,

     Your correct, the data in logical drive 2 is striped across all drives
     in group B. Since the customer is using RAID 0, a RAID level with no
     redundancy, for logical drive 2 if any of the drives within drive group B
     should fail or fault, logical drive 2 would become unavailable and
     NT would "keel over" !! 

     My suggestion would be to tell the customer to change logical drive 2
     to RAID level 0+1.  Also be careful using "write back" on a logical
     drive, use it only if the controller has battery backup.

Bill