Hello,

my PegII has a sii3114 SATA interface, connected is a 640GB harddrive. Recently I discovered temporary system freezes while reading data from that hard disk (some partitions, mostly SFS2 formatted, the biggest one 480 GB).

After some seconds of no operation/total freeze the read process went on without an error and all user actions seemed to be buffered, so all my desparte clicks were processed.

An immidiate call of Dumpdebugbuffer reports:

[sii3114ide/ata_rw_blocks] unit 1 lba28 read error 73, retries left : 4

What's going on here?

@cha05e90

I had the same on my sam with a samsung spinpoint 360gb drive and dumpdebugbuffer reported the same line hundreds of times. I replaced the drive and the problem disappered.

@Severin

Had same issue. New HDD fixed it.

@MickJT

Ok, thanks for your infos - so it seems to be a hard drive issue - I hope the current one will live long enough to make a full backup to the new one...

@cha05e90

Also I have problems with the sii3114ide.device in this case with the SATA DVD RW drive(before with IDE DVD), when I put one music CD or one DVD and play them with PlayCDA or DvPlayer my Sam is freezes and appear the message with problem with this device, this device isn't bad with bugs?

@Templario

It could be the drive, they also suffer from companies ignoring the standards, try another one if you can. at least they are cheap to replace now, iirc my current one (a sony optiarc AD-7170S) was less then £20. I've always used nec drives before this one and never had a problem with them.

@Severin
I've been a fan of Lite-On drives myself (they OEM for Toshiba IIRC). Their firmware is excellent.

The worst drives I've tried have been LG drives (due to bad firmware).

The mechanical part of the drives don't really matter. It is all about the quality of the firmware.

@ssolie

Lite-On drives in europe aren't brilliant, I've had 2 that have died after writing a couple of hundred dvd's. although they were ide rather than sata.

@cha05e90
Definitely sounds like your HD is dying - get it changed ASAP! Best to avoid using it until it is replaced.

@Severin
I've used Lite-On drives without trouble, so you may have been unlucky...

@Severin

I've had the same experience but with Lite-On CD drives and it was far from writing to CD's one hundred times. Replaced them both with LG writers which are still going strong.

@ChrisH

Ok - new hard drive attached, prepped and partioned. Full backup done. Puh! But before the problematic HD will hit the bin, I'll install some new SATA cable. Just to be sure...

BTW: While doing my backup yesterday I noticed, that the error *seems* to show only (sporadically) when accessing my big 480GB-SFS/2-partition. While working on the other partition on that HD (~100GB, SFS/0) I did *not* notice any problems - this might be by coincidence, but nevertheless: Is there anything known about partition size, SFS/2 et al. that might lead to this problematic behaviour?

@Severin

+1 I've bought a NEC DVD-burner a couple of years ago for my A1 and the pricetag then was 30 euro. I can just say that it has worked without a single issue.

@cha05e90

"before the problematic HD will hit the bin" maybe you should try to first re-format / partitioning / initialize / low-level format it?

If it were out of order, it shouldn't work at all and not only sometimes not. So, my guess is that it's the filesystem's fault (maybe broken?). Harddrives should also be able to mark out defective sectors which will not be used any further afterwards (as far as I know).

I wouldn't give up so soon
However, backups are a wise decision, I guess

nexus

@cha05e90 Quote:
My guess:

Some semi-random locations on your HD are dying, with lots of errors. When the HD sees these locations it has to stop what it is doing & perform a lot of error correction. So you get the data, it just takes a long time, and these are the freezes you get.

These semi-random locations are more likely to be found in your big partition, precisely because it is so big; imagine throwing a dart at stamp next to a postcard - you are more likely to hit the (bigger) postcard! These locations may also be biased to one part of your HD, which happens to be used by your big partition. So no need to worry about it being specific to SFS!

@cha05e90

Quote:

Don't bin it, just chuck it in a pc or mac if you have one, my problem drive is working fine in my mac now.

NEVER low level format a sata/ide drive that was only ever needed for old (pre 1990) scsi and ide drives when amigas and seperate hard drive controllers that used custom formats. Don't even attempt to full format a hard drive, big drives can take many hours to format and will probably fail before they've finished (possibly due to the heat generated by constantly accessing for hours on end). just quick format, any bad sectors are automatically remapped by the drive itself (which doesn't help if there's already data on dodgy sectors though).

@Severin

Quote:
No, I usually show great and deep-hearted loyalty to all technical things, that surround me! I am an Amiga User!

After having a (hopefully) full functional backup now, I'll give that 640GB-hottie every chance it needs to be a well of joy...

Question: Maybe I'll try some other file systems and/or different partitioning. What are the borders for SFS/0 (GBytes, blocksize)? Is it true, that JFXS needs a specific blocksize?

@Severin Quote:
Only half-true: HDs only remap bad blocks *when they are written to*.

Quote:
This probably appears to work *because* you format it & start from scratch. Thus any bad blocks are written to before being read, thus causing them to be remapped, thus stopping the HD from having to perform slow error-correction every time it reads them.

Thus it may appear to start working again - as long as you don't get more bad sectors. (Depends on the failure mode, but I personally wouldn't risk it.)

It's nothing to do with pcs or macs being better! You could do the same trick on an Amiga, if you feel like playing Russian Roulette with your data :)

@ChrisH

Hm. Not many possibilities left...

- Changing SATA port.
- Changing to a new, shiny SATA Cable.
- Complete re-partitioning/permutation of file systems.
- successive down grading UDMA, testing with PIO access.

...nothing helped, still sporadical sys freezes with the mentioned error in the debug buffer. So'll put in a new hard disk permanently. Hopefully this will *not* show the same behaviour, 'cos then it might be the SATA controller, bad RAM, funny PCI whatsoever...

Oh my dear...

New hard disk - same problem:

[sii3114ide/ata_rw_blocks] unit 1 lba28 read error 73, retries left : 4

Now I'm somewhat lost - what shall I do now? Has someone an idea what I should do next? I suppose now it's *not* a hard disk error but something else, but I don't know where to search the cause now...

@cha05e90

Remove a stick of memory? (If that's possible for you to do).