I've had to have a two day break from this thread because I decided to try my two hard drives on the JMicron interface and the DVD drive on the Intel interface. The mainboard manual says the JMicron interface supports up to 2 devices.
The short version of the story is that I ended up with the file systems on both drives comprehensively trashed! Linux is normally very good at recovering from file system errors but it couldn't do anything with these.
The long story: After swapping interfaces I booted the machine using a Knoppix v5.1 CD, because I'd need to change device names in a couple of files, but went into the BIOS first. The DVD registered as the first primary drive in the first main menu screen as expected but it seems very odd that once in the BIOS any devices connected to the JMicron interface are not listed anywhere. Drives are set as Master and Slave.
They are shown, briefly for about 5 seconds, just before entering the BIOS and it was here that I noticed that only one hard drive was shown after the heading 'HDD:' but the name that followed was just a random collection of letters and numbers, the font was also broken up (characters had gaps in them), and that line of text was flashing.
I left the BIOS and continued with the boot into Linux and tried mounting one of the partitions. On seeing the message that 'the filesystem has errors' I didn't go any further but closed down the system.
I then swapped the cables back to their original positions and tried to boot but it hung a short way into the process. I decided to swap the primary drive and DVD drive back into my old machine and boot from the same Live CD to examine the hard drive. The contents were basically rubbish! I spent quite a lot of yesterday restoring from tape which, as there is a fair amount of data, was a rather slow business. The tape was up to date as far as early yesterday morning so I lost a couple of unimportant emails but that was it.
So I'm not over impressed with the JMicron interface so far.
On Thu, 2007-02-15 at 21:43 +0000, Barry Samuels wrote:
They are shown, briefly for about 5 seconds, just before entering the BIOS and it was here that I noticed that only one hard drive was shown after the heading 'HDD:' but the name that followed was just a random collection of letters and numbers, the font was also broken up (characters had gaps in them), and that line of text was flashing.
I have had that before.
Apart from possibly having a broken IDE controller there, I'd suggest two other possible causes.
Either a mistake was made on setting the master/slave jumpers so that both drives appeared as one or
Are you sure your cables are good ?
Are they the proper UDMA spec 80 conductor cables ? Interfaces should fall back to < UDMA mode 2 but they detect this by looking at signalling errors on the interface so the detection of a 40 conductor cable is not 100% guarantied (I have even seen vice versa where an interface insists that it has a 40 conductor cable when it really has an 80 conductor one)
Also the 80 conductor cables are much more fragile than the old 40 conductor ones...Having learnt the hard way (and given the insignificant cost) I tend to replace any that look anything less than perfect on machines that pass through me as a matter of course.
Thank goodness you backed up huh
I sent this some hours ago from the wrong identity so it went into the Moderator queue. It has not yet appeared on the list so I'm sending it again. If the other one appears then I apologise for the duplication.
On 15/02/07 22:49:31, Wayne Stallwood wrote:
On Thu, 2007-02-15 at 21:43 +0000, Barry Samuels wrote:
They are shown, briefly for about 5 seconds, just before entering the BIOS and it was here that I noticed that only one hard drive
was shown after the heading 'HDD:' but the name that followed was just a random collection of letters and numbers, the font was also broken up (characters had gaps in them), and that line of text was flashing.
I have had that before.
Apart from possibly having a broken IDE controller there, I'd suggest two other possible causes.
Either a mistake was made on setting the master/slave jumpers so that both drives appeared as one
Mistake? Moi?
No - the settings for both drives haven't changed. One is set to master and the other to slave. They work perfectly on my old machine and they work perfectly on the Asus provided they are connected to the Intel IDE interface and not the JMicron.
Are you sure your cables are good ?
Well I'm still using the same cable (on my old machine at the moment) which I was also using on the Asus (Intel) and I haven't, yet, had any other problems.
Are they the proper UDMA spec 80 conductor cables ?
Pass! How do I tell? Why should there be 80 cables when there are only 40 pins?
Interfaces should fall back to < UDMA mode 2 but they detect this by looking at signalling errors on the interface so the detection of a 40 conductor cable is not 100% guarantied (I have even seen vice versa where an interface insists that it has a 40 conductor cable when it really has an 80 conductor one)
Also the 80 conductor cables are much more fragile than the old 40 conductor ones...Having learnt the hard way (and given the insignificant cost) I tend to replace any that look anything less than perfect on machines that pass through me as a matter of course.
What do regard as perfect? No sharp bends, no bends at all, no fraying?
Thank goodness you backed up huh
I've adopted the belt and braces approach. I have had a SCSI DAT drive for many years now, not the same one as they seem to fail regularly, but in the recent past I bought another IDE drive to mirror (not raid) my main drive on the assumption that should the main drive fail I just swap drives and continue on my merry way.
