Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
And then there was the tape backup that functioned fine until you needed it at which point it ate the tape
...or you went off to get a cup of coffee while it was backing up.
Anybody work with a tape archive robot that changed tapes like a juke box?
Our consulting company had one customer with one of our guys as full time system manager responsible for system operation and backups. Huge system, lots of money for things like big tape robots managing dozens of tapes automatically. He complained all the time that those 8mm 4GB data tapes were continually failing to perform backups. Tapes would jam or abort in the middle of a job with verification failures, or the robot mechanism would fail or the backup itself would be incomplete or just fail because the moon was blue.
I myself eventually learned that tape backup to those tiny tapes was Russian Roulette even on the comparatively small systems I was responsible for. Since then and to this day I backup by duplicating files onto hard drives or CD or DVD or static memory. I won't trust modern tiny tapes ever again. The old 1/2 inch, 9-track, 2400 foot reel, drives from the 60s and 70s were great but considering that they only held a few dozen megabytes of data and were being handled by $10,000 drives the probability of failure was not unexpectedly low. But damn, you felt like a real computer person when threading those tape reels and raising the vacuum door and watching the tape be sucked into the vacuum columns. (* Ooh ooh, butch! *)
Speaking of tape storage... anybody ever work with an IBM "Noodle Picker"? I never saw one but have always been fascinated by the concept. I think they were the inspiration for the musical keyboard instrument called the "Mellotron" used in 70s rock bands like "The Moody Blues". Strips of tape that were picked from their hanging rack and zipped past a read head. The Mellotron could read a few seconds of recorded sound but the IBM "Noodle Picker" would read and write data from/to the short tape strips. It was essentially random access on tape. 8-0
We shipped our last IBM 3420-II tape drive back to Comdisco in the late '90s, shortly before they went bankrupt (too many shops returning obsolete gear that they couldn't lease any more). As I understand it, the Chicago warehouse had receiving on one side of the building and shipping on the other. When the open-reel drives came back they'd check to be sure it was the correct serial number and that it was packed and certified by IBM as being maintenance qualified. And then 4 or 5 of the guys would start pushing it as fast as they could across the floor, with the object being to see how far out into the dumpster they could get it.
Never worked with one - the insurance companies loved them, as each tape strip held 2 KB; something like 40 tapes to the cell, and 6(?) cells to the drive - with special sand-filled dummy cells to use for balance if you didn't have that many tape cells. (IBM 2321 was the model number, IIRC).
Never worked with one - the insurance companies loved them, as each tape strip held 2 KB; something like 40 tapes to the cell, and 6(?) cells to the drive - with special sand-filled dummy cells to use for balance if you didn't have that many tape cells. (IBM 2321 was the model number, IIRC).
WOW! Almost a half megabyte of random access storage. More than anybody would ever need! (* eye roll *)
I remember in the early 80s Seagate sold a 10 Megabyte hard drive and a 20 Megabyte hard drive and it turned out they were both the same hard drive. It was cheaper just to have one assembly line making the same hard drive than two different lines making two different drives. So there was a jumper on the hard drives. Remove it and it would activate a second set of hard drive heads to read the other side of the platters. Once word of this got out the 10 Megabyte drive was just pulled from the market.
IBM also had a Quietwriter printer line that had a low priced one that printed at 80cps and a high speed one that printed at 160cps. If you bought the low speed on you could get it upgraded latter for a fee and a technician would come an install the upgrade. The upgrade was simply moving the belt that drove the print head from one set of pulleys to another and then replacing the brand sticker on the front of the printer.
...had one of those 10/20MB Seagates.
Interesting about the printer.
Yeah, they probably figures most people were not savvy enough to figure things out like that themselves.
Yeah, people were still having trouble with their slide-out cup holders, and finding the "any" key. (* eye roll *)
WOW! Almost a half megabyte of random access storage. More than anybody would ever need! (* eye roll *)
Yeah - I was wrong. :-) That's what I get for relying on 40-year old memories. :-)
The 2321 had 10 cells with 200 strips of tape 2.25 inches wide and 13 inches long - for a total storage of 400 MB. And given that the other storage device on early 360 series systems was the 2311 removable disk pack, with a whopping 7.25 MB per pack - you can see why Insurance companies (and some banks) loved them.
References: Wikipedia and IBM Archives - note that they disagree on the number of cells on a unit, but agree on total data capacity.
OK, I believed your first figures but 400MB sounds much better.
I worked on an IBM 1130 that had the removable IBM 2315 disk cartridges that had about a half megabyte of storage. I'd hoped that the Noodle Picker had more storage.
Ah, 50 year old memories. Picture it, 1967, Melbourne, Florida. The newly renamed, tiny college (Florida Institute of Technology) had acquired their first computer. The IBM 1130. I was a Sophomore and took my first and only formal programming course (FORTRAN IV) then started working in the computer room. The computer area was in the basement of one of the new dormitories. The computer was in the room in the far half of the basement. The one and only TV (b/w) on campus and some folding chairs were in the room in the front half of the basement. Friday nights, every geek on campus would gather to watch new episodes of Star Trek (the original) while the computer hummed and clacked and chattered in the background.
