fuzzy's ongoing Linux server/router question thread

[silverpig]

Okay, the new fstab looks good.

do a

dmesg | grep hdc

 

[fuzzybabybunny]

Originally posted by: silverpig
Okay, the new fstab looks good.

do a

dmesg | grep hdc

Ahhhhhhhh!!!

[ 86.316668] hdc: task_in_intr: error=0x04 { DriveStatusError }
[ 86.317560] hdc: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
[ 86.317564] hdc: task_in_intr: error=0x04 { DriveStatusError }
[ 86.317960] hdc: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
[ 86.317964] hdc: task_in_intr: error=0x04 { DriveStatusError }
[ 86.367786] hdc: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }

And it goes on and on and on. What happened?

 

[silverpig]

Hmm, I've never seen that error myself, but a little googling suggests it might be you don't have the driver properly installed. What driver are you using? Can you post the output of

lsmod

and maybe do a

modprobe ntfs

If you didn't have ntfs listed in lsmod before the modprobe command, and then see it there after another lsmod, see if you get anything different with that dmesg command again.

 

[fuzzybabybunny]

I don't get this.

I did lsmod, and it showed ntfs under module, 101376 under size, and 2 under Used by the first time before doing modprobe ntfs.

Then when I go to highlight the lsmod output to copy to this message, it hard locked again in the middle of highlighting. The mouse pointer still moves, but nothing responds.

 

[silverpig]

So ntfs is loaded. That seems alright.

But if you say it locks and nothing responds, but the mouse moves, then that just means X crapped out on you. If it does that again, try doing a

<ctrl><alt><bksp>

to kill the X server. Also, you can switch to a terminal by pressing

<ctrl><alt><F2>

or F3, F4...

 

[Nothinman]

[ 86.317560] hdc: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }

I would run the manfacturer's drive test stuff on that drive, there's a chance the drive is dying.

I did lsmod, and it showed ntfs under module, 101376 under size, and 2 under Used by the first time before doing modprobe ntfs.

If it's in use then something NTFS is already mounted.

 

[fuzzybabybunny]

Originally posted by: silverpig
So ntfs is loaded. That seems alright.

But if you say it locks and nothing responds, but the mouse moves, then that just means X crapped out on you. If it does that again, try doing a

<ctrl><alt><bksp>

to kill the X server. Also, you can switch to a terminal by pressing

<ctrl><alt><F2>

or F3, F4...

Thanks, but I already tried refreshing X, and the keyboard commands don't respond.

 

[fuzzybabybunny]

Originally posted by: Nothinman

[ 86.317560] hdc: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }

I would run the manfacturer's drive test stuff on that drive, there's a chance the drive is dying.

It was working fine before 🙁

*tear*

 

[Nothinman]

It was working fine before

That doesn't mean it can't die, drives have moving parts that wear out eventually. And I said it's a possibility, the manufacture's SMART tests will give you an idea whether it's really happening or not.

You never did tell us whether this was the 64-bit or 32-bit version of Linux. And you should try the stuff without X running so if the kernel does oops you get the see the output.

 

[silverpig]

Oh, can you log in remotely via ssh?

Whenever I mess around with one of my systems (often) I almost always mess something up. Having a laptop to ssh into my main box to fix it when my keyboard and mouse stop responding does wonders.

 

[fuzzybabybunny]

Originally posted by: Nothinman

It was working fine before

That doesn't mean it can't die, drives have moving parts that wear out eventually. And I said it's a possibility, the manufacture's SMART tests will give you an idea whether it's really happening or not.

You never did tell us whether this was the 64-bit or 32-bit version of Linux. And you should try the stuff without X running so if the kernel does oops you get the see the output.

I actually don't have any idea which version it is... I think what I'm going to do is just let it sit for now, go order up a couple of 500GB hard drives, back of the data to a Windows machine, and then start this system up from scratch with newly formatted IDE drives.

 

[xtknight]

Type:

$ uname -a

This will give us some stats, including the kernel version and machine architecture.

Last time I checked ntfs-3g was not 64-bit safe. I compiled it on my 64-bit machine before I realized that but it didn't write to the drive properly.

 

[drag]

Originally posted by: fuzzybabybunny

Originally posted by: Nothinman

[ 86.317560] hdc: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }

I would run the manfacturer's drive test stuff on that drive, there's a chance the drive is dying.

It was working fine before 🙁

*tear*

Yep.

I can be caused by a bad cable, bad driver, bad drive.

