Our experience with disk pools on RAID5 volumes has indicated that slow performance is the norm rather than the exception. This is especially true when more than one operation was running at a time, i.e., more than one client session, more than one migration or backup stg operation. We have taken to putting disk pool volumes on JBOD disks rather than on RAID5 disks. Dramatic performance improvements when we made this change.
So, that might indicate why your disk pools are slow. I'll also concur with Paul: tape is not necessarily slower than disk. In fact, our experience, again, has shown that backup directly to tape is as fast as backup to disk. That said, if you have gazillions (technical term) of small files, this might not be the case. But I'm thinking client to disk won't be all that fast either. We are routinely seeing 35-40 GB/hour data transfer rates to tape (all varieties: LTO, SDLT, AIT3) either from disk to tape operations or client operations. Kelly J. Lipp Storage Solutions Specialists, Inc. PO Box 51313 Colorado Springs, CO 80949 [EMAIL PROTECTED] or [EMAIL PROTECTED] www.storsol.com or www.storserver.com (719)531-5926 Fax: (240)539-7175 -----Original Message----- From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED]]On Behalf Of Seay, Paul Sent: Wednesday, July 24, 2002 10:05 PM To: [EMAIL PROTECTED] Subject: Re: diskpool performance You are probably running at source disk speed in both cases. Just because it if fibre disk does not mean it is fast. It really depends on how you configured it, what type it is, etc. Paul D. Seay, Jr. Technical Specialist Naptheon Inc. 757-688-8180 -----Original Message----- From: Orville Lantto [mailto:[EMAIL PROTECTED]] Sent: Wednesday, July 24, 2002 2:03 PM To: [EMAIL PROTECTED] Subject: Re: diskpool performance What are your throughput values? Tape is not necessarily slower than disk, especially when the data is compressed on the tape drive. Orville L. Lantto Datatrend Technologies, Inc. (http://www.datatrend.com) IBM Premier Business Partner 121 Cheshire Lane, Suite 700 Minnetonka, MN 55305 Email: [EMAIL PROTECTED] Steve Freeman <[EMAIL PROTECTED]> Sent by: "ADSM: Dist Stor Manager" <[EMAIL PROTECTED]> 07/24/02 11:00 AM Please respond to "ADSM: Dist Stor Manager" To: [EMAIL PROTECTED] cc: Subject: diskpool performance Please could someone advise me on the following issue we have with tsm. I am a TSM novice and would appreciate some pointers etc. Environment : TSM 4.2.0 AIX4.3.3 ML09 Server and Client on same Server using Sharedmem 2 Stgpools defined diskpool and tapepool we have created a diskpool on RAID5 and created random access volumes Total Diskpool size is 150GB. Our nextpool is tapepool for migration upon high thresholds being reached. We have compression turned off on the client and set the selftune parameters in dsmserv.sys SELFTUNEBUF SELFTUNETXN Issue: IF we initiate a client backup using sharedmem to the diskpool, it seems to take the same time as it does to backup the client straight to tapepool. I would have thought the diskpool access would be quicker ? The Diskpool is located on direct fibre attached storage with large read/write cache. The tapepool is SCSI attached ATL library with 2 drives Any suggestions or pointers to increase the throughput to a respectable level would be appreciated. kind regards steve freeman [EMAIL PROTECTED]