I have seen many comments on the Internet that using a second drive is the cheapest way of keeping a backup these days. If I'd relied on that idea I wouldn't have had anything to restore.
On Fri, 2007-02-16 at 18:33 +0000, Barry Samuels wrote:
Pass! How do I tell? Why should there be 80 cables when there are only 40 pins?
The 80 conductor cables have every other conductor connected to the ground pin. so compared to a 40 pin cable they go pin1 ground pin2 ground etc. The ribs look much finer than a 40 conductor one...if the construction of the cable itself looks the similar to what you find on a floppy drive then you have the wrong cables.
The reason for the extra wires is to provide shielding between the data lines...once IDE got to 33 MB/s with UDMA mode 2 they started having real problems with reflections/crosstalk in the cable, anything beyond that requires the extra ground wires to work properly.
This is the same reason why you can't really mess about with IDE cables (like cutting down a master/slave cable to make a shorter single device one) Doing things like that over 66MB/s yields very unpredictable results, even using the connector nearest the main board and leaving the furthest one away unconnected can cause problems with some drives/controllers.
What do regard as perfect? No sharp bends, no bends at all, no fraying?
Well bends that were made once to neatly arrange the cable in the machine are fine...dents and creases that have resulted from rough handling or the machine being pulled to bits a lot are usually enough to make me want to change it.
Unless yours look like they have been screwed up in a ball at some point they are probably fine.
It's just one of those things where it is cheaper for me to spend a few quid per machine changing even slightly suspect cables than take the risk of having to have the machine back on the bench again under post repair warranty with some hard to trace fault.
On 2/16/07, Barry Samuels bjsamuels@beenthere-donethat.org.uk wrote:
I've adopted the belt and braces approach. I have had a SCSI DAT drive for many years now, not the same one as they seem to fail regularly, but in the recent past I bought another IDE drive to mirror (not raid) my main drive on the assumption that should the main drive fail I just swap drives and continue on my merry way.
I have seen many comments on the Internet that using a second drive is the cheapest way of keeping a backup these days. If I'd relied on that idea I wouldn't have had anything to restore.
Relying on only one backup device is risky - I bet you have more than 1 DAT tape. You even mention the DAT drives fail regularly!
A number of removable harddrives (IDE, USB, or Firewire) used in rotation would give you the same strengths as a tape library, with great capacity per media item. Using an IDE (PATA or SATA) removable caddy will be faster than tape too. USB and Firewire harddrive enclosures usually only let you attach one device to the IDE connector inside, so there is no chance of Master/Slave screw ups or motherboard IDE weirdness.
Regards, Tim.
On Thu, Feb 15, 2007 at 09:43:04PM +0000, Barry Samuels wrote:
I've had to have a two day break from this thread because I decided to try my two hard drives on the JMicron interface and the DVD drive on the Intel interface. The mainboard manual says the JMicron interface supports up to 2 devices.
All IDA/PATA interfaces support two drives I think, it's inherent in the interface specification.
The short version of the story is that I ended up with the file systems on both drives comprehensively trashed! Linux is normally very good at recovering from file system errors but it couldn't do anything with these.
The long story: After swapping interfaces I booted the machine using a Knoppix v5.1 CD, because I'd need to change device names in a couple of files, but went into the BIOS first. The DVD registered as the first primary drive in the first main menu screen as expected but it seems very odd that once in the BIOS any devices connected to the JMicron interface are not listed anywhere. Drives are set as Master and Slave.
Are you sure the JMicron is supposed to work at the same time as the other PATA interface?
They are shown, briefly for about 5 seconds, just before entering the BIOS and it was here that I noticed that only one hard drive was shown after the heading 'HDD:' but the name that followed was just a random collection of letters and numbers, the font was also broken up (characters had gaps in them), and that line of text was flashing.
Yes, this is the interface BIOS I think, mine does that too, the other disk interfaces make a momentary appearance before the main BIOS pops up too.
I left the BIOS and continued with the boot into Linux and tried mounting one of the partitions. On seeing the message that 'the filesystem has errors' I didn't go any further but closed down the system.
I then swapped the cables back to their original positions and tried to boot but it hung a short way into the process. I decided to swap the primary drive and DVD drive back into my old machine and boot from the same Live CD to examine the hard drive. The contents were basically rubbish! I spent quite a lot of yesterday restoring from tape which, as there is a fair amount of data, was a rather slow business. The tape was up to date as far as early yesterday morning so I lost a couple of unimportant emails but that was it.
So I'm not over impressed with the JMicron interface so far.
It sounds more like a geometry issue or something like that (disk geometry that is).
On my system the JMicron interface is the *only* IDE/PATA interface, all the others are SATA so I don't have the possibility of moving drives from one PATA interface to another.
-
Barry Samuels -
Eur Ing Chris Green -
Tim Green -
Wayne Stallwood