(Bad cable could be a cable that is too long. A cable that has been 'rounded'. A cable that has a bad crimp or has been smooshed between metal bits. A cable who may not have both ends seated correctly in the respective devices or the stuff may be to dusty for a good connection. etc)

Those sort of errors will definately cause your system to hang periodicaly. I'd probably make sure that that gets resolved before working on other stuff.


In windows when it sees these problems it will disable DMA access to drives by default. This will cause a much slower drive, but it's safer for the data on the drive.... So it easily could of caused your machine to hang or lockup while running Windows once, but then after a reboot it went away.

So use the manufacturer's utilities to check it. Or you can use SMART support in linux to check it out if the drivers and harddrive supports that sort of things.

to disable dma access you go
sudo hdparm -d0 /dev/hdc

And that may help stabilise your system until you get that issue resolved.

 

[silverpig]

Originally posted by: fuzzybabybunny
I actually don't have any idea which version it is...

uname -m

🙂

And the fresh format is a good idea. I mean, it'll work with what you want, but just in case you decide to write to the drives remotely, it'd be beneficial to use a native linux filesystem.

 

[drag]

Also if that is a cdrom drive you can get funky things like that if you have a disk in there that it's trying to access, but has a error on it.

So if it's a cdrom don't assume that it's a big problem, just remove any media you have in it.


(my parents had a computer with a bad cdrom drive that would cause the system to hang periodicly when running Linux. also on my laptop it (used to) hang momentarially when it encountered funky media. Both times it would print errors similar to what you posted.)

 

[Nothinman]

I actually don't have any idea which version it is...

But you downloaded and install it it so was it the AMD64 or the i386 version?

I think what I'm going to do is just let it sit for now, go order up a couple of 500GB hard drives, back of the data to a Windows machine, and then start this system up from scratch with newly formatted IDE drives.

That's the only real solution anyway, I didn't think you were going to use the NTFS volumes for anything more than a data migration.

In windows when it sees these problems it will disable DMA access to drives by default. This will cause a much slower drive, but it's safer for the data on the drive.... So it easily could of caused your machine to hang or lockup while running Windows once, but then after a reboot it went away.

Linux will do the same thing, once there's a DMA transfer error it'll get knocked down to PIO. I guess it's possible he was just being impatient though and thinking it had hung when it really hadn't.

 

[drag]

Ya that's what makes difficult to troubleshoot problems like this on forums.

It's very difficult because he doesn't know the common commands and terminology so then it's difficult to communicate to us what is going wrong.

It's also tough when you get hit by couple problems at the same time.. ie: something is causing hanging on his computer + Ubuntu's disk gui app is mangling the mount points.

tough. Could be related or not. Hard to tell from our perspective.


edit:

This is why IRC is often much better for troubleshooting these sorts of problems. Question and answers happen realtime and there is much less chance for real confusion...

 

[Nothinman]

Speaking of IRC, if he wants to try again in "real-time" he can stop by #linux irc.arstechnica.com, there's almost always someone willing to help out there.

 

[silverpig]

Originally posted by: drag
Also if that is a cdrom drive you can get funky things like that if you have a disk in there that it's trying to access, but has a error on it.

Yeah it's odd. I've had that problem too.

 

[nweaver]

could the NTFS driver be causing the hard locks? I know it's not "stable", but I'm not sure what that means, as I have never had to load the NTFS module.


Wait, I take that back, Ubuntu on my lapotop does RO ntfs for my windows partition.

 

[Nothinman]

could the NTFS driver be causing the hard locks? I know it's not "stable", but I'm not sure what that means, as I have never had to load the NTFS module.

Anything that runs in the kernel can do anything and with a filesystem hard locks are the nicest thing for it to do when it crashes.

 

[fuzzybabybunny]

Originally posted by: silverpig

Originally posted by: fuzzybabybunny
I actually don't have any idea which version it is...

uname -m

🙂

And the fresh format is a good idea. I mean, it'll work with what you want, but just in case you decide to write to the drives remotely, it'd be beneficial to use a native linux filesystem.

It's x86_64 🙁

 

[Nothinman]

It's x86_64

Grab the i386 image and reinstall with that one then, it might not make a difference but at this point it can't hurt =)

 

[nweaver]

but bottom line is to get off NTFS, there is NO reason to run NTFS on a linux server. It's bad bad bad stuff.

 

[silverpig]

Either way, I applaud your resolve with having problems yet still trying 🙂