FIT had the first accredited course program in "Space Science" many of the professors were engineers at the space center. It was indeed a school for rocket scientists. FIT is now a big player in several fields of ocean research, bio medical research, aeronautics, etc.
I recently noticed a photo of a night time rocket launch and had a feeling that I knew where that picture had been taken. After a little close examination I realized that it was taken from the roof of the tallest building on the FIT campus (the "Science Tower". Originally it was on the southern end of the campus but with land donations and purchases now sits in the middle of the long narrow primary campus.
http://www.fit.edu/gallery/
My cost the year I started (1966) was $800 per quarter for 3 quarters and that included tuition, room & board. Current yearly price for tuition, room & board is over $50,000 what a change 49 years makes! 8-O I've never been to an alumni event but I think next year I'll go for the 50th anniversary of the school changing its name to FIT as well as being 50 years since I started. A lot of things happened for the school that year.
I worked with two Automated Tape Libraries (ATLs) - neither used quarter-inch or 8 mm tapes. The first was a Storage Tech Timberwolf library with 9 Quantum DLT7000 tape drives at the end; 35 GB native, 70 GB if the data compressed on a 4 inch by 4 inch cartridge; the tape was 3/4 inch. The second was an IBM 3584, using LTO (Linear Tape Option) cartridges, slightly smaller than the DLT tapes. First generation was 100/200 GB and each generation of tapes and drives doubled the density. As I understand it, my successor has upgraded the library to 24 tape drives, all LTO 4 fiber attached. Two bays in the library, with only one pick head - can't remember that total tape capacity, somewhere around 380 tapes. Over the 10 years I worked with it we had 3 or 4 tape drive failures and only 8 or 10 tapes that went bad.
I love hearing about those small setups... (Your "total tape capacity" is our "dangerously low on scratch tapes.")
I love hearing about those small setups... (Your "total tape capacity" is our "dangerously low on scratch tapes.")
We used IBM's TSM (Tivoli Storage Manager, renamed after IBM bought Tivoli and transferred ADSM - Advanced Data Storage Management- to them)- key to TSM is that it handles backups which are held for generations, and only include files that have changed, and archives which are complete file systems retained for a specific number of days. Everything is recorded in a disk database, with pointers to which tape and what position on the tape has the data. Every day a process runs which walks the database, deleting references to expired (by age or cycle count) files.
This, of course, leaves tapes with 'holes' in them - so another process can be configured to 'reclaim' tapes base on percentage used. The reclaim process reads the tape with the 'holes' and appends the valid data to another tape in the storage pool. If the tape in question is an off-site copy, the on-site originals are read to create a new copy - that we would check out the next day, as well as running a list of all off-site tapes that were logically empty and could come back.
In any case -tapes were continuously appended to until full. The key to the functionality was to split data into pools of like information. Our system backups, for example, were in pools determined by system type and retention count. Our database 'backups' (actually TSM archives) fell into several pools, with 8-day retention for the online backups and 21-day for the offlines we ran Sunday just after Saturday midnight.
On the average day we'd check out between 22 and 30 tapes, including two TSM database backups, and have a pull list of about the same size. On the day I retired our available scratch tape pool was stable at about 120 tapes. We found that an Oracle database compresses nicely - on the LTO-4 tapes (800/1600 GB) we were usually getting between 1850 GB and 2 TB - and our big database at the time was only 1.4 TB that we split into 4 concurrent backup streams to 4 concurrent tapes just to get it done in under two hours. Of course we had 5 copies of that database . . .
One of the nicer aspects of TSM is that you don't just write a tape and put it on the shelf for 7 years. Every tape is subject to expiration and reclamation, and you only send copies (technically 'storage pool backups') off-site. I put all our fiscal retention archive data in one storage pool copy - for the off-site; mixed data with a retention of 30 days on up to 7 years. So every month or so, that off-site pool would get reclaimed and the data written to a different tape to go off-site.
And in reference to 'small sites' - back in 1997, when we cut over to SAP/Oracle there were three groups of IBM RS/6000 users: the group that thought 2 GB of data was large; the group running home-grown Oracle databases, that knew 10 GB was BIG - and those of us running SAP on Oracle who damn well knew that 120 GB was small. We had two best-guesses from consultants on how big our system would be at the end of the first year - one was something around 40 GB and the second was "you can't get enough disk on one system for it". We went with the 40 GB, allowed some fudge room, and started with 60 GB on the main system. We hit 120 GB at the nine-month point . . . mirrored, so 240 GB of disk on the system in early 1998. Mostly in 4.5 GB drives.
Uh - yeah, I am/was a professional geek . . . :-)
Yeah, people were still having trouble with their slide-out cup holders, and finding the "any" key. (* eye roll *)
...the keyboard on my workstation i had at my old development job had several blank programmeble keys which had a slipcover into which you could place label. Naturally I had to make one of them the "Any" key.