thanks roger

and it's ok to be drunk :-)

so i should do it for :

/dev/hda
/dev/hdb
/dev/hdc

?

If you want to do both drives then hda + hdc. No hdb.

why not hdb? i have seen hdb among the errors

The Mk1 had two IDE channels so your two drives ended up as hda and hdc.

The Mk2/Mk2A had 1 IDE channel so your two drives ended up as hda and hdb. However to make it easier, the device nodes in /dev were changed around slightly so you could still refer to the 2nd drive as /dev/hdc.

Yeah. Blame me (and Linus) for the skip naming, and blame Hugo for his twisted simplification (of empeg s/w) in taking advantage of it.

The kernel refers to the second drive in a MkII/MkIIa as hdb, but the non-kernel software uses /dev/hdc for the same unit. This is a peculiar, yet normal, sort of thing for a POSIX style operating system to be able to do.

EDIT: The purists would prefer that we should have used 3,0 and 3,16 in the kernel, instead of hda and hdb. But I argued against that, as being rather unfriendly for common uses and users.

Cheers

funny thing is that when i zero fill hda at some point it goes crazy and are writing stuff so fast that i cant read it.

if i reboot the player it hits me with a kernel panic

doing the disk build stuff and everything is normal again, but it is only stress testing hda....

anyway cat /dev/zero > /dev/hdb can't be done, it tells me that /dev/hdb is read only

a little challange for mark: build a disk test thing in to hijack, with options to test and repair disk, and also stress all installed drives

maybe as a special version that only do that.. or maybe a "bug fix" hijack that can be used for fixing bugs and repairs ?

As I said above, don't do hdb.

There is a finite number of spare blocks available on the drive to remap bad blocks. If your drive is really badly broken then you'll have to replace it.

but there are errors on hdb... guess i can take them out of the player and zero fill them on my laptop

Errors on hdb get fixed when you "cat /dev/zero > /dev/hdc". If you open the player, count the drives carefully! there will be at most two, not three

I think, Mark, your clarification itself contained a typo: didn't you mean:

..?

Peter

there are 2 drives, i installed the second one myself,

also i used some clear silicon glue on the ide connector on the drives, this becuase the silicon stays flexible and is easy to remove.. i used the clear type as it is the most weak of all types.. you can get greay, white and red types...

i prefered silicon to hot melt glue do to that vibrations could break of the hot melt glue as i cools down and get hard

Hi,

Sometimes that type of silicone RTV will leech into the connector mating area and contaminte the gas-tight seal of the pin and socket.

Make sure this is not the problem too.

Ross

i'm sure it's not the problem, i only applied so little that it would only touch the surface of the connector on the drive and on the ide cable

Silicone uses an acid solvent, which can corrode contacts.

Yup. Typo now fixed by an EDIT in the original post.

Thanks.

I'd have preferred Solaris-style c0t0 and c0t1 namings. It's so easy to figure out what's what under Solaris in a way that it is not under most other Unices, Linux included.

so it's 100% safe as long óne does not put it on the contact pins

Vapours.

i did some heavy vent. wkile i dried up

anyway after some diskbuilds and zero fills it all comes down to this:


empeg:/empeg/bin# cat /dev/hda > /dev/null
empeg:/empeg/bin# cat /dev/hdc > /dev/null
hdb: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
hdb: read_intr: error=0x40 { UncorrectableError }, LBAsect=115459753, sector=115459753
end_request: I/O error, dev 03:40 (hdb), sector 115459753
cat: /dev/hdc: Input/output error
empeg:/empeg/bin#

i will be fixing the temp sensor later, so the unit will get a very close inspection while i have it open

as for the Silicone i have only used it on the drive end of the IDE cable... i could never dream of using the stuff on the mainboard.. there is no need to either as the connector is more a theight fit than on the drives

just loaded ver. 3 alpha 11.. i feel like an idiot, i can't find the auto eq feature

do i have to plug in the mic to have the menu option displayed?

The smartctl program is the only real way to know if the drive is the issue.

I've placed a binary copy (for the empeg) on the Hijack site (links at top of this page and that page.

Cheers

how do i use the tool? sorry for asking stupid

Use it the same as you would on any other Linux system.
The man page is here, courtesy of Google.
The normal sequence is to do something like this:

## Enable the S.M.A.R.T. monitoring in the drive firmware:
smartctl -s on /dev/hdc

## Check current drive status. "-a" means "all", or "everything":
smartctl -a /dev/hdc

Then scour the output looking for error reports.
(and post the complete output here for us to help with)

If you don't see any errors, then do the "cat /dev/zero > /dev/hdc"
thing again (sorry), and this time the drive will record errors
for later retrieval with smartctl -a

Cheers

i'm a windows man so i don't know much about linux, can you make a very quick step by step?

The FAQ has a step-by-step here, but I haven't sanity-checked it in a long time, so it might be outdated. Worth a try though.

i will say it does work or not

damn.. i cant connect to ftp... :-(
newest hijack installed

still no go at the http or ftp in hijack

i installed ver 2.01 kernel and newest hijack

Check the ethernet connections. Go to the "about" screen on the empeg front-panel, and see what IP address it is using.

You know the drill by now.. post a serial log here along with the IP address info.