Unfortunately no one can be told what fun_plug is - you have to see it for yourself.
You are not logged in.
Hi!
I'm running Debian chroot on a USB-Stick... when the system get's under heavy load, like an rsync of thousands of files or compiling the kernel via "make -j4" the DLINK often freezes...
When this happens you can't logon via ssh anymore... you'll get the error "ssh_exchange_identification: Connection closed by remote host".
Using the serial console I noticed a lot of those messages popping up...
----------
SCSI error : <4 0 0 0> return code = 0x70000
end_request: I/O error, dev sdc, sector 559173
----------
breaking into the shell I noticed that my USB stick got unmounted automatically.
I tried a lot of different USB sticks since then - but it doesn't seem to make any difference.
Did anyone else ever experience this problem? And if... were you able to solve it??
Thanks in advance!
Paul
Offline
WHY do you do a make -j4?!?!?
This is a low end box and has only a cpu with one core. -j4 will create 4 threads. No wonder your box is giving up.
Offline
*g* the -j4 was just for me trying to reproduce the error - because it doesn't always happen during rsync... I run daily backups from my servers on that device - and about once a week it happens.
Offline
I run my box from a cheapo USB stick. I have not seen anything like this.
I suppose you are running the usb-storage module without debugging?
If not try that instead.
Offline
Yes I'm running usb-storage without debugging...
also tried my own kernel with built-in usb-storage and some other small modifications - but there was no difference :-(
maybe I give this patch which was included in 2.6.13 a try...
commit 86dbde9cbdfe8bc2c2dfe5d33027d3acc55e0470
Author: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Date: Mon Jun 6 17:22:42 2005 -0700
[PATCH] USB Storage: retry hard errors
This patch started life as as527, and was rediffed by me.
Since the IDE interface doesn't convey much information about types of
errors, many USB-IDE adapters report all low-level errors with SK = 0x04,
which is supposed to be used only for non-recoverable errors. As a result
the SCSI midlayer doesn't retry the command. But quite often a retry
would succeed, whereas an unnecessary retry doesn't really hurt anything.
This patch uses a recently-implemented flag to tell the SCSI midlayer that
such hardware errors should be retried.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Matthew Dharm <mdharm-usb@one-eyed-alien.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Last edited by Paul (2007-09-28 08:21:41)
Offline
Well I incorporated all 3 or 4 patches concerning usb-storage from 2.6.13 into 2.6.12.6 and it kinda looks like my box is more stable now...
It's been running almost a week without the described error. the only thing I noticed was an USB reset once (which I guess would be the fatal error I had otherwise).
[196405.570000] usb 1-1: reset high speed USB device using ehci_platform and address 2
I'll do some more testing ...
Offline