Kevin Gusler wrote:
How many disks are in your array and are they all raid 0?
What kind of controller (if hardware) are you using that is doing a Raid
0+1+1??
Kevin Gusler
Applications Developer
Mortgage Builder Software, Inc.
Main: 800.850.8060 ext. 106
Fax: 248.304.0601
www.mortgagebuilder.com
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Steve Romanow
Sent: Tuesday, June 23, 2009 2:04 PM
To: U2 Users List
Subject: Re: [U2] FW: Valid Backup on Windows Universe
Kevin Gusler wrote:
All -
We are attempting to enact IBM's double mirrored striped set backup
alternative.
This is where your server has a RAID array of 0+1 and then a second Raid 1
that is mirrored to the first mirror.
The process as IBM describes it is:
1. Quiesce the database (pause all writes, data can still be read,
no
loss of data)
2. Split off the second set of mirrored drives
3. Restart the database
4. Backup system using split off hard-drives
5. Rejoin the mirrors
Has anyone used this scenario in a windows environment?
Wondering what types of RAID systems people are using?
- (3) Arrays all of which are Raid 0 and in the OS you mirror
those
RAID 0's
- A Raid 0+1 Array and another Raid 0+1 and then in the OS mirror
the
2 Arrays
- (3) Arrays Raid 0+1 and in the OS you mirror the 3 arrays
Were there any drawbacks to this scenario?
Any scenarios that works better?
We use this scenario with aix. Should work ok with windows. As long as
you are using HARDWARE raid, I think it is safe. I wouldn't trust
software raid to this task.
_______________________________________________
U2-Users mailing list
[email protected]
http://listserver.u2ug.org/mailman/listinfo/u2-users
_______________________________________________
U2-Users mailing list
[email protected]
http://listserver.u2ug.org/mailman/listinfo/u2-users
I'm sorry, I am more of a user/dev than sysadmin. Our var set it up for
us. What I do know is:
AIX 5.3, external drive enclosure with 15 sas drives (I think). 12 are
used for the udvg, in what they called a triple-mirror. I know that is
not a technical term, but I never pushed the issue to determine what
raid level it is. We use tivoli storage manager (that comes with aix)
and use mkvgback. We basically do a dbpause, run a script called
do-sbom-1, which creates the split mirror /sbom fs, resumes the ud db,
startes the backup. After the backup completes to tape, do-sbom-2 is
called which rejoins the mirror. It takes like 30 seconds for step 1, 4
hours for backup to tape, and a couple of hours to resync the raid. The
users only see a 30second pause.
The failure I see of this system is that even though the db is paused,
it is not in a "safe" state for backup. The only way to achieve that is
to stop all scheduled jobs, logoff oall users, and complete all
transactions. Hard to do in a 24-7 business.
_______________________________________________
U2-Users mailing list
[email protected]
http://listserver.u2ug.org/mailman/listinfo/u2-users
