rbraun@spdcc.COM (Rich Braun) (05/18/91)
My 3.1.5 upgrade came yesterday, and it took forever to install. IBM's two-page instruction guide fails to tell you when you can take extended coffee breaks during the procedure, so I had to baby-sit the thing for about three hours. (It also fails to be explicit about avoiding such things as specifying device /dev/rmt0, which causes a lengthy tape read followed by a cryptic error message which doesn't explain that one just screwed up and should have typed /dev/rmt0.1. Ugh.) Anyway, I'm wondering if I should call up IBM service or just live with my slow tape drive. The model number shown on the front is 7207-001, and the problem is that it takes about 5 minutes to load a tape even if the tape was already rewound. The little green light comes on and the tape spins for this long before one can access the tape (indicated when the light finally goes out and the grinding--er, whirring--noise stops). Once the unit gets going, it can back up a good-sized filesystem in a reasonable amount of time, but the rewind and mount operations are far slower than equivalent 150-Mb cartridge units on my other systems. I miss the old TU-72 units from DEC, which had nice big windows through which one could see exactly what was going on. (I don't miss the 40,000- tape archive room at the U. of Delaware, which was filled with bulky 9-track tapes...) Shall I call IBM service to get this thing fixed, or shall I call IBM sales to get this thing scrapped? -rich
ken@dillman.austin.ibm.com (05/18/91)
Try changing the automatic retension feature of the tape drive to off using SMIT. That should save you a great deal of time every time you load a new tape. ----------------------------------------------------------------------------- Ken Dillman | IBM | Phone: (512) 838-4804 T/L 678-4804 AIX Customer Suppt| 11400 Burnet Road| VNET/PROFS: KDILLMAN at AUSTIN Zip 2830 | Austin, TX 78758| Internet: kdillman@austin.vnet.ibm.com
rbraun@spdcc.COM (Rich Braun) (05/20/91)
ken@dillman.austin.vnet.ibm.com () writes: >Try changing the automatic retension feature of the tape drive to off >using SMIT. That should save you a great deal of time every time you >load a new tape. Thanks for posting this. I got several e-mail replies suggesting the same thing. A question for IBM: why _in the world_ would a company put out a product which by default would do such a thing? It really makes the system seem cruddy and slow. To the uninitiated (myself, for example), it may not be all that obvious that one should go into the SMIT menu and change a simple, relatively unexplained parameter. After all, it's very easy to screw the system up by changing parameters one does not understand; my inclination is to leave things well enough alone if they aren't broken. This truly amazes me. -rich
shair@ux1.cso.uiuc.edu (Bob Shair) (05/20/91)
rbraun@spdcc.COM (Rich Braun) writes: >A question for IBM: why _in the world_ would a company put out a product >which by default would do such a thing? It really makes the system seem >cruddy and slow. To the uninitiated (myself, for example), it may not >be all that obvious that one should go into the SMIT menu and change a >simple, relatively unexplained parameter. After all, it's very easy to >screw the system up by changing parameters one does not understand; my >inclination is to leave things well enough alone if they aren't broken. >This truly amazes me. >-rich Rich, I usually refuse to answer "why" questions, and I'd like to emphasize that this is my personal guess. I don't know anything about this (some people won't be surprised). Automatically retensioning maximizes the chance of long-term error-free operation. In Info-Explorer, "retension the tape" is listed as a PERIODIC "systems management task" for the QIC tape drive. Under the tctl command, it says that retension "Moves the tape to the beginning, the end, and back to the beginning of the tape. If you have excessive read errors during a restore operation, you should run the retension subcommand. If the tape has been exposed to environmental extremes, you should run the retension command before the save operation." In short, retension should be done when there's something wrong with the tape, and also every now and then (for luck?). Not trusting the user to do this, we appear to have opted for reliability over performance. This is very IBM-like. I tell my customers to 1) turn off automatic retensioning, and 2) remember to retension the tape every now and then. -- Bob Shair shair@chgvmic1.iinus1.ibm.com Scientific Computing Specialist SHAIR@UIUCVMD (bitnet) IBM Champaign
bowman@uiatma.atmos.uiuc.edu () (05/21/91)
In article <7572@spdcc.SPDCC.COM> rbraun@spdcc.COM (Rich Braun) writes: >A question for IBM: why _in the world_ would a company put out a product >which by default would do such a thing? It really makes the system seem >cruddy and slow. I'm sure they do it because it's better for the tape. Retensioning undoubtedly reduces tape i/o errors, but takes longer. If you want to turn it off, you can take the chance. --------------------------------------------------------------------------- Dr. Kenneth P. Bowman, Assistant Professor bowman@uiatma.atmos.uiuc.edu University of Illinois at Urbana-Champaign bowman@uiucvmd.bitnet Department of Atmospheric Sciences 217-333-7105 105 S. Gregory Avenue 217-244-4393 fax Urbana, IL 61801
mbrown@testsys.austin.ibm.com (Mark Brown) (05/21/91)
rbraun@spdcc.COM (Rich Braun) writes: |ken@dillman.austin.vnet.ibm.com () writes: |>Try changing the automatic retension feature of the tape drive to off |>using SMIT. That should save you a great deal of time every time you |>load a new tape. |Thanks for posting this. I got several e-mail replies suggesting the |same thing. | |A question for IBM: why _in the world_ would a company put out a product |which by default would do such a thing? It really makes the system seem |-rich Because new, "un-tensioned" tapes can cause problems. If I were to read the little piece of paper that I throw away after unwrapping a brand-new tape, I'd find something like "please unwind and rewind the tape before use", otherwise known as "preconditioning". Given IBM's predilection for trying to do everything *for* the customer, it's no surprise to me that retensioning is the default. Perhaps it shouldn't be.... Mark Brown IBM PSP Austin, TX. (512) 823-3741 VNET: MBROWN@AUSVMQ MAIL: mbrown@testsys.austin.ibm.com OR uunet!testsys.austin.ibm.com!mbrown DISCLAIMER: Any personal opinions stated here are just that. A straw vote only shows which way the hot air blows. -- O. Henry
moody@snap.austin.ibm.com (05/21/91)
In article <7572@spdcc.SPDCC.COM> rbraun@spdcc.COM (Rich Braun) writes: >A question for IBM: why _in the world_ would a company put out a product >which by default would do such a thing? It really makes the system seem >cruddy and slow. I believe that the folks who make the tape drives felt like the error rate would be too high if the tape didn't get retensioned(sp?). In other words, manufacturing insisted that it be done. At least that's the story I heard back when software developement folks were asking the same question about a year ago. Just in case I'm wrong, remember, I speak only for myself. DISCLAIM, DISCLAIM... >-rich -- James Moody aixnet:moody@moody.austin.ibm.com Personal Systems Programming Austin VNET:MOODY@AUSVMQ AIX Field Support - Level 3 internet:moody@aixwiz.austin.ibm.com