errors 105 and 106

Question

Level 2

257 points

errors 105 and 106

I'm getting an increasing number of 105 and 106 errors from my external HD. I have it hooked up to the Performa by SCSI. The purpose of the external HD is backup. For a couple of years now my Retrospect 4.0 backup file has failed with one or the other error when the file got too big, say, after the second or third backup. Tonight I tried something different. The external HD is quite a bit bigger than the main, internal HD (which is partitioned), and I tried just dragging one partition's icon over to the external HD. The copying processs stopped after transferring maybe a fourth of the partition. (I'm guessing that partition currently holds 80 megs or so.) I found that the transfer could work if I dragged individual folders over, but that's slow. Does anyone have any guesses as to why transferring large amounts of data across SCSI causes a stall? TIA

Performa 6400, Mac OS 9.1.x

Posted on Apr 9, 2010 6:02 PM

Reply

Answer 1

Apr 10, 2010 11:29 AM in response to R_55a

On a quick search, +105 and +106 and -105 and -106 do not seem to be valids Apple System Error Codes.

I use Retrospect under Mac OS X (using version 6). Those numbers remind me of 205 (lost connection to backup device) and 206 (data errors on backup device). The retrospect manual suggests in these cases you check for proper Termination and keep your SCSI cables under 20 feet. You ARE using a discrete External Terminator on the second SCSI connector of your External Drive case, right?

I would also add that you may have Bad Blocks on your SCSI Drive. You can run a Test Disk pass using Drive Setup. Test Disk is nominally non-destructive, provided your Drive is not about to crap out anyway.

You can re-write every block and substitute spares, but it takes an Initialize with Write Zeroes (all data will be lost) followed by a Test Disk pass. Both steps are somewhat time consuming and losing all your data is certainly inconvenient.
------

OS 9 Finder builds the tables in memory to do small moves and copies, and for large moves and copies it, well, does the same thing. That is a tragic flaw in OS 9 Finder -- it will crash if you ask it to move huge amounts of data at once.

Message was edited by: Grant Bennet-Alder

Reply

Answer 2

R_55a Author

Level 2

257 points

Apr 10, 2010 4:27 PM in response to Grant Bennet-Alder

Yes, the external HD does have an external terminator on it. Do those things ever fail?

I had thought about the bad-block problem. I may well have to reformat the HD.

Thanks for the info about the problem with transferring large amounts of data in OS 9. That sounds exactly like what happened.

Reply

Answer 3

Apr 10, 2010 4:49 PM in response to R_55a

Older External Terminators are just a bunch of resistors. Newer "Active" Terminators have a Voltage regulator inside, but it is not under any stress. The best active terminators (if you decide to get another) have an LED that shows when Termination is working properly. In general, Terminators are not a likely source of failure.

The reason I bring up the Terminator at all, is some folks have their SCSI chains (sort of) working without an external Terminator. With that shaky set-up, when you go to move a lot of data, that's when it will fail on you.

You can run Drive Setup's "Test Disk" without reformatting to get an idea of whether you have bad blocks or not. It takes a while, so don't start it when you have other work to do.

I get 206 data errors when I am using Retrospect and backing up to CD and DVD using anything less than excellent media. If you are making backups, Retrospect wants that data to be good, so it reports any errors. For CDs and DVDs, it means you just burned another coaster.

Reply

Answer 4

R_55a Author

Level 2

257 points

Apr 12, 2010 5:16 PM in response to Grant Bennet-Alder

I haven't had time to run Disk Test. But here is a sample log from an attempted back-up:

+ Executing Immediate Backup at 4/9/10 4:34 PM
To StorageSet StorageSet HD…

- 4/9/10 4:34:52 PM: Copying a) office
Trouble writing: “1-StorageSet HD” (15961600), error 206 (drive
reported a failure: dirty heads, bad media, etc.)
4/9/10 4:35:31 PM: Comparing a) office
4/9/10 4:35:54 PM: 1 execution errors
Completed: 256 files, 15.2 MB
Performance: 30.3 MB/minute (24.6 copy, 39.5 compare)
Duration: 00:01:02 (00:00:02 idle/loading/preparing)

+ Executing Recatalog at 4/9/10 4:36 PM
To StorageSet StorageSet HD…
Bad StorageSet Header found (0xffffffff at 15,589)
4/9/10 4:36:53 PM: 1 execution errors
Duration: 00:00:03

+ Executing Recatalog at 4/9/10 4:41 PM
To StorageSet StorageSet HD…
Device trouble: “1-StorageSet HD”, error 105 (unexpected end of data)
4/9/10 4:41:14 PM: 1 execution errors
Duration: 00:00:02

+ Executing Recatalog at 4/9/10 4:41 PM
To StorageSet StorageSet HD…
Device trouble: “1-StorageSet HD”, error 105 (unexpected end of data)
4/9/10 4:41:23 PM: 1 execution errors
Duration: 00:00:02

Does any of that look diagnostic? Thanks.

Reply

Answer 5

Apr 13, 2010 9:21 AM in response to R_55a

You may have directory damage on that drive, from the unexpected end of data. An examples is cross-liked files where the directory lost track of the next block. You can check with Disk First Aid.

You may have at least one bad block, from 206 data error. Retrospect has a lot of support for tape drives, so it has a tape drive slant in its error messages. Disk drives do not get dirty heads, but they could have developed a Bad Block.

Reply

Answer 6

R_55a Author

Level 2

257 points

Apr 13, 2010 3:08 PM in response to Grant Bennet-Alder

Well I ran the Test Disk function of Drive Set-up. And nothing happened. At the end of the test, the application stopped. Is it supposed to give me a report of some sort?

I've been running Disk First Aid and Disk Warrior for years. Maybe 40% of the time, the former finds some problem, mostly a problem with volumes, but almost as often "Mount Check finds a serious error." Sometimes Disk First Aid can repair that serious error, but, when it does not, that is when I run Disk Warrior. Other times Disk First Aid reports that the HD is "OK."

One other item to note: whenever Retrospect hits a problem and stops working, I always hear a comparatively loud speeding up of the HD. The increased spinning drops off after about two seconds.

I experimented with the terminator. It is, I believe, an active one. At least, it has an illuminated green LED on it. I had had the terminator on my old, occasionally active DAT drive at the end of the SCSI chain. I disconnected the SCSI connection to the DAT and moved the terminator to the now empty port on the HD case. There was no change in the symptoms of my Retrospect.

Reply

Answer 7

Apr 13, 2010 3:25 PM in response to R_55a

Is it supposed to give me a report of some sort?

You usually get one of:
No Bad Blocks encountered during test -OR-
Bad Blocks encountered during test

Other than that, there is no report.

Maybe 40% of the time, the former finds some problem, mostly a problem with volumes, but almost as often "Mount Check finds a serious error."

You should not see these sorts of problems on a properly-functioning drive, except very, very rarely. It may be time to initialize with Write Zeroes, followed by Test Disk (allow writes). These two together will substitute spares for any Bad Blocks detected during the process. Both are needed (together) to substitute spares in a SCSI drive.

Reply

Answer 8

R_55a Author

Level 2

257 points

Apr 22, 2010 3:02 PM in response to Grant Bennet-Alder

I ran "Test Drive" again, and it reported no problems. Now what?

Reply

Answer 9

Apr 22, 2010 3:41 PM in response to R_55a

If it were mine, I would run the "Write Zeroes" followed by "Test Disk". At the conclusion of that pair, you know every block on a SCSI drive is good. (and removing some uncertainty is what you are after, in my estimation.)

Reply