Re: usb hard disk / ext3 partition - permissions change automatically(!) from read-write to read-only
by Ilias Paraponiaris May 26, 2007; 07:14pm
I would appreciate any help to the following issue:
My system is debian-testing (not stable=etch).
I use linux exclusivelly only for 3 months and I have no programming
background, so I am rather newbe than expert.
I have an external usb hard disk Seagate "FreeAgent" 250GB connected to
my 5 years old laptop (which supports usb1.0 not usb2.0). I have created
two partitions, the original ntfs (preformatted by the manufacturer) and
a 30GB ext3 partition that I created on my own.
=> Regarding the 200GB ntfs partition i) when it is automounted (without
any entry in fstab) I have read-only permissions, but ii) when it is
mounted with an entry in fstab ("/dev/sda1 /media/usbntfs ntfs-3g
defaults,force,locale=en_US.utf8 0 2") I have read-write permissions and
it works perfectly (I use the "force" option because ntfs-3g is not
supported sufficiently by the installed kernel 2.6.18-4-686 and asks for
kernel 2.6.20 or newer. Do you think that it is "safe" and "recommended"
to compile my own vanilla kernel? What how-to, guides, etc do you
propose?). A minor issue is that in the second case I see the ntfs
partition in /media/usbntfs mounted, but in desktop/"computer" I see the
"FreeAgent" unmounted, I see the ext3 partition mounted, but I do not
see at all the ntfs partition (why? how can I fix it?).
=> Regarding the 30GB ext3 partition both i) when it is automounted
(without any entry in fstab) and ii) when it is mounted with an entry in
fstab ("/dev/sda2 /media/usbext3 ext3 defaults,users,auto,rw 0 2") I
constantly face the same issue: IMMEDIATELY AFTER THE PARTITION HAS BEEN
MOUNTED I HAVE READ-WRITE PERMISSIONS BUT WHEN THE EXTERNAL HARD DISK
REMAINS IDLE FOR A FEW MINUTES THEN MY PERMISSIONS CHANGE AUTOMATICALLY
TO READ-ONLY!!! If I unmount/mount the partition, again iniatially I
have read-write permissions and after a wile they change to read-only.
It is interesting that the issue has to do only with the ext3 partition,
not with the ntfs one although they are both in the same hard disk!
In case this is a bug or in case I have to address this issue to another
list, please advise.
I have the same drive which I completely reformated to ext3, and I am
having the same problem.
Here are some relevant log entries reported by logcheck:
Security Events for kernel
=-=-=-=-=-=-=-=-=-=-=-=-=-
May 27 08:07:40 corvus kernel: end_request: I/O error, dev sda, sector 208748607
May 27 08:09:06 corvus kernel: end_request: I/O error, dev sda, sector 208748607
May 27 08:09:06 corvus kernel: end_request: I/O error, dev sda, sector 208748607
May 27 08:09:06 corvus kernel: end_request: I/O error, dev sda, sector 208748607
System Events
=-=-=-=-=-=-=
May 27 08:07:40 corvus kernel: sd 1:0:0:0: Device not ready: <6>: Current: sense key: Not Ready
May 27 08:07:40 corvus kernel: Additional sense: Logical unit not ready, initializing command required
May 27 08:07:40 corvus kernel: ext3_abort called.
May 27 08:07:40 corvus kernel: EXT3-fs error (device sda1): ext3_journal_start_sb: Detected aborted journal
May 27 08:07:40 corvus kernel: Remounting filesystem read-only
May 27 08:09:06 corvus kernel: sd 1:0:0:0: Device not ready: <6>: Current: sense key: Not Ready
May 27 08:09:06 corvus kernel: Additional sense: Logical unit not ready, initializing command required
May 27 08:09:06 corvus kernel: sd 1:0:0:0: Device not ready: <6>: Current: sense key: Not Ready
May 27 08:09:06 corvus kernel: Additional sense: Logical unit not ready, initializing command required
May 27 08:09:06 corvus kernel: sd 1:0:0:0: Device not ready: <6>: Current: sense key: Not Ready
May 27 08:09:06 corvus kernel: Additional sense: Logical unit not ready, initializing command required
---------------------
Security Events for kernel
=-=-=-=-=-=-=-=-=-=-=-=-=-
May 27 11:31:20 corvus kernel: end_request: I/O error, dev sda, sector 12375
May 27 11:31:20 corvus kernel: end_request: I/O error, dev sda, sector 8279
May 27 11:31:20 corvus kernel: end_request: I/O error, dev sda, sector 12423
Security Events
=-=-=-=-=-=-=-=
May 27 11:31:26 corvus kernel: EXT3-fs error (device sda1) in ext3_reserve_inode_write: IO failure
May 27 11:31:26 corvus kernel: EXT3-fs error (device sda1) in ext3_dirty_inode: IO failure
May 27 11:45:47 corvus kernel: EXT3-fs warning (device sda1): ext3_clear_journal_err: Filesystem error recorded from previous mount: IO failure
System Events
=-=-=-=-=-=-=
May 27 11:31:20 corvus kernel: sd 2:0:0:0: Device not ready: <6>: Current: sense key: Not Ready
May 27 11:31:20 corvus kernel: Additional sense: Logical unit not ready, initializing command required
May 27 11:31:20 corvus kernel: sd 2:0:0:0: Device not ready: <6>: Current: sense key: Not Ready
May 27 11:31:20 corvus kernel: Additional sense: Logical unit not ready, initializing command required
May 27 11:31:20 corvus kernel: EXT3-fs error (device sda1): ext3_get_inode_loc: unable to read inode block - inode=2, block=1027
May 27 11:31:20 corvus kernel: Aborting journal on device sda1.
May 27 11:31:20 corvus kernel: sd 2:0:0:0: Device not ready: <6>: Current: sense key: Not Ready
May 27 11:31:20 corvus kernel: Additional sense: Logical unit not ready, initializing command required
May 27 11:31:20 corvus kernel: Buffer I/O error on device sda1, logical block 1545
May 27 11:31:20 corvus kernel: lost page write due to I/O error on sda1
May 27 11:45:47 corvus kernel: EXT3-fs warning (device sda1): ext3_clear_journal_err: Marking fs in need of filesystem check.
May 27 11:45:47 corvus kernel: EXT3-fs warning: mounting fs with errors, running e2fsck is recommended
----------------------------------
I notice the line:
"Logical unit not ready, initializing command required"
I wonder if there is a mount option to use to insure that an
initializing command is sent before other commands.
If I unmount and disconnect the drive, then reconnect and mount it,
things work until some unspecified period of inactivity, then blow up
when I try to read or write.
Initially it was on an usb port on the MB. I put it on a usb2/pci
card. That didn't change anything. I've built the file system three
times (once with "-c -c" - no bad blocks reported).
Any help is appreciated. Thanks.
Don
--
Don.Hayward at pomobuli.net -- debian/rules -- Registered Linux user #374806
- History is the best antidote to illusions of omnipotence and omniscience.
- It should forever remind us of the limitations of our passing perspectives.
- It should strengthen us to resist the pressure to convert momentary
- interests into moral absolutes. -- Arthur Schlesinger Jr.
Reply to: