Printable View
Ok heres the deal, have a dual boot system running Win200 Pro & Psyche Kernel 2.4.20-20.9 on an i686
hda1 39205 MB ntfs
hdb
hdb1 102 MB ext3
hdb2 56949 MB ext3
hdb3 188 Linux-swap
The problem that I am having is this;
mount /dev/hda1 /mnt/win -t ntfs -r -o umask=02222
now I can access hda, can read hda, can copy from it, but I cant write to it ?
any idea why????
not essential but would like to understand why (But thats what were here for is it not)
Well I am NOT the best person to answer this, I suppose, as my *nix is rather out of date :( .
As I understand it various versions of *nix either won't write to an NTFS partition, or you shouldn't try it because of the danger of data corruption and/or screwing your partition.
I have never heard of your linux brand, but would suggest that it is "smart" enough to detect the NTFS partition and is blocked against attempting to write to it because of the issues I just mentioned? I think that the inhibitor may well be in your linux distro?
You can, of corse set up a FAT32 partition which linux should be able to use without problem, as will WIN 2000. So that might be the way to go for shared files?
As I warned you, that is a bit out of date, but may still be the case?
Cheers
Linux NTFS Project should answer your question.
You beat me to it MsMittens; good job.
I also couldnt mount my ntfs partitions and did a lot of searching and found the website MsMittens mentioned.
Solved all my problems.
http://linux-ntfs.sourceforge.net/info/ntfs.html#4.10 Tells you how to mount the partitions automatically at startup.
EXCELLENT, EXCELLENT
Thats just the kind of info that I needed, thxs
nihil
Psyche = Redhat 8.0
But is it nemore I have made so many changes to the base that it might not be 8.0 ;)
The options may be different in your flavor of unix, but in solaris '-r' means mount read-only...Quote:
Originally posted here by s33ka
Ok heres the deal, have a dual boot system running Win200 Pro & Psyche Kernel 2.4.20-20.9 on an i686
hda1 39205 MB ntfs
hdb
hdb1 102 MB ext3
hdb2 56949 MB ext3
hdb3 188 Linux-swap
The problem that I am having is this;
mount /dev/hda1 /mnt/win -t ntfs -r -o umask=02222
now I can access hda, can read hda, can copy from it, but I cant write to it ?
any idea why????
not essential but would like to understand why (But thats what were here for is it not)
-r Mount the file system read-only.
Just a thought...
Perhaps but the problem with Linux is that it doesn't fully support write (basically do it at your own risk and potentially pooch your system). Solaris may fully support NTFS but I wouldn't risk it on anything but a full play box.Quote:
-r Mount the file system read-only.
Actually, I don't think Solaris supports NTFS, at least not natively.Quote:
Originally posted here by MsMittens
Perhaps but the problem with Linux is that it doesn't fully support write (basically do it at your own risk and potentially pooch your system). Solaris may fully support NTFS but I wouldn't risk it on anything but a full play box.
mount_cachefs(1M), mount_hsfs(1M), mount_nfs(1M), mount_pcfs(1M), mount_tmpfs(1M), mount_ufs(1M),
Are what I see in there right now, but that wasn't really the point. I was merely trying to state the obvious of he/she mounted the partition in read-only mode, so of course they couldn't write. Whether or not that was a good idea or supported, I left for someone else to talk about :)
I just updated my kernel to 2.6.4 on my slack box, and NTFS write support was not marked experimental in menuconfig. It was my understanding that it was, and this thread backs this up. I did not compile it in and would not need to as all my partitions are either FAT32 or ext3, so has anyone actually tried the NTFS write support in 2.6.x? If it's not supposed to work fully yet, shouldn't they put "(EXPERIMENTAL)" next to it like they do in all other experimental features?
You'd figure that knowing the risks before compiling it in and using it would be pretty important!
TheRepublican, this thread at LinuxQuestions.org might answer your question better. It's reiterates what is on the Linux NTFS Project's website:
Quote:
This NTFS driver was written by Anton Altaparmakov from scratch, supporting all versions of NTFS. It is included in the current development kernels since 2.5.11 and also has a backported version for the latest 2.4 stable kernel. What are the advantages of the new driver compared to the old one?
* Stable: no known problems in the implemented functionality, it's also multi-processor and reentrant safe.
* Supports all NTFS cluster sizes from 512 bytes up to 64 kB. The old driver is limited to maximum 4 kB cluster size.
* Full support for sparse and compressed files also on Windows XP/2003/2000.
* Supports mmap() thus Wine users can start applications from an NTFS partition.
* Still read-only, but with safe file overwrite support on all Windows versions without changes to the file size and if the file is over 1 kB.
* One can setup a loopback device on an NTFS file. TopologiLinux and others use this feature to run Linux from a Windows NTFS partition with full read-write support.
* Much better performance.
Hmmm,
I guess it goes back to my original post?
AFAIK it is safe to use FAT32 but you use *nix to write on NTFS AT YOUR OWN RISK, AND WITH NO RECOURSE?
I have a .308 Winchester........................it has no "experimental" warnings on it either :D .......open source stuff is usually covered by a general liability disclaimer anyway?....I would generally translate "experimental" as something like "alpha" or "beta" release????
I wonder if SCO realise the potential liabilities that come with trying to hijack open source? :p
If you are going to do this, my friend, I suggest you make sure that you have a nice mirror backup, or do it on something that doesn't matter ;) I do not believe that the technology is quite developed enough?
It is your choice.................
Good luck
EDIT: I have looked at some of my kit and had some thought on how I might try to do it safely:
NTFS partition #1.....................Windows OS
NTFS partition #2.....................Windows applications
NTFS partition #3.....................Windows Data
NTFS aprtition #4.....................Mirror of Windows Data
Linux Partitions.........as required
I would then let the *nix share the "mirror" (NTFS#4) data file (NOT a RAID1 array, BTW.........just a copy)
Until you are happy that you have something stable that won't corrupt? In any case there is absolutely no need for the *nix to access the first two NTFS partitions?
Cheers
Looks like I started quite a decent thread, for a change ;)
Thanks guys , to be honest I think my best bet would be to re-biuld my MS box with FAT32, never did like NTFS anyway but then who does.... It was meant to be tighter security wise is it I never noticed Muhahaha ;)
Quote:
Originally posted here by s33ka
Looks like I started quite a decent thread, for a change ;)
Thanks guys , to be honest I think my best bet would be to re-biuld my MS box with FAT32, never did like NTFS anyway but then who does.... It was meant to be tighter security wise is it I never noticed Muhahaha ;)
I wouldn't scoff at the security benefits to NTFS. Without NTFS there are no such thing as ownership rights and limitations on file access (whala Win95), along with all of the other security benefits such things provide, assuming the person doing the admin work uses them properly. I have also found NTFS to be a more stable and reliable FS than FAT32; however, you do of course have the problem of a limited number of OSs that read/write it properly. If you are strictly going to use it in a Win2k/XP/2k3 environment, I wouldn't use anything but NTFS....