cancel
Showing results for 
Search instead for 
Did you mean: 

compression as low as .262:1 Could this be my drive?

tryeager
Not applicable
Backup Exec 9.0 for Windows Servers
Windows 2000 Server
Seagate Archive Python 4106xxx (STD 224000N)
 
My 20G server finally accumulated enough data to require a second 12G tape.    No surprise, here.  I had been getting compression ratio of around 1.8 for a few years.  The general nature of the files have not changed,  It's mostly general office use with about 8G of the 20G being a SQL Anywhere 7 database.  The open database files are not backed up.   
 
When I learned that the 2nd tape was taking hours to finish, I discovered compression ratios around .4:1 to .5:1.  Original compression setting was software (I believe) and that was changed to hardware.   Ratios increased to nearly 1:1 on a 20M test job of closed database files, but was back to .5:1 on the full server. 
 
Countless experiments with block size/buffer size/buffer count/high water count yielded much improved  performance (up to 1.9:1) on the 20M data base files test and on a 1.6G mixed test (C:\program files).  But the same settings only got about .5:1 on a 240M of general user data and as low as .262:1 on a complete 8G general use drive.
 
I am familiar with  http://entsupport.symantec.com/docs/199542 , and I am interested in the hardware compression reason #2 "The system may not be able to keep up with the tape drive".  How do you determine if this is the problem?
 
Most of my test have been on an old tape (800 hours) but the 2nd tapes were all new and some tests were done on tapes of various ages with no noticiable difference in results.  Small jobs were run multiple times and the reuslts were identical, so there is consistency and repeatability to the problem. 
 
The drive has 4600 hours on it and 2.7 million soft write errors, but only 1200 soft read errors, 1 hard read error, and no seeks or hard write errors.  I am only now watching at how fast the soft errors accumulate.  Preliminary figures are about 20,000 added in 40 minutes on the 1.6G job.   That's about 87 times the rate of the drive over its full life.  So, maybe I have found the culprit, but I would like to hear some opions before I order a new drive. 
 
A cleaning tape is run every week and always has been, but there is a constant error that the drive needs cleaning.  That's always there, too, even right after a cleaning, and I have never figured it out.  Any thoughts on that?  Can the drive be deep cleaned with a swab and alcohol?
 
 
 
 
1 REPLY 1

Ken_Putnam
Level 6